Intensive Care Society State of the Art 2018 Abstracts

Selected abstracts were presented as orals within the conference programme and the remaining accepted abstracts were presented as electronic posters.

Abstracts selected as oral presentations

Winner of the 2018 SOA Oral Presentation Award

Computed tomography measurement of skeletal muscle mass in adults receiving veno-venous extracorporeal membrane oxygenation

Danielle Bear^1,2,3,4, Liisa Macgowan⁵, Zudin Puthucheary^6,7, Rebeka Wright¹, Luigi Camporota^1,4, Stephen Harridge⁴, Nicholas Hart^3,4,8, Kevin Whelan⁹ and Nicholas Barrett^1,4

¹Department of Critical Care, Guy's and St Thomas' NHS Foundation Trust, London, UK

²Department of Nutrition and Dietetics, Guy's and St Thomas' NHS Foundation Trust, London, UK

³Lane Fox Clinical Respiratory Physiology Research Centre, Guy’s and St. Thomas’ NHS Foundation Trust, London, UK

⁴Centre for Human and Applied Physiological Sciences, King’s College London, London, UK

⁵Department of Nutrition and Dietetics, Imperial College London, London, UK

⁶Centre for Human Health and Performance, Department of Medicine, University College London, London, United Kingdom

⁷Adult Intensive Care Unit, Royal Free NHS Foundation Trust, London, UK

⁸Lane Fox Respiratory Service, Guy's and St Thomas' NHS Foundation Trust, London, UK

⁹Department of Nutritional Sciences, School of Life Course Sciences, King's College London, London, UK

Abstract

Background: There has been a growth in the use of veno-venous extracorporeal membrane oxygenation (vv-ECMO) for severe acute respiratory distress syndrome over recent years. However, there are limited data available on nutritional status and the impact on outcome in these patients.

Objectives: 1. To investigate the prevalence of low muscle mass, using computed tomography (CT) scans at the third lumbar vertebra (L3), on commencement of vv-ECMO. 2. To investigate the relationship between muscle mass at the commencement of vv-ECMO and length of stay in hospital prior to ECMO commencement (LOSpreECMO), age, gender, body mass index (BMI), number of co-morbidities, APACHE II and SOFA score.

Methods: Patients receiving vv-ECMO between September 2010 – June 2017 were retrospectively screened for eligibility. Eligibility criteria were adult (³18 years) at commencement of vv-ECMO and a minimum of 72 hours on ECMO support with a CT scan at the level of L3 of sufficient quality to analyse. Skeletal muscle cross-sectional area (SMA cm²) was determined using Slice-O-Matic V 5.0 (TomoVision, Montreal, Canada) and compared with published criteria (low SMA defined as males <170 cm² and females <110 cm²). Demographic and anthropometric data were collected from electronic medical records (Phillips, ICIP, Baltimore, USA). Data are presented as mean (SD) or number (%). Chi-squared or independent t-tests were used as appropriate. Backwards stepwise (Wald) logistic regression was performed to determine the relationship between presence of low SMA and LOSpreECMO, age, co-morbidities, gender, body mass index (BMI), number of co-morbidities, APACHE II and SOFA score.

Results: A total of 215 patients (age, 45.2 (13.7), 56.7% male, APACHE II score 18.2 (4.2), SOFA score 7.9 (4.3) BMI 29.5 kg/m² (7.8), LOSpreECMO 5.0 (5.2) days, duration of vv-ECMO 14.2 (17.1) days) were included in the analysis. The overall prevalence of low SMA on commencement of vv-ECMO was 51.6%. This was higher in males than in females (62.3% vs. 37.6%, X²= 31.28, p < 0.001). LOSpreECMO, BMI and gender were associated with the presence of low muscle mass on commencement of vv-ECMO (Table 1). There were no associations observed with number of co-morbidities, age, APACHE II or SOFA score. Patients with low muscle mass required vv-ECMO for a longer period (16.4 (21.7) vs. 11.8 (9.8) days, p = 0.046)

Conclusion: Over 50% of patients had low SMA on commencement of vv-ECMO. Length of stay in hospital prior to ECMO, BMI and gender (male) were all predictive of low SMA on commencement of vv-ECMO. Patients with low muscle mass required vv-ECMO for a longer duration. These associations should be explored further.

A review of deaths within forty-eight hours of ICU admission using the Royal College of Physicians Structured Judgement Review method

Matthew Luney, Emma Bryden and Clare Stapleton

Intensive Therapy Unit, Wexham Park Hospital, Frimley Health NHS Foundation Trust, Slough, UK

Abstract

Introduction: In our ICU we had noted an increase in 48 hour mortality rate (8.7% to 10.9% of admissions). The reasons we considered were: the admission of patients who were inevitably going to die and therefore should perhaps not have been admitted, or a problem in the quality of care delivered on Intensive Care. To help answer this question we performed a retrospective case record review of a proportion of these patients.

Table 1.

Logistic regression analysis for low SMA.

	OR	95% CI	P-value
Gender (male)	0.327	0.17–0.622	0.001
BMI	0.864	0.817–0.913	0.0005
LOSpreECMO	1.092	1.092–1.017	0.015

CI = confidence interval, OR = odds ratio, BMI = body mass index; LOSpreECMO = days in hospital prior to commencement of vv-ECMO.

The Royal College of Physicians’ Structured Judgement Review (SJR) tool¹ is designed to assess the quality of care delivered to adult patients. It is the basis for mortality review nationally. It allows quantitive and qualitative judgements to be made and the review to highlight good practice and care that should be improved.

Methodology: Patients who died within 48 hours of admission to ICU in the 12 months until 30^th March 2018 were identified. The cases were taken from each month and had a wide spread of Apache II scores. The reviewers assessed admission and initial management, ongoing care, perioperative care, and end-of-life care. Explicit judgements were recorded as well as a numerical score for each phase of care and for the overall care.

Results: 96 patients died within 48 hours, representing 12.7% ICU admissions and 53.6% of all ICU deaths. They were all emergency admissions. Median age 72 years (IQR60-81), 60% male, median APACHE II 25 (IQR20-29). 48 patients were selected, full case notes were available for 33 patients. The commonest reasons for admission to ICU were in- and out- of hospital cardiac arrest n = 32 (33.3%), pneumonia n = 21 (21.8%) and catastrophic brain injury n = 13 (13.5%). Of the 33 cases reviewed 15 were found to have excellent or good overall care, but 13 were found to have a significant problem in care. The themes identified were: under resuscitation despite persistent physiologic derangements, delay in administration of antibiotics, failure to recognise/respond to blood gas abnormalities, and delay in review by relevant clinicians. These all occurred prior to admission to ICU. There was also 2 cases when ICU admission was delayed and a deterioration which was not recognised while on ITU. These cases had also been reviewed as part of the Trusts mortality review programme and the reviewers had indentified similar themes.

Discussion: A significant proportion of these patients were admitted for post resuscitation care or as part of the catastrophic brain injury pathway. Many of these deaths were not unexpected and the patients received good end of life care. There were however cases where a problem in care was identified the majority of these occurring prior to admission to ICU.

Conclusion: This review confirmed to us that the SJR tool is an effective and practical way of assessing quality of care delivered to our patients. We did not identify a recurring or significant problem in ICU care delivery or senior decision making. However there was learning to disemminate to ward teams particularly concerning response to deteriorating patients. This is part of a larger improvement plan within our trust.

Post-traumatic stress disorder and debriefing after in-hospital cardiac arrests

Stephen Spencer¹, Mike Osborn², Jerry Nolan² and Andy Georgiou²

¹North Bristol NHS Trust, Bristol, UK

²Royal United Hospital, Bath, UK

Abstract

The incidence of in-hospital cardiac arrest (IHCA) attended by a resuscitation team is estimated at 1.5 per 1000 hospital admissions in the UK (National Cardiac Arrest Audit).¹ Health-care staff working in intensive care units (ICUs), acute medical units (AMUs) and emergency departments (EDs) who attend IHCAs are therefore exposed to potentially traumatic experiences as a routine part of their work. Whilst it is recognised that traumatic experiences can result in stress symptoms and post-traumatic stress disorder (PTSD), the true impact of attending IHCA on healthcare workers has not been studied. Debriefing after IHCA is advocated by the International Liaison Committee on Resuscitation (ILCOR) and the Resuscitation Council UK (RCUK) for learning and improvement,² and can provide some psychological support.³ This cross-sectional study examined cardiac arrest debriefing practices and the psychological burden of attending IHCA on both nursing and medical staff who work in acute areas of the hospital.

This was a single centre study set in an acute, 732 bed district general hospital in the UK, between April-August 2018. A 33 questionnaire-survey was sent to 517 doctors (of all grades), nurses and health-care assistants (HCAs) who worked in either the ED, AMU or ICU. The survey had three sections: demographics; cardiac arrest and debriefing practices; and the trauma-screening questionnaire (TSQ) as a validated screening tool for PTSD.⁴

Four hundred and fourteen staff responded to the questionnaire (response rate of 80.1%), of whom 312/414 (75.4%) had been involved with an IHCA. A total of 1463 cardiac arrests had been experienced by 312 staff in the preceding 12-months and 258/1463 (17.6%) were debriefed. Fatal cardiac arrests were debriefed more frequently than non-fatal cardiac arrests (202/1079 (18.7%), 56/384 (14.6%) respectively) and 196/312 (62.8%) of staff had never been to a debrief following an IHCA. The TSQ, answered by 300 staff identified that 29/300 (9.67%) had PTSD and 168/300 (56.0%) had one or more symptoms of trauma. Debriefing was not associated with increased risk of PTSD (p = 0.931). HCAs and FY1s had significantly higher TSQ scores than nurses or more senior doctors (p = 0.039, p = 0.046, respectively). Having a break after an IHCA was associated with significantly lower TSQ scores than never having taken a break (p = 0.027). Of the resuscitation team leaders, 8/67 (11.9%) had previously been trained in debriefing techniques and 53/62 (85.5%) said debriefing training would be useful and would increase the frequency with which they would instigate a debrief (55/63 (87.3%)).

Our results reveal that approximately 10% of acute care staff have suffered PTSD as a result of attending an IHCA, with HCAs and junior doctors being most at risk of developing trauma symptoms. Furthermore, a debrief occurred in only 17.6% of arrests, possibly because of very little debrief training amongst cardiac arrest team leaders. This study suggests that more support is required for acute care nursing and medical staff following an IHCA. Training in the effective delivery of a debrief may be the first step to provide such support. Further study in this area warranted.

Hyperoxia is not associated with increased ICU mortality in patients with ARDS

Susanna Gilliland¹, Andrew Boyle^1,2, David Holmes^1,2, Stefania Di Gangi³, Cecilia O'Kane², Paul Young⁴ and Danny McAuley^1,2

¹Regional Intensive Care Unit, Royal Victoria Hospital, Belfast, UK

²Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

³Institute of Primary Care, University Hospital Zurich, Zurich, Switzerland

⁴Department of Intensive Care Medicine, Wellington Regional Hospital, Wellington, New Zealand

Abstract

Background: Emerging data suggests there may be harm associated with hyperoxia and excess oxygen exposure, however most studies of hyperoxia are in a general ICU population. Therefore, we performed an observational study to evaluate the hypothesis that hyperoxia is associated with harm in patients with ARDS.

Methods: Patients with ARDS were prospectively identified from a single, mixed medical and surgical intensive care unit (ICU). Demographic, clinical and laboratory variables were recorded retrospectively. Oxygen exposure variables were collected at 6-hourly intervals, for a maximum of 7 days, censored at the discontinuation of assisted breathing. The primary exposure was the average time-weighted PaO₂ [3]. This was calculated over a maximum of 7 days from meeting ARDS criteria. Secondary exposure variables included a) number of hyperoxia episodes (defined as a PaO₂ > 14 kPa) per 24 hours, b) excess oxygen exposure episodes (defined as any FiO2 value >0.5 in patients with a PaO2 > 10.7 kPa calculated at 24-hour intervals) and c) cumulative excess oxygen exposure. The primary outcome was ICU mortality. Univariable and multivariable logistic regression analysis were performed to assess the impact of exposure variables on clinical outcomes. Results are presented as odds ratio [95% confidence interval] and median [interquartile range].

Table 1.

Multivariable analysis for ICU mortality.

	Odds Ratio [95% CI]	p-value
Average time-weighted  PaO2 (kPa)	0.81 [0.66, 0.99]	0.05
Age (years)	1.04 [1.02, 1.06]	0.001
Lactate (mmol/L)	1.19 [0.99, 1.45]	0.08
Non-respiratory SOFA	1.08 [0.97, 1.19]	0.15
PEEP (cmH2O)	0.89 [0.78, 1.01]	0.08

Results: In the 203 patients included in the analysis overall ICU mortality was 34%. Average time-weighted PaO₂ was higher in ICU survivors vs. ICU non-survivors (12.0 [10.9–12.7] vs. 11.3 [10.2–12.5]; p = 0.03). In multivariable analysis, higher average time-weighted PaO₂ was associated with lower ICU mortality (OR 0.81 [0.66, 0.99]; p = 0.05) (Table 1). The number of hyperoxia episodes per patient within the first 7 days of ARDS was low across the study population, although more frequent in ICU survivors (survivors 4 [1–6], non-survivors 2 [0–4]; p = 0.01). In multivariable analysis, the number of hyperoxia episodes per 24-hour period was not associated with ICU mortality (0.75 [0.44, 1.23]; p = 0.27). Between ICU survivors and non-survivors, there was no difference in the proportion of patients exposed to an episode of excess oxygen (54% vs. 57%; p = 0.88). Similarly, there was no difference in the cumulative excess oxygen exposure between survivors and non-survivors (survivors 0.1 [0.0–0.3], non-survivors 0.1 [0.0–0.2] kPa; p = 0.82).

Conclusion: Contrary to our a priori hypothesis, hyperoxia in the first seven days following a diagnosis of ARDS was not associated with increased ICU mortality. Within our cohort there were few episodes of hyperoxia, and limited excess oxygen exposure, suggesting that oxygen is already titrated to more conservative targets within our ICU.

Acetate versus Lactate buffered balanced Infusates on Hemodynamic Stability in Patients undergoing Cardiac Surgery – a randomized controlled double-blind Trial

Carmen Pfortmueller¹, L. Faeh¹, M. Müller^2,3, B. Eberle⁴, H. Jenni⁵, B. Zante⁵, J. Prazak¹, L. Englberger⁵, J. Takala¹ and S.M. Jakob¹

¹Department for Intensive Care Medicine, Bern University Hospital, University of Bern, Bern, Bern, Switzerland

²Institute of Health Economics and Clinical Epidemiology, University Hospital of Cologne, Cologne, Germany

³Department for Emergency Medicine, Bern University Hospital, University of Bern, Bern, Switzerland

⁴Department of Anesthesiology and Pain Medicine, Bern University Hospital, University of Bern, Bern, Switzerland

⁵Department of Cardiovascular Surgery, Bern University Hospital, University of Bern, Bern, Switzerland

Abstract

Patients undergoing cardiac surgery require several hours of hemodynamic support. Recent evidence suggests that the type of crystalloid fluid used peri-operatively might influence the need for vasoactive agents. Acetate instead of lactate as a buffer molecule has been controversially discussed for many years, with vasodilatation and metabolic alkalosis as the main arguments against its use (1). However, recently two investigations by our group pointed towards acetate-buffered infusions being more beneficial for hemodynamic stabilization in patients undergoing major surgical procedures than 0.9% saline (3). We therefore hypothesized that use of an acetate-based balanced crystalloid solution for perioperative fluid replacement in patients undergoing cardiac surgery results in a lower dose of inopressors, in a shorter total time on inopressors, and in a lesser amount of fluid needed to achieve hemodynamic stability, than use of a lactate-buffered crystalloid solution.

Using a randomized controlled double blind design, we compared acetated Ringer’s (RA) to lactated Ringer’s (RL) with respect to the cumulative dose of inopressors administered until postoperative hemodynamic stabilization. Secondary outcomes were the duration of inopressor administration, the total fluid volume administered, and changes in acid-base homeostasis. Patients undergoing elective valvular +/− bypass surgery were included. Patients with severe cardiac (EF < 30%), renal (eGFR < 30 ml/min/1.73m²), liver dysfunction (bilirubin >3 mg/dl), florid endocarditis or chronic inflammatory disease were excluded from the analysis.

Seventy-five patients were randomly allocated to the RA arm, 73 to RL. Groups were comparable with in respect to age (median 67.5, IQR 58.0−72.5), sex (males = 111, 75.7%), or ASA (ASA IV = 145, 98.0%). In the RL-group more patients underwent a composite graft operation whereas in the RA-group more patients received a single valve replacement (p = 0.024).

Table 1.

Primary and secondary endpoints – vasoactive medication.

Primary endpoint	Acetated Ringer’s (n = 75)	Lactated Ringer’s (n = 73)	p value
	median (IQR)	median (IQR)
Cumulative dose of inopressors (norepinephrine and epinephrine) [µg/Kg/h] until hemodynamic stabilization	2.1 (0.5–8.1)	1.7 (0.7–8.2)	0. 989
Secondary endpoints – vasoactive medication
Cumulative dose of norepinephrine [µg/Kg/h] until hemodynamic stabilization	1.8 (0.5–6.7)	1.5 (0.6–4.7)	0.672
Cumulative dose of epinephrine [µg/Kg/h]* per hour until hemodynamic stabilization	0.0 (0.0–0.0)	0.0 (0.0–0.1)	0.251
Cumulative dose of inopressors (norepinephrine and epinephrine) [µg/Kg/h] per hour on inopressors	4.1 (1.0,11.8)	3.7 (1.6–12.1)	0.959
Cumulative dose of norepinephrine [µg/Kg/h]# per hours on norepinephrine	4.1 (1.1,11.7)	3.4 (1.5–9.5)	0.907
Cumulative dose of epinephrine [µg/Kg/h]§ per hour on epinephrine	0.2 (0.1–9.3)	4.5 (0.4–33.0)	0.047
Cumulative dose of inopressors [µg/Kg]	22 (5–83)	20 (7–114)	0.928
Cumulative dose of norepinephrine [µg/Kg]	19 (5–71)	16 (6–61)	0.726
Cumulative dose of epinephrine [µg/Kg]	22 (5–83)	20 (7–114)	0.928
Time on inopressors, [hours]°	4.8 (4.1–10.7)	6.1 (4.0–10.6)	0.505
Time on norepinephrine [hours]#°	4.8 (4.1–6.4)	4.5 (3.7–7.4)	0.836
Time on epinephrine [hours]*°	0.0 (0.0–0.0)	0.0 (0.0–3.4)	0.209
Cumulative dose of inodilators, ICU [µg]	0 (0.0–0.0)	0 (0.0–0.0)	0.375
Time on inodilators, ICU [min]	0 (0.0–0.0)	0 (0.0–0.0)	0.386
Cumulative dose of vasodilatators [mg]	0 (0.0–2.8)	0 (0.0–4.2)	0.844
Time on vasodilatators, ICU [hours]	0 (0.0–1.7)	0 (0.0–6.4)	0.726
Secondary endpoints – fluids
Total amount of study fluid received (ml)	6,677 (5,325–8,479)	6,103 (4,769–7,855)	0.272
Total amount of fluid received other than study fluid during study period (ml)	393 (112–972)	325 (139–697)	0.811
Total amount of fluid received after initial hemodynamic stabilisation until ICU discharge (ml)	691 (416–1,655)	756 (456–1,276)	0.814
Total amount of fluid received from the start of anesthesia to ICU discharge (ml)	7,761 (5,853–11,106)	7,184 (5,364–9,828)	0.658

^*not including four patients who received one single bolus injection of epinephrine only.

^#not including one patient who received one single bolus injection of norepinephrine only.

⁺any type.

°from intubation to end of study period defined.

The groups did not differ in respect to the cumulative dose of inopressors (RA 2.1 mcg/kg/h, IQR 0.5–8.1 vs. RL 1.7 mcg/kg/h, IQR 0.7–8.2, p = 0.989), see table 1. The evolution of blood pressure, peripheral temperature and urinary output was similar in the acetated Ringer’s and the lactated Ringer’s group. Sensitivity analysis of inopressors showed that all patients received norepinephrine infusions (n = 148, 100%) and 41 (27.7%) of all patients received epinephrine infusions at some time during the perioperative period (for both p > 0.05 between groups). The total norepinephrine dose per hour on norepinephrine did not differ between the RA and the RL group (median 4.1 mcg/kg/h, IQR 1.1–11.7, respectively median 3.7 mcg/kg/h, IQR 1.5–9.5, p = 0.907). However, patients in the RL-group received significantly higher dose of epinephrine per hour of epinephrine when compared to the RA-group (median 4.5 µg/kg/h, IQR 0.4–33.0 vs. median 0.2 µg/kg/h, IQR 0.1–9.3, p = 0.047). If considered only for the period in the ICU, more patients in the lactated Ringer’s group received epinephrine infusions (14 (19.2%) vs. 6 (8.0%), p = 0.047). However, the cumulative dose of epinephrine per hour on epinephrine during the perioperative period was significantly higher in the composite graft group (median 0.1 µg/Kg/h, IQR 0.0–0.6 vs median 0.0 µg/Kg/h, IQR 0.0–0.0, p = 0.008) as were as were the incidence of aortic insufficiency (p = 0.0001) and the duration of hypothermic cardiac arrest (p = 0.0001).

The time on individual and combined inopressors did not differ between the groups. No differences were found in acid-base parameters and their evolution over time. Specifically, there was no occurrence of metabolic alkalosis, see Figure 1.

The main finding of this study is that for use in perioperative fluid management of elective cardiac surgical patients, acetate-buffered balanced Ringer's solution does not differ significantly from lactate-buffered Ringer's with respect to postoperative hemodynamic stability and pharmacological support with inotropic agents. Also, we observed no difference in the cumulative amount of perioperative iv fluids necessary, or in acid-base profiles. More epinephrine use was noted in the group with acetate-buffered infusion, which was however related to a significantly larger proportion of patients with valvular aortic regurgitation and hypothermic cardiac arrests in this group, receiving more composite graft procedures.

Based on this study`s results, we now reject our hypothesis that acetate-buffered is superior to lactated Ringer’s solution with respect to perioperative hemodynamic stability. It is therefore likely that the effect seen in previous studies is due to normal saline-induced excessive chloride load and related acidosis.

Chloride excess was linked to adverse hemodynamic outcomes in several previous human and experimental studies (2, 3) have been correct in implying excess chloride as the culprit responsible for hemodynamic instability, it appears logical that this study, comparing two chloride-reduced balanced solutions, did not observe a difference in this regard.

The hemodynamic effects of acetate are controversially discussed. However, most of this evidence originates from investigations of either sodium-acetate infusion or acetate-buffered dialysis, where the acetate load is much higher than in fluid replacement with acetate-buffered Ringer's solution. Therefore such findings are not generalizable to populations receiving perioperative fluid therapy with acetate-buffered infusions (5).

In this study, the hemodynamic profile of patient groups receiving either RA or RL did not differ significantly, even though large quantities of iv fluid were administered. RA therefore appears as a feasible alternative to RL for fluid resuscitation in the critically ill from a hemodynamic viewpoint.

Another feared side effect of acetate-buffered infusion solutes is the occurrence of metabolic alkalosis. Our study shows that even in patients undergoing cardiac surgery who receive large quantities of intravenous fluid within a short time period, the use of an acetate-buffered infusion solution did not result in metabolic alkalosis. This is in accordance with a recent systematic review by our group on studies comparing acetate buffered solution to other crystalloids were major increases in bicarbonate with acetate buffered solutions were rarely observed (5).

In summary, acetate-buffered and lactated-buffered infusion solutions are equal for early hemodynamic stabilization in patients undergoing cardiac surgery.

Opioid-induced constipation in Intensive Care – how big is the problem?

Abigail Masding and Sundeep Kaul

¹Harefield Hospital, London, UK

Abstract

Background: Opioids are commonly used on the intensive care unit (ICU) but are associated with a high incidence of constipation, often resistant to conventional laxative treatment. Recently developed peripheral mu-opioid receptor antagonists (Naloxegol, Methylnaltrecone, Alvimopan) are new in the treatment of laxative-refractory opioid-induced constipation (OIC). We aimed to identify the prevalence of OIC within a large cardiothoracic ICU, impact on circulatory and ventilator support, cost of laxatives currently used in OIC treatment, and potential cost savings from targeted use of Naloxegol.

Figure 1.

Methods: Retrospective single-centre observational study looking at ICU admissions between October 2015 – September 2016.

Data collected on patient demographics, ICU length of stay (LoS), duration of organ support, one-year mortality, frequency of bowel movement, ventilator and circulatory support and laxative use. OIC defined as bowels not open (BNO) for four or more consecutive days in patients on regular opioids. Vasoactive-inotropic score (VIS) used as a measure of circulatory support. Laxative treatment cost per patient calculated using costs from the hospital’s pharmacy department multiplied by the number of laxative doses administered per patient until bowel opening (BO).

Results: 827 ICU admissions (61 ± 17 years, 68% male), of which 18% (n = 151) had OIC with an average BNO duration of 5.4 days (range 4–15 days). LoS, duration of organ support and one-year mortality rates were higher in the OIC group vs. all ICU admissions (Table 1). Circulatory and ventilator support decreased following BO in OIC patients with BNO duration ≥7 days (Table 2). 75% (n = 112) OIC patients required ≥3 laxatives to induce BO. An average of £5.55 was spent on laxatives per OIC patient until BO.

Table 1.

	LoS (days)	Duration of organ support (days)	1-year mortality n (%)
OIC group	15.4	73.3	64 (42)
All ICU  admissions	6.7	33.5	239 (29)

Table 2.

	24 hours pre BO	24 hours post BO	% decrease
Average VIS	56.4	40.8	29
Average peak  inspiratory  pressure  (cmH20)	24.2	19.8	18

Conclusion: Nearly 20% of ICU patients suffer from OIC which is associated with longer LoS, increased organ support and higher mortality rates. The targeted use of Naloxegol could reduce the prevalence of OIC, reduce ICU LoS, increase ICU patient throughput, reduce time on ventilator support and therefore associated risk of ventilator-associated pneumonia and multi-resistant infection and reduce mortality rates. The predicted cost saving of reducing ICU LoS is £1800 per day, and will have significant implications for pharmacy and ICU budgets.

Successful utilisation of a computer-based prompt to reduce the incidence of hyperoxaemia within a university hospital general critical care unit

Benjamin Shuker, Sarah Shuker and Christopher Bassford

University Hospital Coventry, Coventry, UK

Abstract

Introduction: The deleterious effects of oxygen toxicity have been well known for over a century.¹ However, it is only in recent years that there has been a shift toward using oxygen conservatively in healthcare due to emerging evidence associating its excess with harm.

We previously described a high incidence of excess oxygen therapy within our general critical care unit (GCCU).² We aimed to reduce this.

Method: Arterial blood gas (ABG) data is routinely collated using bedside software on GCCU. At the end of 2016 we implemented an alert highlighting PaO₂ ≥13.3 kPa as abnormal and educated staff about potential harms from excess oxygen. Twelve months following intervention, we retrospectively reviewed 18,530 ABG samples from 679 admissions from the 6-month period prior (July-December 2017).

Mean PaO₂ has been shown to be a useful metric for predicting poor outcomes associated with hyperoxaemia.³ Mean PaO₂ from the duration of each admission was calculated, and this was used to categorise admissions into three groups: hyperoxaemia, normoxaemia, and hypoxaemia. Hyperoxaemia was defined as PaO2 ≥13.3 kPa. Normoxaemia was defined as neither hyperoxaemia nor hypoxaemia (PaO2 <8 kPa).

Comparisons between rounds were made by Chi-squared testing.

Results: In our first audit we retrospectively reviewed data from two years of admissions (2012–2013) (n = 2,916), almost half were in the hyperoxaemia group (48.35%, n = 1,411)¹. A target PaO₂ was added to bedside documentation but this intervention had no significant effect when re-audited (50.46% in hyperoxaemia group, March-June 2016; 1,183 admissions; χ²= 0.038, df = 1, p = 0.8461).

After introduction of the computer-based alert there was a significant reduction of admissions in the hyperoxaemia group (32.84%; n = 151; χ²= 152.971, df = 1, p < 0.0001) and almost two-thirds of admissions were in the normoxaemia group (65.98%; n = 519).

Conclusion: A simple computer-based alert highlighting abnormally high PaO₂ values to nursing staff during routine practice has enabled us to significantly reduce the incidence of sustained exposure to excess oxygen to patients within our general critical care unit.

The one year follow up outcomes of all critical care referrals, and the association between frail and non-frail, accepted and declined referrals: A single-centre long-term prospective study

Megan Kershaw, Alex Steptoe, Jack Parry Jones and James Williams

Aneurin Bevan University Health Board, Newport, UK

Abstract

Keywords: Frailty, Intensive Care, Long Term, Outcomes, Clinical Frailty Scale

Frailty amongst intensive care patients has been shown to have a significant impact on mortality and morbidity. This study evaluates outcomes in all patients referred to intensive care over a two-month period using the Clinical Frailty Score (CFS). Unlike previous studies patients were followed up for a one-year period, regardless of whether they were admitted to intensive care or not.

Primary outcomes were mortality at 30 days, 6 months and 12 months with length of stay in ICU and acute hospital as secondary outcomes. We demonstrated a significant difference in survival between frail (CFS > 4) and non-frail cohorts (63% vs 33% at one year). Frail patients were less likely to be admitted to the ICU (34% vs 58%). Amongst those deemed appropriate for admission to critical care frail patients had a mortality of 48.7% compared with 26.2% for the non-frail. There was also a significant difference in hospital length of stay between frail and non-frail survivors (26 days vs 16 days). There was no significant difference in admission ICNARC and APACHE II scores between the two cohorts. Amongst declined patients those deemed frail and inappropriate for ICU had a one-year mortality of over 80%.

This study supports previous research that suggests frailty is a significant predictor of outcomes in intensive care patients. Unlike previous studies we have demonstrated that frailty would already appear to be linked to escalation decisions.

In the future we are likely to see a rise both in the prevalence of frailty within the population and demand for ICU care. It is vital that we have objective evidence to enable decision making in the best interests of this growing patient group.

Embolic fracture casuing large MCA infarct follwing thrombolysis of pulmonary embolism

Elizabeth Potter¹, Chu-Hao Chiang² and John Dixon²

¹Kings College Hospital, London, UK

²St Helier Hospital, London, UK

Abstract

Introduction: Acute pulmonary embolism (PE) is a common and sometimes fatal consequence of venous thromboembolism. Within the United Kingdom, the incidence of diagnosed PE is 3–4/10,000 though it is likely to be underestimated.[1] Internationally, the incidence is 4–21/10,000.[2] It is associated with a 30 day mortality of 4% and a 1 year mortality of 13%.[3] Within the UK, it is thought to account for 60,000 deaths annually.[4] Management options include thrombolysis, thrombectomy, and anticoagulation with supportive management. The main indications for thrombolysis are haemodynamic instability manifesting as: Severe or worsening right ventricular dysfunction or cardiopulmonary arrest due to PE. Currently catheter directed thrombolysis is still only to be considered in PE with haemodynamic instability, although reduction in mortality remains unclear following both NICE and Cochrane reviews.[5,6] There are no set recommendations for surgical embolectomy, due to limited evidence in the literature.

Case Summary: A 52 year old gentleman with a past medical history of diabetes mellitus type 2, hypertension, oesophagitis, and paranoid personality disorder presented with acute dyspnea and was diagnosed with PE on CTPA. He was prescribed rivaroxaban. Eight weeks later he was readmitted with haemoptysis and a headache. No focal neurological signs were present. He had been non-compliant with medications. A CT Head revealed no abnormalities, but a CTPA demonstrated extension of the PE and right heart strain. Within 24 hours, he desaturated to 60% and was intubated and ventilated after a cardiac arrest. He was admitted to ITU and thrombolysed. His initial echocardiogram demonstrated biventricular failure and, over the subsequent 3 days, cardiac output monitoring suggested improved function. On day 3 post thrombolysis, he did not wake up following a sedation hold and was not moving his right arm. An urgent CT Head revealed substantial thrombus in the Left MCA and extensive infarction in the left hemisphere with mass effect, and a right occipito-parietal infarct. A best interest meeting with IMCA concluded in palliation.

Conclusion: Evidence should guide practice as to management of PE. However, there may be specific patient factors which increase an individual’s risk for interventions and decisions should be made on a case-by-case basis. The risk of haemorrhage and haemorrhagic stroke is a well-known complication following thrombolysis. However, thromboembolic events are rarely reported: there are currently no publications reporting ischaemic stroke secondary to embolic fracture from a PE. The risk of haemorrhage may be attenuated with catheter directed thrombolysis when compared to systematic thrombolysis but this is associated with the additional risk of pulmonary artery rupture. NICE is currently encouraging research into Ultrasound enhanced catheter directed thrombolysis. Surgical thrombectomy is currently reserved for patients in whom thrombolysis is contraindicated or where thrombolysis has failed. Ischaemic strokes may demonstrate no change in an initial CT scan and thus can be falsely reassuring if one were only contemplating haemorrhagic strokes. A high index of suspicion is required when managing these patients, particularly whilst the evidence base is evolving.

Evolution of Induction – The Adult Intensive Care Scavenger Hunt

Reema Patel and Paul James

Queen's Medical Centre, Nottingham, UK

Abstract

BACKGROUND: Traditionally, a new rotation in the hospital begins with a day of induction consisting of power point presentations, talks by invested parties and perhaps even a demonstration of common procedures expected to take place. However, the importance of induction and it’s effect on workforce outcomes is often underestimated.

Research from the corporate world has highlighted benefits of a good induction process to all involved including increasing the employee's feeling of ‘belonging’ and reducing insecurity associated with a new job^1,2

The intensity of Intensive Care Medicine can lead to burnout, emotional fatigue and a lack of job satisfaction therefore attempting to address some of these issues from the outset may lead to an improved morale, increased workforce commitment and retention of the best trainees.

Aims:

1. Discover a novel means by which to introduce trainees to their environment and colleagues they will be collaborating with on a daily basis - the Adult Intensive Care (AICU) Scavenger Hunt.

2. Assess the efficacy of this method at encouraging early team work and evaluating trainees and their colleagues skills set.

Methods:

1. AICU Scavenger Hunt trialled with 15 participants over 2 induction sessions.

2. Participants divided into two groups and asked to locate and name photographs of airway items situated in commonly visited locations within the hospital including the High Dependancy Unit and Accident and Emergency Resuscitation Department.

3. Pre and post hunt questionnaires completed with emphasis of questions on enjoyment of the activity, team work and group skills set.

Results: 12 of 15 participants had worked at QMC prior to induction with 10 stating they had never worked in the AICU before. 7 participants had worked with some of the other doctors in the room prior to induction. 13 participants had never taken part in a Scavenger Hunt before.

On enjoyment and generic usefulness of the Scavenger Hunt:

1. Enjoyment: 4.4/ 5 (average)

2. Usefulness of Scavenger Hunt as an ICE Breaker: 4.2/5 (average)

3. Usefulness of AICU Scavenger Hunt in understanding layout of the hospital: 4/5 (average)

4. Encouraging other departments to include the activity as part of their induction process: 4/5

On the Scavenger Hunt in understanding colleagues and team work:

1. Understanding roles and skills set of other members of the team 4/5 (from 3/5 pre hunt)

2. Feeling comfortable working with other doctors in the room 4/5 (from 4/5 pre hunt)

3. Helping to determine strengths and weakness of the participant 2.9/5

4. Helping to determine strengths and weaknesses of colleagues 2.8/5

The top 3 factors that contribute to effective teamwork (Pre hunt): Communication Understanding skills set of the team Being open to suggestions

The top 3 factors that contribute to effective teamwork (Post hunt): Communication Understanding skills set of the team Listening to other members of the team

Conclusion: We have demonstrated an enjoyable method by which induction can help doctors familiarise themselves with their environment and colleagues. Recommendations from this include evolving this process to include simulated tasks for teams to complete and formulating a questionnaire for participants at the end of their rotation in order to assess longer term effects of the AICU Scavenger Hunt on the team work, morale and job satisfaction.

The accuracy of hand-held echocardiography performed by non-cardiologists with limited echocardiography training

Julia Harrington¹, Beth Whittington¹, David Cusack¹, Joanne Mccarley¹, John Hung² and Alan Japp^1,2

¹St. John's Hospital, Livingston, UK

²Royal Infirmary of Edinburgh, Edinburgh, UK

Abstract

Background: Bedside echocardiography using hand-held portable devices is reported to have favourable image quality and diagnostic accuracy when compared against standard transthoracic echocardiography (TTE). Most studies with hand-held echocardiography (HHE) have employed experienced cardiologists with formal echocardiography accreditation. We sought to explore the accuracy and clinical utility of HHE performed in a real-world clinical setting by non-cardiologists following a defined, limited period of echocardiography training.

Methods: In this prospective cross-sectional study, inpatients referred for TTE from January 2017 to January 2018 underwent bedside HHE with a Vscan (GE Healthcare), performed by one of two middle grade doctors (anaesthetic CT1/medical CT1), prior to departmental TTE. Both assessors received 30 hours of hands-on echocardiography training by a British Society of Echocardiography (BSE) trainer prior to study participation, and neither had previous echocardiography experience. Each study assessed four specific areas, broadly based on the Focused Intensive Care Echo (FICE) minimum dataset, with addition of valvular assessment. Referrals were considered for HHE if within these areas, but excluded if the use of spectral doppler was needed (e.g. for quantification of valve gradients). All patients subsequently underwent formal TTE carried out by a BSE accredited sonographer blinded to the results of the HHE. Accuracy of HHE for detection of i) pericardial effusion, ii) RV dilatation/dysfunction, iii) left ventricular systolic dysfunction (LVSD) of at least moderate severity and iv) valve disease of at least moderate severity was determined using TTE. Alongside each HHE report, the assessor specified whether they would recommend departmental TTE.

Results: Paired HHE and TTE was performed on 42 acute medical and intensive care patients (21 each assessor). Assessment for LV dysfunction and pericardial effusion was consistently accurate with HHE (table 1). Although RV dysfunction and valvular regurgitation were occasionally underestimated, full departmental scan had been recommended in each case. In all cases of significant discrepancy between HHE and TTE, the scanner had recommended departmental TTE. No cases of severe LVSD, severe valve disease or pericardial effusion were missed by HHE. In the 13 patients where the assessor would not have recommended departmental TTE, there were no significant discrepancies between HHE and TTE.

Table 1.

Diagnostic Accuracy of HHE.

Measure	Pericardial effusion	Significant Valvular Disease	Left Ventricular Function		Right Ventricle
			<Moderate	≥Moderate	Dilatation	Systolic Impairment
Sensitivity (%)	100	77	100	93	80	69
Specificity (%)	100	89	93	100	93	100
PPV (%)	100	77	96	100	80	100
NPV (%)	100	89	100	96	93	86
(n)	1	16	26	16	10	13

^*(n) = Numbers of times abnormality detected on TTE.

Conclusion: HHE has comparable sensitivity and specificity for specified indications in real-world clinical practice, even when employed by non-cardiologists with only a limited period of echo training. If replicated in larger studies, these findings would suggest that focused training of physicians in HHE in the acute medical or intensive care setting could help to guide real-time clinical decision-making, reduce demand for departmental TTEs and complement the FICE accreditation process.

Evaluating the neuroprotective effect of melatonin on patients with hemorrhagic stroke using serum S100B protein as a prognostic marker

Mehrnoush Dianatkhah¹, Mojtaba Mojtahedzadeh¹, Atabak Najafi², Hamidreza Sharifnia² and Farhad Najmeddin¹

¹Department of Clinical Pharmacy, Tehran University of Medical Sciences, Tehran, Iran, Tehran, Islamic Republic Of Iran

²Department of Anesthesiology and Critical Care, Sina Hospital, Tehran University of Medical Sciences, Tehran, Iran, Tehran, Islamic Republic Of Iran

Abstract

Background: Intracerebral hemorrhage (ICH) is the second most common cause of stroke following ischemic stroke. Current data show that early initiation of neuroprotective treatment might protect threatened neurons and improve outcome. The aim of this study is evaluating the neuroprotective effect of melatonin on the hemorrhagic stroke.

Method: Forty Adult patients with nontraumatic ICH, who was admitted to the ICU within 24 hours of onset were enrolled in this study. Subjects in the melatonin group received 30 mg of melatonin every night for 5 days. In order to measure the severity of neuronal injury, S100B was assessed once on day 1 and, day 5 post ICU admission. Additionally, the length of ICU stay, mortality and the duration of mechanical ventilation were also recorded.

Result: In both groups the plasma levels of S100B decreased after 5 days in comparison with their baseline values. However the decrement was more significant for the melatonin compare to the control group. The duration of mechanical ventilation and length of ICU stay was shorter in patients who received melatonin, and this difference was statistically significant for the length of ICU stay, and marginally significant for the duration of mechanical ventilation. Although not statistically significant, the mortality rate of the control group was 30%, almost double that of the study group (15%).

Conclusion: The result of our research has shown that melatonin can increase the clearance of S100B in ICH patients, and it can be hypothesized that melatonin may be a safe and effective nueroprotective agent with some exclusive properties which has made it an appropriate medication as an adjuvant therapy in the critically ill intubated patients.

IRCT: IRCT2016100530164N1.

Analysis of A-line Patterns Seen on Lung Ultrasound Scans in Healthy Volunteers Following Spontaneous Breathing and High Flow Nasal Cannula Therapy

Jing Yi Kwan^1,2 and Wojciech Wierzejski²

¹University of Manchester, Manchester, UK

²Lancashire Teaching Hospital NHS Trust, Preston, UK

Abstract

A-lines are a type of ultrasonographic artefact seen as horizontal lines arising at regular intervals from the pleural line. The presence of A-lines can either be a variant of normality or pathological conditions like pneumothorax.¹ Currently, there is a lack of guidelines and recommendation about the usage of A-lines analysis in the diagnosis and monitoring of respiratory conditions.² This study is designed to determine whether hyperinflation of the lungs results in a difference in the number and echogenicity of A-line artefacts on lung ultrasound scans.

We performed a prospective before-and-after trial on 37 healthy volunteers. Lung ultrasound scans were performed before and right after 15 minutes of high-flow nasal cannula (HFNC) air therapy, which was used to increase the end-expiratory volume of the lungs to create a state of hyperinflation. Two variables were analysed: the number and echogenicity of A-line artefacts.

Changes in the number of A-line artefacts before and after HFNC were analysed using a paired t-test. Out of 37 healthy participants, 28 showed a significant increase in the number of A-lines [1.27, 95% CI 0.82 to 1.72, P < 0.0001] after 15 minutes of HFNC. Conversely, a majority 51.4% of participants (19 participants) showed no changes in echogenicity. However, a two-sided P value of P = 0.0127 (P < 0.05) obtained using a sign test indicates that if there is any change, it is more likely to be an increase in echogenicity.

Since hyperinflation of the lungs results in a significant increase in the number of A-lines, lung ultrasonography has the potential to be used as a point of care tool for the monitoring of the degree of inflation of the lungs, and thus the severity of Chronic Obstructive Pulmonary Disease (COPD) and asthma attacks. It can also be used in the monitoring of mechanical ventilation and prevention of ventilator-induced lung injury (VILI).

Investigating the accuracy and effect of post-surgical antimicrobial administration practices on an Intensive Care Unit

Rhys Wenlock¹, Caroline Gerrard² and Alain Vuylsteke²

¹School of Clinical Medicine, University of Cambridge, Cambridge, UK

²Royal Papworth Hospital, Papworth Everard, UK

Abstract

Background: Antimicrobial prophylaxis is routine for Cardiac Surgery in the UK. This study set out to explore the accuracy to which antimicrobial doses are administered on an Intensive Care Unit (ICU) and, if present, the impact of delayed administration on patient outcomes.

Methodology: We extracted information from the electronic record of consecutive post-surgical patients admitted onto a single cardiothoracic ICU between 2011 and 2016. Demographics, predicted surgical risk, time, quantity and name of each antibiotic dose administered and the presence of Surgical Site Infection (SSI) at 30 days post-operatively were analysed. The 4 most commonly prescribed antimicrobials in this setting (Flucloxacillin, Ciprofloxacin, Vancomycin and Tazocin), were included in the analysis. The intervals been antimicrobial doses were calculated for the first three prophylactic doses on ICU and the difference between “actual” and “expected” dose interval was calculated (Figure 1). The mean delay in antimicrobial dosing was analysed, using the Wilcoxon signed-rank test, between SSI-Positive and SSI-negative groups, in addition to sub-group analysis with groups stratified by surgical risk. This analysis reported a power of 80% to detect an effect size of 30%.

Results: Of 10,764 patients, 3647 received at least 3 prophylactic antimicrobial doses after arrival in ICU. Flucloxacillin represented the largest sample with 2441 patients. 36.3% of antimicrobial doses were delayed with a mean delay by patient of 29 minutes, or by dose of 15 minutes.

Seventy-nine of the 2441 patients receiving prophylactic Flucloxacillin had a SSI at 30 days post-operatively. There was no significant difference in the delay in antimicrobial administration between any of the SSI-positive or SSI-negative groups/subgroups.

Discussion: Across a large dataset, antimicrobials are consistently administered late with a mean delay of 15 minutes per dose. Our analysis, shows that it is unlikely that such practice has a large effect on the rate of post-operative surgical site infections.

Role of fragmented QRS complex in the prediction of the extent of myocardial damage and prognosis following acute coronary syndrome

Nora Abbas, Ahmed Salah Younes, Moataz ElHalag and Mahmoud ElBadry

Critical Care department Cairo University, Cairo, Egypt

Abstract

Background: Coronary heart disease (CHD) is one of the major and leading causes of death worldwide.¹ Fragmented QRS (f-QRS) is a pattern of QRS complex in 12 leads surface ECG which has a great diagnostic and prognostic role in cardiac diseases including coronary heart disease.^2,3

Objective: to investigate the role of using f-QRS in acute coronary syndrome (ACS) as a non-invasive and easily accessible tool for the prediction of myocardial damage.

Patients & Methods: Retrospective study of 84 patients with ACS were divided into 46 patients with f-QRS (fragmented group) and 38 patients without f-QRS (non-fragmented group) excluding prior history of major ischemic events (MI, PCI & CABG), permanent AF or ischemic and non-ischemic cardiomyopathy.

Results: General demographic characteristics, major risk factors of CHD, Killip class, updated GRACE risk score of ACS, cardiac biomarkers, wall motion score index (WMSI), left ventricle ejection fraction (LVEF), diastolic dysfunction (DD), mitral regurgitation (MR), Gensini score and in-hospital death showed no significant differences between both groups. Only LVEDD was significantly higher in fragmented group than non-fragmented group (P = 0.033). The optimum cut off value for f-QRS leads was >3 leads for predicting in-hospital death with 83.3% sensitivity and 72.5% specificity. In the fragmented group, patients were divided into 2 subgroups according to the numbers of f-QRS leads; Subgroup (A1) including patients with >3 f-QRS leads &subgroup (A2) including patients ≤3 f-QRS leads. Subgroup (A1) showed significant difference than subgroup (A2); a lower SBP (111.33 ± 25.03 vs. 139 ± 38.89, P = 0.016), a higher HR (93.81 ± 19.13 vs. 80.77 ± 14.91, P = 0.014), a higher updated GRACE risk score (6.81 ± 12 vs. 3.22 ± 6.95, P = 0.048),a lower LVEF (48.08 ± 13.07 vs. 56.14 ± 10.92, P = 0.049), a higher WMSI (1.55 ± 0.33 vs. 1.27 ± 0.27, P = 0.007), a higher Gensini score (86.12 ± 47.2 vs. 55.08 ± 35.97, P = 0.030) and a higher incidence of in-hospital death (5/16 vs. 1/30, P = 0.015). The different locations of f-QRS had different impacts on SBP, HR, Killip (IV), LVEF, WMSI, updated GRACE score, Gensini score and in hospital death. Anterior f-QRS showed significant differences than non-anterior f-QRS; with a lower SBP (P = 0.006), a higher HR (P = 0.040), a higher incidence of Killip (IV) (P = 0.030), a lower LVEF (P = 0.039), a higher WMSI (P = 0.004), a higher updated GRACE risk score (P = 0.033), a higher Gensini score (P = 0.016) and a higher incidence of in-hospital mortality (P = 0.004).

Conclusion: Fragmented QRS on 12 leads surface ECG is not an uncommon phenomenon among the patients with acute coronary syndrome (ACS). The location and the number of f-QRS can be used as a cheap non-invasive and easily accessible tool to predict the extent of myocardial damage.

Electrical Impedance Tomography: Clinical Applications for the Paediatric Respiratory Physiotherapist

Simon Gates, Catarina Silvestre and Patrick Davies

Nottingham Children's Hospital, Nottingham, UK

Abstract

Aims and Objectives: Electrical impedance tomography (EIT) is an established method of measuring global lung expansion, regional ventilation and ventilatory compliance. It has been utilised frequently in the research setting but has had relatively little impact in the clinical setting to date.

Respiratory physiotherapy in paediatric intensive care aims to optimise respiratory function by promoting airway clearance and lung recruitment. As such, EIT may have a role in the physiotherapy management of patients in paediatric intensive care but little information is available.

Methods: Cases where EIT was in place during physiotherapy intervention in the paediatric intensive care unit (PICU) were reviewed. Information collected included real time recordings of EIT data, screen shots and opinions of the physiotherapy team on their knowledge of EIT and whether it influenced their intervention.

Results: The utility of EIT for physiotherapists in this period could be divided into 3 primary categories

1. Guidance in assessment

2. Monitoring of treatment effect

3. Providing real time feedback to incentivise a patient’s inspiratory effort

Key cases were selected as examples for practice and illustrated with real time recordings.

Feedback from the physiotherapy team highlighted that interpreting basic tidal images was a relatively easily developed skill but more in depth analysis of breath to breath monitoring required significantly more time and educational input.

Conclusion: Real time evaluation of lung expansion is a potentially valuable addition to the respiratory physiotherapy assessment of children in PICU. EIT monitoring can influence and assist respiratory physiotherapy management in PICU.

Currently, there is little guidance available to physiotherapists to inform how EIT may be useful in practice. The information in this review of clinical cases provides a basis for the use of EIT by physiotherapists and highlights potential areas for the focus of future research. It also highlighted the importance of a robust education and competency programme to ensure EIT is utilised to its maximum potential in the physiotherapy setting. This is currently in development.

Percutaneous Tracheostomy Practice in a U.K. Major Trauma Centre: A Report on 519 Procedures

Adam Green¹, Paul Bramley², Matt Thomas³, Jules Brown³ and Alex Manara³

¹Clinical Fellow, Intensive Care Unit, Southmead Hospital, Bristol, UK

²Academic Foundation Year 2 Doctor, Intensive Care Unit, Southmead Hospital, Bristol, UK

³Consultant, Intensive Care Unit, Southmead Hospital, Bristol, UK

Abstract

Southmead Hospital has a 46 bed neurotrauma Intensive Care Unit (ICU). Percutaneous tracheostomy was introduced to our ICU in 1991 and has become one of most commonly performed invasive procedures undertaken. Recent practice recommendations from NCEPOD have aimed to improve the safety surrounding tracheostomy care _[1]. There is ongoing debate over the optimum timing for tracheostomy placement following the TRACMAN study _[2], especially in patients with neurological injury where reduced mortality has been demonstrated with an earlier approach to tracheostomy (<10 days) _[3]. We report our recent experience of more than 500 patients.

Records were obtained for all percutaneous tracheostomies undertaken on the ICU between May 2014 and September 2017. Neurological illness was considered as a subgroup. Patients that had had a surgical tracheostomy, or those admitted with a tracheostomy in situ were excluded from analysis.

519 patients (of 2069 ventilated patients) underwent percutaneous tracheostomy. Patients with neurological illness were over represented in the tracheostomy cohort, comprising 53.3% of all tracheostomies but only 34.4% of total admissions. For all diagnostic groups combined the median (interquartile range) APACHE II score was 15 (10–21) with a 6 month mortality rate of 16.8%. The median duration to tracheostomy and decannulation was 8 (5–11) and 14 (10–21) days respectively. The average length of stay on the ICU was 23 (17–31) days. The combined major and minor complication rate was 5.3%, inclusive of one catastrophic haemorrhage on decannulation leading to the death of a patient. When compared to all other diagnostic groups the neurological illness subset was a younger cohort with a median age of 51 (37–65) vs 64 (52–75) and similar proportions of males and females in both groups. This group also had a lower APACHE II score of 12 (8–19) vs 18 (13–23) and a lower 6 month mortality of 10.8% vs 22.1%. ICU length of stay was lower at 21 (15–28) days as opposed to 25 (20–34) days for all other diagnostic groups. However, the median duration to tracheostomy was the same at 8 days for both subsets. Likewise, the duration to decannulation was similar at 14 (9–23) days for the neurological illness group vs 15 (10–21) days for all other groups. Weaning from mechanical ventilation was the most common indication for tracheostomy insertion in both groups at 48.9% for the neurological illness subset and 77.4% for all other combined diagnostic groups, with airway protection being the next most common indication at 36.9% and 14.5% respectively.

This retrospective study demonstrates our local percutaneous tracheostomy practice over a 3 year period. The differences between the neurological and non-neurological group suggest that these are indeed distinct patient populations. With this in mind it is interesting that the median time to tracheostomy does not differ. This large case series gives an indication of contemporary U.K. tracheostomy practice and indicates a ‘real world’ incidence of associated complications outside of clinical trials. The use of retrospective data, from a single institution alongside the potential for missing data are recognised limitations.

LP QIP: Quality improvement for precious samples on ICU

Scott Bradburn

University Hospital of Wales, Cardiff, UK

Abstract

Introduction: Cerebrospinal fluid samples are often sent for diagnostic purposes on intensive care. In our institution there is a variation as to the tests requested, how they are sent and the process as to how they get to the biochemistry lab. This results in inappropriate tests being requested, lost samples and wasted time obtaining the desired equipment for sampling.

Method: A cohort of 12 doctors working on ICU were surveyed with regard to their practice when requesting tests on CSF samples, how many bottles they would send, how much CSF in each bottle, have they had a sample go missing or had a delay in its processing, do they find all equipment easy to find, do they think an LP pack similar to CVC pack would be useful?

Results: The response to the survey showed a huge variability of tests requested by individual doctors.

The number of bottles sent ranged from 2–6

The number of drops of CSF ranged from 5 drops to >1 ml CSF.

Half of all respondents had experienced a sample of CSF going missing.

The majority of respondents found LP equipment difficult to find.

All repondents thought a lumbar puncture pack would be a useful addition the the equipment kept in ICU.

Discussion: As a result of the survey and discussion with lab staff a standardised lumbar puncture pack is now available for use, it includes a 'quick help guide' demonstrating the tests required on a simple diagnostic CSF sample and a 'precious sample label. It is hoped with these simple interventions will improve ease of access to LP equipment decrease sampling variation, avoid unncessary tests and loss of samples.

Severe respiratory failure rescue: APRV survey, protocol and algorithm

Patrick Colhoun, Scott Bradburn

Cardiff and Vale University Health Board, Cardiff, UK

Abstract

Introduction: Airway Pressure release ventilation is increasingly being used as a rescue technique in patients with severe respiratory failure. It may be used in patients who are unsuitable for proning or paralysis. Despite being a valuable tool in managing severe respiratory failure it is not a ventilation mode many trainees and consultants are familiar with.

Method: A survey was obtained of 32 doctors working on intensive care units within Wales with regards to their experience using APRV. We asked what the respondents stage of training was, how often they have used APRV in the last 12 months. Whether they knew where to find guidance on APRV settings. Whether they were aware of indications and contraindicatiosn for APRV. Whether they found APRV easy to use and whether they could manage a patient becoming increasingly hypoxic or hypercarbic on APRV.

In addition to these questions we asked whether a APRV guidance protocol would be useful.

Results: We received 32 responses from a range of training grades and consultants (Core trainee 3), (Intermediate 7), (Higher 16), (Consultants 6). Of the individual who replied 84% Were unfamilar with APRV and did not know where to find information on initial settings.

In addition 65% of respondents were unaware of indications and contraindications for APRV. 88% Did not find APRV easy to use. 72% Did not know how to manage a patient becoming increasingly hypoxic or hypercarbic on APRV.

75% strongly advocated the development of an APRV protocol.

Discussion: As a result of the survey a standard operating procedure has been developed for use on the intensive care unit.

In addition a quick reference algorhythm has been developed. We show case these tools that will allow APRV to be used as a technique in severe respiratory failure confidently by both trainees and consultants.

A review of abdominal massage in the treatment of gastrointestinal motility disorders

Dr Mark Verlander¹, Dr Stephanie Jones² and Shondipon Laha¹

¹Critical Care Unit, Lancashire Teaching Hospitals NHS Foundation Trust, Preston, UK

²University of Central Lancashire, Preston, UK

Abstract

Background: Abdominal massage is a recognised therapy in several disorders, however it is not currently utilised in the intensive care unit (ICU). ICU Patients are much less active for a variety of reasons. These patients have a high risk of suffering from gastrointestinal motility disorders which can affect mortality and length of stay.

When these disorders arise, clinicians prescribe prokinetics in order to alleviate these problems. However, prokinetics are not without risk, because they have many side effects, a risk of infection when administered and a risk of increasing antibiotic resistance. The aim of this review is to summarise the scientific literature on the use of abdominal massage in hospital and community settings with a view to utilising this intervention in the ICU.

Methods: We performed a review in accordance with PRISMA guidance. MEDLINE and EMBASE were searched, from 2000 to 2017. Any intervention studies were included if they focused on the efficacy and safety of abdominal massage as an intervention in hospitals and community settings. Studies were included if they are published in full and in English. Studies that appeared to meet the inclusion criteria had an in-depth evaluation. Data were extracted from the articles, and risk of bias using a pre-specified tool based on Cochrane guidance and research quality were assessed by two researchers.

Results: Our search identified 268 studies of which 5 met the criteria for inclusion. All 5 were randomised controlled trials. The number of participants involved totalled 419 across 5 studies, median 84 (range 32 to 189). Two studies took place in the community and 3 in hospital settings (neuroscience units, adult ICU). The mean proportion of female patients was 68%, the mean age was 60.9 years (range 23–90).

The most commonly used outcome measure was Gastric Residual Volume (GRV) which was used in 3 studies (3, 4, 5); GRV was used in combinations with vomiting, abdominal distention and abdominal circumference. Studies included a range of outcome measures including constipation symptoms (Constipation Scoring System (CSS), Gastrointestinal Symptoms Rating Scale (GSRS) (2), Bowel Symptoms (7-day bowel diary), Radio Opaque marker transit tests and adherence to massage schedule (massage diary). Only one study (1) measured changes in Neurogenic Bowel Dysfunction (NBD) Scores at baseline and 24 weeks, reporting a non-significant improvement (p < 0.056).

The effect of abdominal massage on reducing the gastric residual volume and abdominal distension was significant (p < 0.05) across all studies using this outcome measure. Studies also reported reduced rates of vomiting and pneumonia, and significant increases in the frequency of bowel movements.

Conclusion: This is the first review that has systematically synthesised the published literature on use of abdominal massage in a range of community and clinical settings. Whilst there is a paucity of research in this area, the findings from five randomised controlled trials suggest that abdominal massage can be effective in improving gastrointestinal motility either by a reduction in gastric residual volume, abdominal distension and a reduction of constipation related symptoms. Further trials are required to establish a clear benefit in ICUs.

A retrospective evaluation of the prevalence, characteristics and management of refractory ventricular fibrillation (rVF) amongst in-hospital cardiac arrests in a District General Hospital

Luke Staveley-Wadham and Tom Standley

Western Sussex Hospitals NHS FT, Worthing, UK

Abstract

Background: Refractory ventricular fibrillation (rVF) is widely accepted as VF persisting despite the proper application of Advanced Life Support (ALS).^[1] Recently rVF has attracted significant discussion, including the proposal of novel management strategies, such as Dual Sequential Defibrillation.^[2] Much of this interest stems from pre-hospital research, with a paucity from in-hospital management, potentially leading to lack of awareness in the hospital domain. Given the availability of in-hospital investigations to guide the correction of reversible causes of rVF and emerging advanced resuscitation strategies, the rationale for this study is to review the prevalence, characteristics and management of rVF in a DGH setting to help inform clinical practice.

Method: In a DGH with in-hours PCI capability cardiac arrest proformas (from 05/01/2016 to 28/01/2017) were individually reviewed (n = 200). Cases with 3 or more consecutive pulse checks with VF/VT were recorded. VF unresponsive to more than 5 defibrillations were categorised as true rVF, with 3 or more impending rVF. Patient notes and electronically stored investigations were reviewed, with pertinent information recorded in spreadsheet format.

Results: Of 200 cardiac arrests reviewed, 35 where shockable-rhythms (VF/VT), with the remainder non-shockable rhythms (PEA/Asystole). 12 cases represented impending rVF, with 4% of total cardiac arrests requiring 3–5 defibrillations (n = 8), and 2% requiring 5 defibrillations (n = 4), all of which reverted on the 5^th defibrillation. One patient (0.5% total arrests) was a true rVF, requiring over 5 defibrillations (16 in total).

Of the impending and true rVF arrests, 46.2% (n = 6) had unclear causes, 15.4% (n = 2) were secondary to cardiac thrombus (1 NSTEMI and 1 STEMI). 15.4% (n = 2) were secondary to hypokalaemia (average Potassium 2.6 mmol), 7.7% (n = 1) secondary to hyperkalaemia (Potassium 6.9 mmol), 7.7% (n = 1) secondary to dysrhythmia and 7.7% (n = 1) secondary to multi-organ failure.

In relation to prognosis, 53.8% (n = 7) achieved return of spontaneous circulation, 23.1% (n = 3) survived to hospital discharge and 15.4% (n = 2) were transferred. Of the patients that survived to discharge the average age was 69, with the causes of arrest being hypokalaemia (n = 2) and cardiac thrombus (STEMI) (n = 1). Although not amendable to statistical analysis, it would appear the greater number of defibrillations correlated to poorer prognosis.

Management was compared with ALS protocols, with some inconsistencies noted, mainly involving administration of Adrenaline and Amiodarone and defibrillation usage. However, interventions may not have been accurately recorded, representing a limitation of this study. One patient received thrombolysis (STEMI), with no other advanced strategies employed (potentially secondary to limited DGH facilities). No novel management strategies were recorded.

Conclusion: A small but significant proportion of cardiac arrests represented impending and true rVF, with some cases surviving to hospital discharge. A cohort of these patients had simple reversible causes for their rVF amenable to essential resuscitation strategies. This study highlights that rVF is prevalent in the hospital domain and importantly the early application of essential in-hospital resuscitation strategies can be improved on when compared with national guidelines. A larger scale multicentered trial would be useful to further evaluate the characteristics of in-hospital rVF and review novel management strategies as outlined in the pre-hospital literature.

A service review of the incidence of laryngeal and tracheal abnormalities in mechanically ventilated spinal cord injured patients and the impact this has on team decisions to commence oral intake and begin weaning

Hannah Chalke and Sarah Morgan

The Royal National Orthopaedic Hospital, Stanmore, UK

Abstract

Introduction: Spinal cord injury (SCI) is a complex condition affecting multiple body systems. Paralysis of respiratory muscles in cervical spinal cord injury (CSCI) can lead to rapid respiratory failure, with high numbers requiring mechanical ventilation (MV). Early insertion of a tracheostomy is recommended as it allows for reduction in intubation time, sedation and facilitates communication and swallowing.¹ Long term MV is associated with increased risk of laryngeal and tracheal injury, with reports being as high as 80%. Without use of endoscopy decisions to wean or decannulate may be unsuccessful.²

Dysphagia in CSCI is common with a reported incidence of 40%. Clinical presentation is subtle with high rates of silent aspiration. Left untreated this could impact on weaning due to respiratory complications associated with aspiration.³ The London Spinal Cord Injury Centre (LSCIC) is one of 11 specialist units in the UK for SCI. In our unit there is a weekly mutli-professional tracheostomy team ward round. We routinely use Fibreoptic Nasendoscopy (FNE) to assess the upper airway and swallowing. Prior to decannulation tracheoscopy assessments are carried out.

Abnormality	Percentage
Vocal Fold (VF) impairment	60%
VF tension	5%
VF trauma	5%
Laryngeal Oedema	20%
Subglottic Stenosis	5%
Subglottic Granuloma	10%
Tracheal Granuloma	20%

Method: A retrospective notes review of mechanically ventilated SCI patients admitted to the LSCIC between 2015 and 2018 was completed. 27 patients underwent a FNE to assess the upper airway and presence of dysphagia. 18 male and 9 female patients aged 4 to 79 were included. 23 patients had a CSCI and 4 a thoracic SCI, injuries were complete and incomplete. Outcomes were collected for presence of laryngeal abnormality, dysphagia and team decisions for weaning.

Results: Of the 27 patients, 74% (n = 20) presented with laryngeal abnormalities and 59% (n = 16) with a dysphagia.

Of the patients with laryngeal abnormalities, 15 presented with a pharyngeal dysphagia, suggesting strong association between laryngeal pathology and dysphagia. At the time of the FNE 96% of patients were nil by mouth. Following FNE 74% were able to start some form of oral intake. 70% of patients with a laryngeal abnormality required further intervention prior to considering weaning and/or decannulation e.g. downsizing tracheostomy tubes or referral to ENT.

Conclusion: Although small numbers this study highlights the high number of laryngeal abnormalities and dysphagia in this complex group. The use of FNE in the context of the MDT allows a proactive approach to the management of these difficult to wean patients, and may facilitate earlier oral intake.

Haemodynamic monitoring and focused echocardiography on the Neurosciences Intensive Care Unit (NICU): A service evaluation

Gary Misselbrook, Eleanor Green and Manprit Waraich

University Hospitals Southampton NHSFT, Southampton, UK

Abstract

Background: The major aim of neurocritical care is the prevention of secondary brain injury after major neurological events. To achieve adequate cerebral perfusion pressures in patients with traumatic brain injury (TBI) and treat cerebral vasospasm in patients with aneurysmal subarachnoid haemorrhage (aSAH), vasopressors are commonly instituted on neurocritical care unit but have the potential to induce myocardial ischaemia, particularly in older patients. Haemodynamic monitoring is therefore indicated to optimise volume status and guide the use of haemodynamic support to maintain sufficient cerebral blood flow and oxygen delivery to tissues. Echocardiography (TTE) is a method of haemodynamic monitoring which has been shown to change the management of patients on ICU in approximately 37–51% of cases.¹ Focused Intensive Care Echocardiography (FICE) is an expanding skill in general intensive care but there is a paucity of information regarding its use in neurointensive care units.

Table 1.

Indications for echocardiogram.

Reason for TTE (May be >1 reason on request)	Number	Could be investigated by FICE Scan
Ischaemic ECG Changes	7	Y
Hypotension/Escalating  vasopressors	5	Y
New murmur	5	N
Pulmonary oedema	4	Y
Post arrest	2	Y
PE	1	Y
Pericardial effusion	1	Y
Heart failure	1	Y
Cause of collapse	1	N
Endocarditis	1	N

Aims: Audit compliance with local guidelines on the institution of haemodynamic monitoring in our unit and quantify the frequency and potential indications for echocardiography performed on neurocritical care locally.

Methods: Patients with aSAH and TBI identified from database maintained by neurovascular and head injury specialist nurses. All patients with aSAH or TBI admitted to Neurosciences Intensive Care Unit (NICU) between 01/01/17–30/06/17 were included in the evaluation. Information was collected using Metavision CIS, paper notes and electronic laboratory reporting systems. Data was then compared to local standards of practice from the Wessex NICU guidelines.

Results: 108 patients were admitted to NICU in 6 months with SAH (n = 56) or TBI (n = 52) with TBI. 25 patients (23.1%) had an TTE performed on NICU (all by cardiac technicians) with a median time from request to time of TTE as 20 hrs. The clinical question submitted for 18 of 25 (72%) TTEs could have been answered by a FICE Scan (Table 1). Of 20 patients with ischaemic changes on ECG, 40% received a TTE and 55% had a troponin I measured. 15 of 29 (52%) patients with noradrenaline infusion at >10mls/hr (>0.16 mcg/kg/min in 70 kg adult) had cardiac output monitoring (LiDCO) instituted.

Conclusion: Local compliance with guidelines was poor highlighting a need for increased education and training to recognise indications for echocardiography and haemodynamic monitoring. There was a median delay of 20 hrs between request and TTE results, by which time cardiac physiology may have changed. If TTE had been requested according to local standards, workload for our cardiac technicians would have been 1.5 times higher. Considering that 72% of clinical questions could have been answered by a FICE scan, training in point-of-care echocardiography is a useful skill to acquire for the neurointensive care unit.

Development of Critical Care Airway Checklist Booklets for use in the ICU and A+E

Kate Hames, David Garry and Ian Edmond

John Radcliffe, Headington, UK

Abstract

After NAP 4 there has been more awareness that Intubation and Extubation outside of theatre is a high risk procedure with the potential for life threatening consequences. During the audit period, 1 in 4 major airway events took place in either the ICU or A + E with lack of planning identified as a major risk factor.

Across the country there have been many changes due to the NAP 4 findings, for example capnography equipment obtained for resus bays and recovery beds. Many different checklists have been developed in the last 10 years to try to reduce human error and prompt better planning prior to an event.

We took the checklist a step further by developing an Airway Booklet that stays with the patient throughout their intubation journey and contains a set of checklists for:

- intubation (including a MACHCHA score chart)

- extubation (including a re-intubation checklist)

- the care of the patient whilst intubated

The booklet features an airway 'flag' on the front to focus all team members on whether the patient has an easy, potentially difficult or known difficult airway. A documentation section splits the intubation into categories: ease of Bag Mask Ventilation, Laryngoscopy score and video laryngoscopy score (Freemantle score). There is an extubation checklist purposefully placed opposite the page where the documentation for intubation has taken place. Towards the end of the booklet are emergency contact details for both our trust sites for the ITU SpR, Max Faxs, ENT and the theatre Anaesthetist. We have also included difficult airway pathways and pictoral prompt for front of neck access as we recognised that in an emergency staff like to be able to see and read through the steps as a team.

The booklet has been running for over 5 months now and we announced their implementation with posters in all the departments explaining key points, email reminders, use of the booklet in simulation scenarios and talking to staff in the department. The team have been more than happy to receive feedback and will develop the booklet to get the highest compliance and efficacy.

A comparison of the use of lung ultrasound and chest radiography in the diagnosis of rib fractures: a systematic review

Ceri Battle^1,2, Simon Hayward³, Sabine Eggert⁴ and Phillip Evans²

¹Physiotherapy Dept, Morriston Hospital, Swansea, UK

²Welsh Institute of Emergency Medicine Research, Morriston Hospital, Swansea, UK

³Physiotherapy Dept, Blackpool Teaching Hospitals NHS Foundation Trust, Blackpool, UK

⁴Ed Major Critical Care Unit. Morriston Hospital, Swansea, UK

Abstract

Introduction: Rib fractures represent the most frequent blunt chest trauma and can be associated with complications such as pneumothorax or lung contusion. Delayed on-set of pulmonary complications following blunt chest wall trauma is common and as a result, prediction of outcome and decisions regarding appropriate management on presentation, can be difficult. Despite its questionable sensitivity, the chest radiograph (CXR) remains the primary diagnostic modality for blunt chest wall trauma. More recently, the use of lung ultrasound (LUS) has been investigated for diagnostic accuracy in the identification of rib fractures following blunt chest wall trauma. The aim of this systematic review is to investigate whether the use of LUS is superior in accuracy to chest radiography, in the diagnosis of rib fractures following blunt chest wall trauma.

Methods: This study was registered with PROSPERO database (CRD:42017067908) and conducted according to the PRISMA guidelines. Electronic databases including MEDLINE, EMBASE, Cochrane and ScienceDirect were searched with no imposed time or language limitations. Reference lists of eligible studies and review articles were hand-searched. The search terms are illustrated in Table 1.

Population: adults aged 16 years or more, patients with blunt chest wall trauma. Interventions: LUS. Comparators: CXR. Study design: Observational studies. Studies not investigating diagnostic accuracy of LUS compared with CXR were excluded from the review. Two authors independently assessed the risk of bias and quality of the included studies using the QUADAS-2 Tool. Due to the high degree of heterogeneity, it was not possible to conduct a meta-analysis.

Results: A total of 4317 citations were identified. Of the 25 full articles that were retrieved and critically appraised, a total of 13 were included in the review (Table 2).

Only one study reported sensitivity and specificity values for LUS, CXR and clinical acumen, with LUS having a higher sensitivity, but lower specificity than both CXR and clinical acumen. All other studies reported proportions of rib fractures identified on diagnostic tests. LUS was reported to identify a higher proportion of patients with rib fractures than CXR in 11 of the included studies. The actual number of rib fractures identified by LUS compared to CXR was also reported to be higher in four of the studies. In contrast, two studies reported that a certain number of fractures were evident on CXR, but missed by LUS.

Discussion: The results of this review demonstrate that LUS may be superior to CXR, but the low quality of the studies means that no definitive statement can be made. High levels of heterogeneity between the studies precluded meta-analysis and the drawing of any definitive conclusions. As computed tomography (CT) is generally agreed to be the gold standard test in the diagnosis of rib fractures, studies directly comparing LUS with CT scan should be completed. It is not currently possible to recommend an immediate change in clinical practice based on the results of this review, due to the low quality of the studies, the lack of a reference standard and the variability in methods used within each study.

A multi-centre randomised feasibility STUdy evaluating the impact of a prognostic model for management of BLunt chest wall trauma patients: STUMBL Trial

Ceri Battle¹, Hayley Hutchings², Timothy Driscoll², Claire O'Neill², Sam Groves³, Zoe Abbott², Fiona Lecky⁴, Phillip Evans¹, Richard Body⁵, James Gagg⁶ and Sally Jones⁷

¹Welsh Centre for Emergency Medicine Research, Morriston Hospital, Swansea, UK

²Swansea University Medical School, Swansea, UK

³College of Human and Health Sciences, Swansea, UK

⁴School of Health and Related Research, Sheffield, UK

⁵Emergency Dept. Manchester Royal Infirmary, Manchester, UK

⁶Emergency Dept. Musgrove Park Hospital, Taunton, UK

⁷Emergency Dept. Royal Gwent Hospital, Newport, UK

Abstract

Objective: The purpose of this trial was to assess the feasibility and acceptability of a definitive impact trial investigating the clinical and cost effectiveness of a new prognostic model for the management of blunt chest wall trauma patients in the Emergency Department.

Design: A multi-centre, cluster-randomised (stepped wedge) feasibility trial.

Setting: Four Emergency Departments in England and Wales

Participants: A total of 176 blunt chest wall trauma patients

Intervention: A new prognostic model to guide the clinicians' clinical decision making

Main outcome measures: Patient outcomes: mortality, pulmonary complications, delayed upgrade in care, unplanned representations to the ED, physical and mental components of quality of life. Organisational outcomes: patient recruitment rate, survey response rate, clinicians’ compliance with use of model. Qualitative outcomes: feedback regarding training and use of the risk score in clinical practice, feedback regarding trial methods. Health economic outcomes: training and travelling time, equipment costs

Results: Recruitment targets were achieved at all sites. The intervention group was significantly older, more severely injured and scored higher on the risk score, than the control group. All progression criteria were achieved with either green (continue to full trial) or amber (minor changes required) status. After adjusting for pre-defined covariates and sensitivity analysis, there were no statistically significant differences in outcomes between the control and intervention period. Qualitative analysis highlighted that STUMBL was a well-received study and clinicians were enthused about the prognostic model and would support a definitive trial to ensure its wider implementation. The process of collecting trial data on outcome and resource use was demonstrated to be feasible and, as a result, the most appropriate framework for Health Economic analysis in the full definitive trial has been established.

Discussion: All our progression criteria were met with amber or green status, which has demonstrated that the full impact trial is feasible. Based on the statistical, qualitative, and health economic data collected, some minor modifications to the methods are necessary when moving forward to the full trial.

The critical care trainee forum: Impact on trainee morale and patient care

Gayathiri Sivasubramaniam, Hina Pattani

Guy's and St Thomas' NHS Trust, London, UK

Abstract

In clinical practice, when harm or potential harm occurs to patients, this can adversely impact upon the morale of staff involved and thereby affect the clinical care delivered to subsequent patients. To counteract this, we introduced a monthly, trainee-led forum to provide trainees with an environment within which to discuss mistakes in a confidential and non-judgemental setting. Our intention was not only to improve trainee morale, but also provide a space within which mistakes could be reflected upon, such that the entire group of trainees could make changes to their practice to avoid recurrence. In doing so, this would also expected to improve patient safety. We surveyed trainees' perceptions of these sessions, before (n = 28) and after (n = 17) their implementation. The results are summarised below in Table 1.

Table 1.

	Pre- trainee forum	Post- trainee forum
Positive emotions about participating in a trainee forum	85%	94%
Worried about legal implications of participating in a trainee forum	14%	0%
Worried about judgement/blame in a trainee forum	43%	0%
Felt a trainee forum had the potential to improve morale	64%	88%
Felt a trainee forum had the potential to improve clinical practice	42%	100%

The results depict that though trainees tended to be positive about a trainee forum, there were a substantial proportion who expressed concerns about judgement from colleagues and legal implications. After attending these sessions, trainees were not concerned and more positive about the potential impact on trainee morale and clinical practice.

Though this is a pilot study, its results are encouraging, especially because it is an intervention that does not require any additional resources other than time and space for a meeting. The main challenge of this project is in quantifying its impact on patient safety. Initial qualitative data suggested that trainees felt it was helpful to be alerted to possible clinical situations that they may need to face in the future, and being aware of what resources were available in the event of an emergency. Ongoing data collection in a larger quality improvement project will be required to quantify this.

Should we aim for hypocaloric feeding in early critical illness? A single centre experience

Gemma Fagan¹, Elizabeth Smith², Alastair Duncan^1,2, Liesl Wandrag² and Danielle Bear²

¹King's College London, London, UK

²Department of Nutrition and Dietetics, Guy's and St Thomas' NHS Foundation Trust, London, UK

Abstract

Background: Recent guidelines for nutrition support in critically ill adults recommend that patients at high nutrition risk are advanced towards goal feeding as quickly as tolerated over 24–48 hours. However, hypocaloric feeding is deemed sufficient for those at low nutrition risk. An enteral feeding protocol is currently in place on the Intensive Care Unit (ICU) in our institution, with the intention of meeting full energy and protein targets within 72 hours. The aim of this study was to examine whether our current protocol and practice need to be adapted to follow guidelines for hypocaloric feeding in critical illness.

Methods: A prospective service evaluation was undertaken over a 5 week period. Patients were included if they were ≥18 years old, mechanically ventilated for ≥72 hours and due to receive nutrition support via the enteral or parenteral route. Along with demographic information, the modified Nutrition Risk in the Critically Ill (NUTRIC) score was calculated retrospectively for those patients with the relevant information available. Nutritional adequacy was assessed based on the energy and protein received from enteral (EN) and parenteral nutrition (PN), and energy from non-nutritional sources (Propofol, IV glucose and regional citrate anticoagulation). We defined adequacy as ≥80% of target energy and protein. The amount of energy and protein received was expressed as a percentage of nutritional targets, both daily and overall at the end of the study period. Data were collected for the full duration of the ICU stay until discharge, death or if the patient progressed from EN or PN to oral diet.

Results: A total of 52 patients were included. The median NUTRIC score on admission was 4 (IQR 2–5) and 26% of patients were considered to be at high nutrition risk. Adequacy of energy and protein did not differ between those with a low and high score (p = 0.724, 0.599). Overall, 84.2% (IQR 69.4–94.7%) of energy targets and 81.1% (IQR 65–93) protein targets were met. However, adequate energy provision did not occur until day 4, and adequate protein provision did not occur until day 6 (Figure 1). Despite adequate energy and protein targets being met overall, underfeeding was still common, with only 36.5% of patients meeting ≥80% of both energy and protein as per pre-defined targets throughout their admission.

Figure 1.

Average percentage of energy and protein targets delivered during first 7 days of ICU admission.

Conclusion: Hypocaloric feeding is occurring in the early days of critical illness on this ICU, despite the presence of a feeding protocol which aims to meet nutritional targets. Hypocaloric feeding should therefore not be an aim of nutritional practice on this ICU, however, further exploration of differences of nutrition adequacy and outcome in patients at low and high nutrition risk is required.

Correcting serum acetaminophen deficiency with intravenous paracetamol in critical care: just because we can, doesn't mean we should

Robert Lyndon Jones, Maryam Crews, James Bordoni and Jennifer Hughes

Royal Liverpool University Hospital, Liverpool, UK

Abstract

Intravenous paracetamol is a drug given commonly in the intensive care setting for both its analgesic and antipyretic properties and due to its perceived safety as an ‘innocent’ drug, its use has gained widespread popularity. However, the recent literature has raised questions over its benefit in treating pyrexia, its potential to cause hypotension and also its cost compared to oral and nasogastric preparations. Controversy exists over parenteral dosing in patients with chronic malnutrition and liver disease, and clinician unawareness of adjusted dosing in patients weighing less than 50 kg has led to iatrogenic overdose and even death in one case. This audit aimed to assess trends in intravenous paracetamol prescribing among critical care clinicians in a large teaching hospital.

Information concerning the prescription of intravenous paracetamol was collected as a snapshot over a 9 day period in April 2018, from a total of 31 patients based in critical care areas within the Royal Liverpool University Hospital. Collection of data was retrospective and involved analysis of both paracetamol prescriptions and patient observation charts. From the gathered information, an impression of prescribing trends amongst physicians was composed.

The vast majority of patients (87%, n = 31) had a prescription for parenteral paracetamol. This was inclusive of 13 surgical patients, who received regular intravenous paracetamol 4 times daily for a cumulative total of 79 out of 84 days spent in a higher level care setting. In the surgical cohort, this equates to 316 doses (1000mg per 100 mL preparation), and a volume averaging 2.4 litres per patient contributing to their overall fluid balance.

Of those with a prescription for intravenous paracetamol, only 10 experienced a temperature of greater than 37.9°C (and only 3 of these spiked a temperature of greater than 38.9°C) – all 10 patients had an alternative route (i.e. oral or nasogastric) available for drug administration.

Considering that intravenous paracetamol is over 6 times more expensive than a nasogastric-suitable preparation and almost 60 times more expensive than oral preparations, we concluded that significant savings could be made if intravenous paracetamol was used judiciously rather than being prescribed freely. Following completion this audit, we proposed the introduction of a number of recommendations to consider with regard to the prescription of parenteral paracetamol, including limiting the duration of regular intravenous paracetamol for surgical patients to 48 hours postoperatively and only prescribing an intravenous preparation for a significant pyrexia when the enteral route is unavailable. These recommendations have not only be designed to stem the liberal prescription of a resource that is cumulatively expensive, but also to raise awareness amongst clinicians that intravenous paracetamol is not the risk-free drug that it was once thought to be. Moreover, these recommendations aim to challenge the clinician to streamline their prescribing practice, and the framework could potentially be modified to include other drugs used frequently in the critical care setting.

Rocuronium-induced anaphylaxis causing systolic anterior motion of the mitral valve and haemodynamic collapse – a case report

Alex Smith, Li Tan and Con Giannellis

Epworth Hospital Richmond, Victoria, Australia

Abstract

We present a case of systolic anterior motion of the mitral valve (SAM) caused by anaphylaxis to rocuronium in a patient who presented for an elective total knee replacement (TKR). To our knowledge, this is only the second documented case of rocuronium-induced anaphylaxis causing LVOT obstruction secondary to SAM with the first case reported only last year.

A 67 year old patient presented room for a TKR and received a spinal anaesthetic followed by a general anaesthetic. After administration of rocuronium, he suffered haemodynamic collapse, bronchospasm and ultimately cardiac arrest. Return of spontaneous circulation (ROSC) was achieved after 35 minutes of CPR but he was significantly adrenaline-dependent, requiring an infusion of 80 micrograms/minute to maintain his blood pressure. On arrival in the ICU, a junior trainee performed a focused cardiac ECHO to review for other causes of shock, given that the patient remained critically unstable. The left ventricular size was small and hyperdynamic with severe concentric LV hypertrophy. On close inspection of the valves, the anterior leaflet of the mitral valve was seen obstructing the left ventricular outflow tract (LVOT) in the parasternal long axis view (PLAX) in systole. There was evidence of mitral regurgitation on colour flow Doppler. Following the findings of this study, a noradrenaline infusion was commenced, 1.5L of IV 4% albumin was administered and the adrenaline infusion was gradually weaned off. The patient improved over the coming days and recovered without significant sequelae.

Anaesthetic agents can cause a reduction in cardiac output by means of reduction in systemic vascular resistance (SVR). This change in loading conditions results in a compensatory increase in LV contractility and alterations in flow across the LV outflow tract (LVOT). As a result of drag forces and the Venturi effect, blood flow velocity increases and the movement of the anterior mitral valve leaflet into the LVOT during systole results in obstruction to forward flow. In this case, the increased gradient across the LVOT was exacerbated by the use of spinal anaesthetic prior to induction of general anaesthesia. The haemodynamic balance was already compromised and then the circulation was challenged further by an anaphylactic reaction to rocuronium, resulting in cardiac arrest. This theory was supported by raised mast cell tryptases.

This case is important and worthy of presentation to the wider intensive care community for several reasons. Firstly, to increase clinical suspicion of both rocuronium anaphylaxis and SAM in the perioperative patient who presents with shock. This is particularly relevant given the publication of NAP6 earlier this year which reports on serious perioperative allergic reactions just like this case. Secondly, it is a rare cause of shock with a very different treatment strategy. Finally, and most importantly, we present this case to highlight the importance of focussed echocardiography in the shocked patient. This diagnosis was found by a junior trainee whose ECHO findings caught the diagnosis in time and changed the management. It is a fantastic example of POCUS in the hands of trained practitioners improving the care of our critically ill patients.

Quality of ICU patient discharge medication review: Effect of a mandatory discharge checklist

James Parker, Joanne Barton, Heather Mee and Richard Bourne

Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK

Abstract

Interfaces of care pose increased patient risks, including medication errors in ICU patients transferred to the ward.¹ A review of local ICU medicines-related incidents with adverse patient events identified the potential for improved planning and communication of medication therapy upon ward discharge. Using a multi-professional group, a microsystem coaching academy approach was employed to process-map patient medication review on ICU to ward discharge.² Interventions reported to improve discharge medication reviews and planning were considered.³ A checklist intervention was chosen and then designed to address a perceived lack of standardised medication review, prescription and communication. This quality improvement project aimed to explore the effect of implementation of the discharge checklist on the quality of patient medication discharge documentation.

Quality improvement programme with continuous patient discharge medication documentation data covering baseline and post-implementation of the mandatory checklist periods. Critical care outreach sisters were educated by pharmacists to review discharge prescriptions and information according to 6 criteria: Level 2 medicines reconciliation documented⁴?; Chronic medication prescribed or plan for review/re-introduction?; Prescriptions annotated correctly (O/A, new start date, dose change)?; Specialist charts prescribed (if required)?; Any inappropriate medicines stopped/weaning plan documented?; and Antibiotic review date/therapy codes (if required)?

Figure 1.

WinChart demonstrating significant process change after implementation of mandatory discharge checklist.

The outreach sisters then reviewed all ICU patients (with a length of stay over 48 hours) discharged to the ward, conducting a 30-week continuous study to evaluate the effects of the checklist implementation. The mandatory electronic discharge checklist was incorporated into the MetaVision® clinical information system on 30/01/17 comprising 4 essential actions requiring clinician signature prior to allowing patient discharge (Table 1). The checklist was supported by an education programme and supporting information was linked. Discharge medication reviews were graded on a 4-point scale (0 = “not completed” to 3 = “fully completed”) by the outreach sister with pharmacist advice if required. “Not applicable” responses were excluded from the results. The “completeness score” was reported as a percentage of the total marks available and results were analysed using the statistical process control software WinChart^©.

Baseline: 88 patient discharges were continuously reviewed with a mean “completeness score” of 87.8% and lower confidence limit of 54.3%; lowest score 50%.

Checklist implementation: 89 further patient discharges were reviewed after checklist implementation. By 14/02/17 a significant process change, derived from Shewhart’s principles (p = 0.005), was identified in the discharge system (Figure 1).⁵ 10 consecutive discharges had an improved mean “completeness score” of 93.9%, and lower confidence limit of 78.2%. However, after these 10 discharges mean “completeness scores” decreased and variance increased, returning to pre-checklist performance levels.

Overall, the checklist transiently improved mean completeness scores and reduced variance compared to baseline, but this was not sustainable. ICU discharge medication review is complex and a simple checklist appears unlikely to improve quality consistently. It has identified the need to further explore discharge processes and as such, the multi-professional group are now focusing on further process developments and spread.

Improving the documentation of intra-hospital transfers of critically ill adults at The Christie NHS Foundation Trust

Sam Clark and Vidya Kasipandian

The Christie NHS Foundation Trust, Manchester, UK

Abstract

Background: Critically ill adults (those who require a level of care greater than that normally provided on a standard hospital ward) may be moved within the same site multiple times per admission (e.g. for escalation of care or investigations) – each constituting an intra-hospital transfer. The same level of preparation, care and supervision is required for these transfers as is from one hospital to another, and standardised documentation should be used ^[1,2].

Methods: A retrospective case note audit was performed initially selecting ICU patients who had either CT or MR imaging. The paper case record and electronic patient record (EPR) were examined for documentation of the transfer against agreed standards. Staff were also surveyed for their awareness of their responsibilities.

Following this initial audit, change was implemented involving raising awareness to staff; updating the standard operating procedure in line with guidelines; and reproducing a regional checklist with an accessible version in the EPR. A re-audit was performed to examine the impact which included outreach admissions to ICU.

Results: 40 responses to the staff survey were received. More staff recalled having performed an intra-hospital transfer than could recall having completed transfer training (n = 33 vs. n = 7). Only 18/40 staff knew of the checklist for documentation.

Data was collected from 20 and 27 patient records respectively (original audit and re-audit). There was improvement across 6/8 standards: details of transferring personnel documented (13% vs. 30%); pre-departure checks of patient documented (0 vs. 22%); pre-departure checks of equipment documented (0 vs. 19%); observations for the duration of the transfer documented (7% vs. 11%); clinical note made on return of transfer and any significant events (13% vs. 41%); and the regional transfer checklist completed electronically or paper copy filed in the notes (0 vs. 15%).

In 2 standards compliance deteriorated: there is documentation in some form that a transfer took place (73% vs. 70%); and the transfer was initiated on the EPR (13% vs. 0).

Discussion: Following implementation of change there was a small improvement in compliance. This may partially be attributed to the inclusion of admissions to critical care by outreach staff who were noted to meet some standards through free text documentation in the EPR.

Table 1.

Serial ABG results.

	Day 1			Day 2		Day 3		Day 4
	1	2	3	4	5	6	7	8	9
pH	7.78	7.75	7.76	7.58	7.55	7.72	7.5	7.51	7.5
pCO2	2.8	3.1	3.6	4.6	5.1	3.1	4.9	4.9	5
K	1.9	1.6	1.6	2.5	2.9	2.6	3.4	3.7	4.4
Lac	6.4	5.8	5.8	3.8	1.5	3.1	3.3	1.9	2.3
HCO3	34.4	35.4	35.4	32.8	32.7	32	29.1	29.4	28.8

Three domains may have contributed to a failure to achieve significant improvement:

1. A lack of availability of the paper transfer checklist (although the electronic version remains ‘live’ since implementation)

2. Equipment issues. The observations are not retrospectively uploaded to the EPR from the monitoring system due to manufacturer incompatibility, the fact of which some staff seem still to be unaware. Furthermore, different equipment (transfer bags and monitors) are used by the critical care staff and the outreach staff

3. Staff training may be inadequate. There has been no practical transfer training or formal training on the transfer section of the EPR.

It has been agreed to update the transfer section of the EPR with a localised checklist and to provide practical transfer training to staff during a month dedicated to transfer education. A re-audit will follow.

Lactic alkalosis – a rare consequence of hypokalaemia in a malnourished patient?

Bhaskar Narayan and Theophilus Samuels

Surrey and Sussex Healthcare NHS Trust, Redhill, UK

Abstract

Introduction: We describe a patient presenting with the interesting and highly unusual combination of severe alkalosis and hyperlactataemia in the context of severe hypokalaemia; possible causes and mechanisms for such a metabolic disturbance are discussed below.

Case: A 56-year-old man attended the emergency department complaining of generalised weakness. He gave a history of sustained heavy alcohol consumption and poor diet but specifically denied any vomiting. Fluid resuscitation was initiated and blood samples drawn for blood gas and laboratory analysis.

Initial ABGs showed severe mixed metabolic and respiratory alkalosis, hypokalaemia and hyperlactataemia. These results (and subsequent ABGs) are summarised in Table 1.

After resuscitation with 3 Litres intravenous crystalloid (including potassium replacement), he remained hypotensive. The alkalosis and raised lactate persisted despite fluid resuscitation and vasopressor support.

By Day 2, lactate started to decrease, serum potassium was rising, and the pH was decreasing. Vasopressor support was weaned off. However, on Day 3, there was a transient dip in serum potassium, associated with a relapse in the alkalosis and rise in lactate. After further potassium replacement, these changes reversed. (see Table 1).

Results: Serum potassium was strongly inversely correlated with pH (r = −0.89) and lactate (r = −0.82). pH and lactate showed a strong positive correlation (r = 0.84; the opposite of what would normally be expected).

Discussion: Severe alkalosis and hyperlactataemia is an unusual combination. Although it has been previously described in patients with large volume vomiting¹ or lower GI losses², our patient did not have any such history. To the best of our knowledge, this is the first documented case of such a presentation. We suggest that the biochemical abnormalities resulted from potassium depletion due to poor oral intake and renal losses secondary to excessive alcohol consumption.

Severe hypokalaemia can cause metabolic alkalosis by increasing the uptake of hydrogen ions from the extracellular to intracellular compartment, in exchange for potassium ions, and possibly also increasing renal ammonium secretion.³ Although the raised lactate was partly due to hypovolaemia and hypoperfusion (Type A; liver synthetic function was preserved), an alkalaemic state may in itself promote a rise in serum lactate. An increase in pH shifts the oxygen–haemoglobin dissociation curve to the left, leading to decreased oxygen release in tissues^2,4. Raised pH may also increase intracellular lactate production as part of a feedback loop to maintain normal intracellular pH.²

Learning points:

• Malnourished patients can have severe electrolyte deficiencies.

• Severe hypokalaemia affects acid-base homeostasis and can cause metabolic alkalosis.

• While Type A and Type B lactic acidosis are widely recognised, severe metabolic alkalosis is a third cause of raised serum lactate.

Understanding why trials in critical care work (or not): the role of process evaluation

Lydia Emerson¹, Danny McAuley¹, Mike Clarke², Thomas P Hellyer³, A John Simpson³ and Bronagh Blackwood¹

¹Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

²Centre for Public Health, Queen's University Belfast, Belfast, UK

³Institute of Cellular Medicine, Newcastle University, Newcastle Upon Tyne, UK

Abstract

Background: Process evaluation in clinical trials explores the characteristics of the setting, adherence to the components of the intervention, and elicits information about barriers and facilitators to implementation. To date, there have been no process evaluations of clinical trial interventions within the ICU, with previous efforts focusing on quality improvement. We hypothesised that the ICU environment presents specific complex challenges for trial implementation and delivery, and sought to develop and test a framework to provide guidance for conducting a process evaluation alongside multi-centre randomised trials in UK ICUs.

Methods: Our design was influenced by Steckler and Linnan’s process evaluation guidance (1) and drew on the principle that interventions work (or not) in part, based on how those responsible for their implementation/delivery respond to them (2). We examined attitudinal factors to understand how staff engage with trials and interventions. We took an ‘all-trial’ approach, interviewing staff throughout the entire trial trajectory, which we hypothesised would detect any change in implementation/delivery over time. Logic models illustrate how programmes or interventions are intended to work to produce desired outcomes. We interviewed staff periodically throughout the trial, using a logic model to inform our interview guides; exploring pathways between intervention implementation, delivery, and outcomes. We also examined routinely collected trial data relating to screening and recruitment. We used the multi-centre trial VAPrapid-2 (ISRCTN65937227) to develop and test the preliminary process evaluation framework.

Results: Our findings suggest that process evaluations in critical care should adopt a three-phase approach.

1. Baseline (pre-trial): Less routine use of bronchoalveolar lavage (BAL) in the diagnosis of ventilator-associated pneumonia (VAP), and a culture of not actively de-escalating antibiotics appeared to correlate with lower recruitment. This shows the influence that pre-trial clinical practice/attitudes has on the acceptability and uptake of new interventions.

2. Exploration (during-trial): The concept of ‘trial champions’ emerged at this stage, the absence of which appeared to correlate with lower recruitment rates and a lack of engagement in units. This highlights the importance of identifying emerging barriers and facilitators, and examining attitudes and perceptions towards the intervention and trial.

3. Clarification (end-of-trial): Lower recruiting units discussed a greater perception of risk for BAL (favoring less invasive methods outside the trial protocol) and for discontinuing antibiotics (favoring antibiotic use as the perceived low-risk approach). Phase three data are central in understanding how and why interventions are implemented, delivered, and sustained or not over time.

Attitudinal factors emerged as key factors in critical care trials, directly impacting upon staff willingness to engage with trials, and the degree to which they deliver interventions as intended. Trial logic models should be used to inform interview guides, as they enabled us to identify and test potential risk-points within the study design. Routinely collected trial data pertaining to recruitment were good indicators of the degree to which units engaged with trials.

Conclusion: The framework was a good fit for complex critical care trials. Data from process evaluations undertaken alongside critical care trials will inform design of future such trials or help to implement trial findings.

Benefits and pitfalls of pop-up simulation on the intensive care unit

Hannah Burns, Hajnalka Hideg

St Georges Hospital, London, UK

Abstract

Background & Aims: Simulation is a growing resource utilised throughout different areas of healthcare. It is useful for improving clinical skills, but also beneficial for non-clinical skills such as communication, leadership and situational awareness.¹ Simulation can be used for a number of clinical situations but often lends itself to emergencies.

We wanted to adopt simulation into the education programme within our intensive care unit and after assessing for the wants of MDT we initiated a program with the aim of delivering regular, relevant simulation sessions. We wanted to make it reproducible, to have multiple events, involving the whole cohort of doctors and nurses and to become engrained within the ethos of the unit.

Methods: As a result of a questionnaire to learners we developed four scenarios based on clinical situations that were frequent on the ICU and where staff may feel less confident. These were cardiac tamponade, ventricular fibrillation, major haemorrhage and ventilator failure.

Co-developers of the program were junior doctors, nurse educators and a consultant. Assistance from the simulation centre was obtained for equipment and advice. These pop-up simulations were to be run regularly with a schedule adopted between the key team members involved.

Feedback questionnaires were given to participants and observers. They rated a number of components such as the content, structure, simulation and overall feedback of the situation. The rating system was from 1–5 with 5 being strongly agreeing the statement.

Results/Experience: We facilitated 10 sessions across a four month period. There were weeks missed due to unit busyness, staffing levels being inadequate, or not having simulation leads on shift. Most scenarios were run with low-fidelity simulations and for some we borrowed high-fidelity mannequins.

All questions were rated positively with a rating of 3.9 to 4.9 equating to agree to strongly agree. The statements that were most positively rated were:

• “Content was relevant to my training”,

• “The simulation was designed appropriately to my learning level”,

• “Knowledge I gained from the session will be helpful to me in my practice”

• “The debriefing enhanced my knowledge”

The least positive statements were:

• “The simulation felt real”

• “There was a clear introduction to the subject”

In the educational feedback at the end of this placement, simulation was rated as most valued educational method amongst trainees, whereas in previous cohorts it was rated 5^th.

Conclusions: This service received well by the learners, the common issues we have identified to aid similar units wishing to create a regular pop-up simulation programme:

• Running a pop-up simulation on a busy unit meant planned training was occasionally cancelled or rescheduled.

• Using core trainers requires a rota and dedication to be available outside working hours

• Students feeling our low-fidelity simulations were not realistic (common in simulation but learning outcomes still valuable¹)

• Pop-up nature meant learners felt there was a lack of introduction to the subject- we placed the “patients” on handover lists to try rectify this

Our unit has developed this further by dedicating scheduled time for simulation in the overall teaching programme.

lung ultrasound and the role of lung aeration score in patients with acute respiratory distress syndrome on extracorporeal membrane oxygenation: a prospective observational study

Stefanie Curry¹, Luna Gargani², Oraina Ng³, Andrew Roscoe¹, Kiran Salaunkey¹, Angus Butchart¹, Bobby Agrawal¹ and Antonio Rubino¹

¹Royal Papworth Hospital, Papworth Everard, UK

²Institute of Clinical Physiology CNR, Pisa, Italy

³Singapore General Hospital, Singapore, Singapore

Abstract

Acute respiratory distress syndrome (ARDS) is a respiratory condition characterised by progressive loss of lung aeration and impaired gas exchange. In its most severe form, ARDS may require veno-venous extracorporeal membrane oxygenation (VV-ECMO) to maintain adequate physiological parameters of ventilation. Strict imaging protocols play an important role in clinical management of patients receiving VV-ECMO. Lung ultrasound (LUS) represents an emerging bedside imaging technique, allowing for quantification of daily aeration scores based on the observed pathological aspects of lung injury in ARDS. The potential use of LUS as a tool to monitor ARDS pathology represents a significant opportunity to enhance patient care and optimise treatment delivery. The aim of this study is to determine if lung aeration scores calculated using LUS correlate with the severity of lung injury identified on bedside chest radiography (CXR) in patients with ARDS receiving VV-ECMO, using computed tomography (CT) as the reference standard.

This single-centre prospective, observational study investigated consecutive adult patients diagnosed with ARDS receiving VV-ECMO admitted to Royal Papworth Hospital between October 2017-February 2018. LUS assessments were conducted to examine 12-lung regions within 24-hours of patient admission and daily thereafter until liberated from VV-ECMO. Daily lung aeration scores were calculated according to appearance of four defined ultrasound patterns. CT and CXR were conducted as per hospital protocol and an independent consultant radiologist calculated aeration scores retrospectively. Correlations between LUS aeration scores and those of CXR and CT were tested using Spearman correlation rank analysis and assessed for bias and level of agreement using the Bland-Altman method. Inter and intra-observer agreements were assessed using kappa coefficient tests. Statistical significance was fixed at 0.05.

Consecutive adult patients (n = 30) were prospectively screened for inclusion and data were collected from 18 patients (42yrs ±16.07). Correlations in aeration scores indicated statistically significant results for LUS vs CT (rho = .868, p = .002) and for LUS vs CXR (rho = .498, p = .018) with good agreement and no evidence of proportional bias (B = .059; p = .664; B = .085; p = .683). Highly statistically significant correlations were found for identification of pleural effusion (rho = .601; p < .010) and pneumothorax (rho = 792; p < .001). Inter-observer reliability indicated good reliability between all operators (K = .726, 95% CI:.602-.830, p < 0.005; K = .699, 95% CI:.577-.810, p < 0.005; K = .640, 95% CI:.489-.766, p < 0.005) and good intra-observer reliability for individual operators (K = .700, 95% CI:.569 –.807, p < 0.005; K = .642, 95% CI:.517 –.750, p < 0.005).

Lung ultrasound was demonstrated to correlate with CT for quantifying lung aeration with a good level of agreement and no evidence of proportional bias. Correlations of aeration scores between LUS and bedside CXR were also demonstrated. For identification of pleural effusions and pneumothorax, LUS was shown to be accurate with a good level of sensitivity and specificity. These results quantify a statistically significant relationship between the contemporary lung imaging techniques and LUS. In clinical practice, LUS offers clinicians a reliable and reproducible bedside imaging technique for real-time evaluation of lung pathology. In conjunction with traditional imaging techniques, routine use of LUS should be considered for assessment and monitoring in this patient group.

Case report of extracorporeal CO2 removal in a District Hospital

Jonathan Pang and Raghib Malik

Romford Queen's Hospital, London, UK

Abstract

65 years old gentleman with 50pack years smoking history, stroke was admitted with increasingly dysponea over 2–3months with weight loss and appetite. His ABG showed Type1 Resp failure with metabolic acidosis on 15L/min.

He was diagnosed as community acquired pneumonia and supported with high flow nasal oxygen therapy with increasing oxygen requirment and subsequently intubated. Day 3, mechanical ventilation with worsening inflammatory markers with renal impariment and a CT chest showed {fig 1}, bronchoscopic Alveolar lavage: Inflamatory changes with no malignancy cell. Sputum grew Coliform growth. His antibiotics were escalated to Tazocin due to rising in inflammatory markers on Day9. With negative auto-immune screen and fungal screen, AIP was diagnosed and pulsed methylprednisolone was initiated with little response. Discussion with Extra-coporeal Membrance Oxygen Therapy (ECMO) centre gave the opinion he was not a suitable candidate due to possibility of malignancy. Day 12 he was proned to improve lung compliance with little response. Day14 Extracorporeal CO2 Removal (ECCO2-R) was started and the reduced the tidal volume and PIP to 32 to 35 (previously was upto 50 to 60). In Lung cancer MDT, consensus was more likely to be inflamatory process than neoplasia but malignancy could still not be ruled out. He continued to deteriorated further and decision was made to withdrawn and he passed away on 8/10/2017.

Discussion: This case report aims to inform options of ECCO2-R in district Hospital when tertiary centres have turned down patients for ECMO. One crucial beneficiary effect of ECCO2-R is to super-protect ventilation for patients with ARDS by reducing tidal volume and minute volume preventing ventilator induced lung injury {2}. From the study, Bein & team found that ECCO2-R impacted morbidity as more hypoxaemic patients had more ventilator free days with this threapy. Looking into the future, Italy is conducting clinical trials investigating ECCO2-R as treatment option for COPD patients who have hypercapnia respiratory acidosis failing to respond to non-invasive ventilation{3}

Free fatty acids level In CSF, stroke severity and comorbidity indices as independent short-term prognostic factors in acute ischemic stroke (AIS) in ICU

Sherine Mohamed Ibrahim¹, Mahmoud Mohamed Kenawi², El Sayed Gaber Ali³ and Nora Abbas³

¹Ahmed Maher Teaching Hospital, ICU, Cairo, Egypt

²Kasr ALAIny school of Medicine, Cairo, Egypt

³KAsr alAiny Faculty of Medicine, Cairo, Egypt

Abstract

Abbreviations

• AF=atrial fibrillation

• AIS=acute ischemic stroke

• AUC= Area under the curve

• CSF= cerebrospinal fluid

• FFA=Free fatty acid

• GCS = Glasgow Coma Scale

• ICU= Intensive Care Unit

• mRS = modified Rankin Scale

• NIHSS=National Institutes of Health Stroke Scale

• ROC = Receiver operating characteristic

Abstract

Background: There are limited data about the prognostic value of free fatty acid (FFA) levels in cerebrospinal fluid (CSF) in AIS. The purpose of this study was to investigate the prognostic value of FFA in CSF as an independent biomarker in addition to the co-morbidity indices for assessing the functional outcome and mortality of AIS.

Methods: In a prospective cohort study, from November 2016 to May 2017, we recruited 80 critically ill acute ischemic stroke (AIS) patients, admitted to ICU. FFA levels in CSF were measured on admission; NIHSS, GCS and mRS were evaluated on admission and at 30 days. Univariate and multivariate analysis were used to evaluate stroke outcome according to FFA levels in CSF.

Results: There was positive correlation that linked the levels of FFA in CSF to the NIHSS score ≥16 at 30 days (P value 0.049) and to GCS <7 at 30 days (P value 0.048).

There were positive significant correlation between levels of FFA and the infarct volume ≥145 ml was found ((P value 0.001).

Also, there was a significant linear correlation between FFA levels in CSF and functional outcome as defined by mRS score >2 at 30 days (P value 0.037).

Our study also showed that there was statistically significant linear correlation between FFA levels in CSF and all-cause mortality, (P = 0.007). Based on the Receiver operating characteristic (ROC) curve, the optimal cut-off value of CSF FFA levels as a predictor for mortality in AIS was 0.27 nmol/µl, with specificity 72.2% and sensitivity 62.9%.

Conclusion: Our study revealed that FFA levels in CSF on admission were correlated with stroke severity, functional outcome, and infarct volume and mortality among AIS patients admitted to the ICU.

Keywords: FFA in CSF, AIS, mortality, CCI, NIHSS, GCS, mRS

Audit of Initial Mechanical Ventilation (IMV) Settings on the Intensive Care Unit at Royal Surrey County Hospital

Idrisu Sanusi, Ben Creagh-Brown, Chia Ling Tey and Michael Webb

Royal Surrey County Hospital, Guildford, UK

Abstract

Introduction:

• Research findings advocate the use of ‘lung protective ventilation’ strategy, by targeting a tidal volume (TV) of 6mls/kg in mechanically ventilated patients

• This strategy reduces the risk of developing ventilator induced lung injury and its associated morbidity and mortality.

• We performed an audit to analyse if we have been achieving the desired low TV strategy and transitioning patient into spontaneous ventilation at the earliest possible opportunity.

Methodology:

• 143 patients identified using WardWatcher

Over 6 months (April to September 2016)

Of newly intubated patients in ICU or arrived intubated.

• Initial ventilation settings data were collected from the electronic patient record system (ICIP).

• Data was plotted into a Microsoft Excel spreadsheet with set formula to calculate the approximate dynamic compliance, TV as per actual body weight (ABW), the driving pressure and the median number of hours of mandatory ventilation.

Results:

• When using patient's ABW:

73% of patients received a TV of less than 8mls/kg.

85% had a PeAP <30 cmH2O.

75% had a driving pressure of ≤15 cmH2O (Ppeak-PEEP).

• Median number hours of an entirely mandatory ventilation: 15 hours.

• However, when using estimated ideal body weight:

47% of patients received a TV <8mls/kg; in which 25% received TV >9.6 ml/kg

Discussion:

• There is significant room for improvement in achieving low tidal volumes based on ideal body weight in mechanically ventilated patients on our unit.

• We developed an initial ventilation protocol aiming to guide ICU nurses, ICU physiotherapists and trainee ICU doctors

• The aim of this protocol is to achieve lung protective ventilation targets to minimise potentially injurious ventilation with high TVs and to avoid patients being ventilated in mandatory mode longer than necessary to prevent respiratory muscle wasting.

CC3N: National Critical Care Nursing and Outreach Workforce Survey

Claire Horsfield^1,2, Andrea Berry³, Chris Hill⁴, Angela Himsworth⁵, Julie Platten⁶ and Stephen Littleson⁷

¹Critical Care Networks -National Nurse Leads, Preston, Lancashire, UK

²Lancashire and South Cumbria Critical Care Network, Preston, Lancashire, UK

³Critical Care and Major Trauma Network Manager, West Yorkshire, UK

⁴Network Nurse Lead, North East North Central London, UK

⁵Network Nurse Lead, Midlands Critical Care and Major Trauma Network, UK

⁶Critical Care Network Manager, North of England, UK

⁷Midlands Critical care and Major Trauma Network, Birmingham, UK

Abstract

Overview: In 2017, the National Networks Nurse Leads Group (CC3N) facilitated a National Critical Care Nursing and Outreach Workforce Survey across England, Wales, and Northern Ireland. This poster presentation summarises the results and presents changes that have occurred over time.

Methods: The survey tool was developed in 2015 utilising Microsoft Excel 2003. It was devised around the standards detailed within the Guidelines for the Provision of Intensive Care Services (ICS, FICM, 2015) and the Core Standards for Intensive Care Units (ICS, FICM, November 2013) and the then named draft D16 (now D05) Service Specification Framework for Adult Critical Care (NHSE, February 2014). The survey was distributed via the network nurse leads and returned to a central point for collation and analysis.

Results: Nursing workforce data was returned from 242/288 clinical areas (84%) an increase from 2015 (70%). At the time of the survey, there were a total of 1944 level 3 and 1341 level 2 beds with 17767.88 whole time equivalent (WTE) registered nursing posts to staff these beds. A significant proportion of posts were vacant at the time (8.3 %), with an increased number of units seeking to recruit from EU and non-EU countries (n = 57 in 2015, n = 80 in 2017). The proportion of registered nursing staff from EU and non-EU countries varies considerable by region, with some areas employing only 1.7% of the workforce from abroad, others have up to 66.4% of the nursing workforce from overseas. Despite recruitment challenges, almost all respondents stated they met the requirement to allocate 1 registered nurse to care for one level 3 patients (100%) and 99.5% stated they met the requirement for 1 registered nurse to care for 2 level 2 patients. There has been an increase in the provision of a supernumerary clinical coordinator across all shifts, although larger units are more able to meet this standard. Staff turnover and the % of staff over 50 years of age has reduced since the previous survey along with the use of agency staff to supplement nursing numbers. Of note is the increased % of clinical educators alongside the increased uptake of the STEP competency framework (CC3N, 2012), with 97% of units now utilising some or all part of the competency package. With regard to outreach provision, 85% of hospitals have an outreach service and results show an increase in the number of shifts/hours covered, but the composition of teams remain extremely varied.

Conclusion: Recruitment to critical care nursing posts continues to be a challenge and this survey has helped to quantify the amount of vacancies along with the changes over time in relation to turnover, overseas recruitment, agency use and proportion of posts at each band. Despite the perceived challenges, there has been demonstrable standardisation of critical care nurse education supported by an increased investment in educator provision. Regional variation exists for a variety of workforce issues and subsequent surveys will track changes in the future.

Patients’ support needs during and after the critical illness event: a scoping review

Judy King¹, Brenda O’Neill², Pam Ramsay³, Mark Linden⁴, Alia Darweish Medniuk⁵, Joanne Outtrim⁶ and Bronagh Blackwood⁷

¹School of Rehabilitation Sciences, Faculty of Health Sciences, University of Ottawa, Ottawa, Canada

²Centre for Health and Rehabilitation Technologies, INHR, Ulster University, Newtownabbey, UK

³School of Health and Social Care, Edinburgh Napier University, Edinburgh, UK

⁴School of Nursing and Midwifery, Queen’s University Belfast, Belfast, UK

⁵Department of Anaesthesia, Southmeade Hospital. North Bristol NHS trust, Bristol, UK

⁶Division of Anaesthesia, Department of Medicine, University of Cambridge, Cambridge, UK

⁷School of Medicine, Dentistry and Biomedical Sciences, Queen's University Belfast, Belfast, UK

Abstract

Background: Intensive care survivors suffer chronic and potentially life-changing physical, psychosocial and cognitive sequelae, and supporting recovery is an international priority. As survivors transition from the intensive care unit to home, their support needs develop and change. An understanding of needs at different transition periods would help inform health service support for survivors.

Objective: To identify patients’ support needs (informational, emotional, instrumental, appraisal) following an ICU admission. We mapped these against the Timing it Right framework reflecting the patient’s transition from intensive care (event/diagnosis); to ward (stabilisation/preparation); and discharge home (implementation/adaptation).¹

Methods: We conducted a scoping review of published qualitative research studies using the Arksey and O’Malley and Levac frameworks.^2,3 We searched the major databases from 2000 to 2017. We included qualitative research studies that reported adult critical care survivors’ experiences of care and identified support needs with reference to the Timing it Right framework. Two reviewers independently screened, extracted and coded data. Data were analysed using a thematic framework approach. The review team reviewed and affirmed findings.

Table 1.

Patients’ support needs during transition from ICU to post-hospital discharge.

Information needs Initial event/ICU: what happened, diagnosis, prognosis, and illness event. Stabilisation/preparation phases: medical progress; treatments and medications for ongoing recovery. Implementation/adaptation (home/community): coping with long-term sequelae of the illness and stress; pamphlets/booklets.

Emotional needs Initial event/ICU: coping with confusion, anxiety, memories; need for comfort and security. Stabilisation/preparation phases: security, family attendance; coping with relocation anxiety, isolation, depression about lack of physical progress. Implementation/adaptation (home/community): coping with vivid memories, terrifying dreams, worry, with feeling excluded from family; psychological counselling, support from community health care providers and a support group.

Instrumental needs Initial event/ICU: managing lack of sleep, fatigue, pain, anxiety; nursing care. Stabilisation/preparation phases: physical and cognitive disabilities, re-establishing premorbid physical strength, training to relearn personal care tasks. Implementation/adaptation (home/community): performing household activities, gaining independence, achieving higher level cognitive function and pre-ICU level of strength, work.

Appraisal needs Stabilisation/preparation phases: progress feedback from staff who knew them, appreciate the mental and physical transformation, and strain on family. Implementation/adaptation (home/community): reassurance from others who had been through the ICU experience and know what was ‘normal’

Findings: From 3035 references we included 32 studies involving 702 patients. Studies were conducted in UK and Europe (n = 17, 53%); Canada and the United States (n = 6, 19%); Australasia (n = 6, 19%); Hong Kong (n = 1, 3%); Jordan (n = 1, 3%) and multi-country (n = 1, 3%). Patient needs were collected at various time-points from admission to 5-years post-hospital discharge. Informational, emotional, instrumental and appraisal social support needs differed when mapped against the time points of the Timing it Right framework and are presented in Table 1.

Conclusions: Our review is the first to identify the change in social support needs among intensive care survivors as they transition from intensive care to the home environment. The mapping of support needs across time may inform service provision.

Acknowledgements: Funded by an Intensive Care Foundation Research Priority Award

Outcomes of elderly patients admitted to ICU with emergency neurosurgical diagnoses

Michael Nesbitt

Salford Royal NHS Foundation Trust, Manchester, UK

Abstract

Aims: Our aim was to determine the outcomes of patients over the age of 75, who were admitted to the ICU at a tertiary neurosurgical centre, with emergency neurosurgical diagnoses between 2015–2017.

Method: This was a retrospective, observational study. Using ICNARC data, we established a database of all patients over the age of 75 who were admitted to our ICU between 2015–2017. We reviewed the electronic patient records (EPR) in order to acquire patient demographic data, and determine the reason for admission. All elective and non-neurosurgical admissions were excluded.

By examining the EPR, we determined the neurosurgical diagnosis and established the co-morbid status of each patient, including his or her WHO performance status and medication history.

In order to measure outcomes, we evaluated the length of stay on critical care, and whether or not the patient was successfully stepped-down to a ward. We also established whether or not the patients suffered any complications during their ICU admission, and the overall ICU mortality.

Results: There were 69 patients (n = 69) included in our study, 30 male (n = 30) and 39 female (n = 39).

Over one third of the patients (38%, n = 26) had a good WHO performance status (0 or 1), and more than half (58%, n = 40) were taking fewer than 5 medications prior to admission.

Interestingly, half of the patients (51%, n = 35) were prescribed an antiplatelet or an anticoagulant drug. The most prevalent diagnosis was subdural haematoma (n = 29), and of those patients a significant proportion (62%, n = 18) were prescribed an antiplatelet or an anticoagulant drug.

ICU mortality was almost 40% for all patients (38%, n = 26). Of those patients that survived, over 80% (81%) had a critical care stay of less than 5 days, and almost all (95%, n = 43) were successfully stepped-down to a ward without re-admission to ICU.

On critical care, respiratory complications were most common, especially hospital-acquired pneumonia (35%, n = 24), and respiratory failure was a frequent cause of death (38%, n = 10). Of the patients that died on ICU, a significant proportion were deemed ‘not for intervention’ on admission (31%, n = 8), and were admitted to ICU in order to facilitate organ donation or family discussions.

Conclusion: A significant number of our patients had a good performance status and were taking fewer than 5 medications prior to admission, this is perhaps indicative of a pre-selected patient cohort. Despite this, WHO performance status and number of medications were not predictive of survival, and mortality in our group was high. This poses interesting questions regarding how we select patients for admission to critical care, and suggests it is difficult to predict outcome by solely using these measures of co-morbid status.

A significant number of our patients were prescribed an antiplatelet or an anticoagulant drug, we believe this highlights the importance of judicious risk vs. benefit analysis before prescribing these drugs to patients in this age group.

Arterial blood gas sampling within Critical Care: An audit of practice

Anita Jones¹, Sarah Quinton², Nitin Arora³ and Marion Palmer¹

¹ACCP. University Hospitals Birmingham, Heartlands., Birmingham, UK

²Nurse Consultant. University Hospitals Birmingham, Heartlands., Birmingham, UK

³Consultant. University Hospitals Birmingham. Heartlands., Birmingham, UK

Abstract

Background: Within the NHS there is an ongoing drive to reduce expenditure and increase efficiency, challenging the practice within critical care.

Arterial blood gases are an essential element of monitoring critically unwell patients, but research suggests that arterial blood gases are over sampled, with potential repercussions such as anaemia, pain and cost, standards should be followed.

The purpose of this audit is to compare the current Trust guidelines identifying existing practice and recommending changes, ensuring best practice for our patients.

Aim: To determine if the critical care team complies with existing arterial blood gas sampling guidelines, measuring practice and identifying areas for improvement.

Method: An audit was conducted at a single centre to measure compliance with the arterial blood gas sampling standard to current practice. 500 critical care charts were reviewed and retrospective, quantitative, anonymous data was obtained from 224 charts over a six-week period. Data was collected using an electronic audit form, reviewing the relationship of ABG sampling and alteration in ventilation or oxygenation strategies.

Results: Within the audit period, 1,393 arterial blood gas samples were taken, giving an average of 6 per every 24 hours, comparison to Trust standard showed oversampling by 100%. Goal directed saturations were only documented on 14.3% charts. Only 14% of critical care charts had an oxygen titration without an arterial blood gas sample performed, despite 100% of charts having saturations monitored. 75% of charts had patient saturations documented at 99–100% exhibiting potential unnecessary over oxygenation. 492 arterial blood gases were taken with no ventilation intervention post sample.

Conclusion: The data demonstrates a necessity to reduce the amount of arterial blood gas samples. The results showed a non-compliance of practice with a measurable deficit to the current guideline resulting in unnecessary cost. Relaunching the guidelines, using them effectively could reduce ABG sampling by half, saving £35,498/year based on an ABG cost £5.68. In addition, over 11,000 mls of blood would be saved in sampling (based on 2mls per sample). It is imperative that an update of the Trust guidelines is completed, consequently decreasing cost while providing excellent evidence-based treatment to the patient.

Perfectionist or Pragmatist? Decision making on the Critical Care Unit and the Associated Prevalence of Personality Traits

Kate Grailey¹, Eleanor Murray² and Stephen Brett¹

¹Imperial College London, London, UK

²Said Business School, Oxford, UK

Abstract

Introduction: The Intensive Care Unit (ICU) requires staff to make decisions quickly and often without adequate resources. Based upon previous research by this team, ⁽¹⁾ we explored the decision-making processes employed by individuals within the ICU. Did they have perfectionist (striving for flawlessness with high performance standards) or pragmatic (realism and compromise in order to obtain a desired outcome) tendencies?⁽²⁾ We considered whether certain personality traits as defined by the 16PF tool ^(3,4) would predominate within individuals and were associated with decision-making tendencies.

Methods: Semi-structured interviews were conducted with ICU staff. These centred around a fictitious staffing scenario, developed by a multidisciplinary steering group. The scenario focused upon staff allocation, and was designed to be challenging but realistic in terms of available resources. Participants were emailed 16PF questionnaires to complete.

Prior exploratory qualitative analysis of the interview transcripts had identified two possible processes of decision making. This data was then re-coded performing a second thematic analysis to further evaluate decision-making tendencies. Statements relating to perfectionist or pragmatic decision making were coded using NVIVO software. Participants were grouped according to whether most of their decisions were perfectionist or pragmatic, or a balance between the two.

16PF data was analysed using Microsoft Excel, and statistical correlations between the presence of personality traits and decision-making tendencies explored using GraphPad Prism 7 software, applying a Spearman R non-parametric test.

Results: Interviews and 16PF questionnaires were completed by 31 ICU workers within Imperial College NHS Trust. 4 individuals were predominantly pragmatists, and 8 were perfectionists. Most participants demonstrated a balance between both pragmatic and perfectionist tendencies, depending upon the nature of the clinical decision they were facing.

Each 16PF trait is provided as a score between 1 and 10. Average scores (between 4–7) were recorded for most participants in all 16 traits. [Figure1]

The data demonstrated emerging trends in the personality traits of those with different decision making tendencies. Those with strong pragmatic or perfectionist tendencies had high scores for emotional stability in comparison to the balanced group. The balanced group did not tend to high or low scores in any of the 16PF traits. One statistically significant correlation between 16PF score was obtained; individuals were more likely to have a balanced decision-making profile if their score for vigilance was low (R = −0.823, p = 0.03) [Figure 2]

Conclusion: This study provides an initial exploration into the differences in decision making tendencies of staff members working in ICU, and how this may be associated with different personality traits. Whilst we acknowledge the low number of participants, the thematic analysis of the interview transcripts indicates that whilst most clinical staff have a balanced approach to decision making, applying both perfectionist and pragmatic ideals according to the clinical need, there are a proportion of individuals who strongly trend towards one type of decision making strategy. Further exploratory studies are required to both examine these associations further, and to evaluate the impact of different decision making strategies and personality traits on effective and safe team working within the ICU.

Ultrasound-guided subclavian venous access: regional survey of experience followed by a multispecialty teaching programme

Benjamin Millette and Nick Suarez

Oxford University Hospitals NHS Foundation Trust, Oxford, UK

Abstract

Introduction: Subclavian central venous access (SCVA) has a lower rate of infective and thrombotic complications compared to femoral and internal jugular venous access(1). Whilst data have shown higher rates of mechanical complications, such as pneumothorax, they often have high rates of insertion guided by landmark technique(1). The subclavian vein’s anatomical position is variable and the rate of mechanical complications is reduced by the use of real time ultrasound scanning (USS)(2).

We hypothesised that few UK anaesthetic or intensive care trainees have received formal training in landmark or ultrasound-guided SCVA, and are not confident in it. We sought to establish if a short teaching session with live human USS and subclavian gel phantom needling practice would improve confidence.

Methods: We conducted a survey of all Anaesthetic and ICM trainees in the Thames valley deanery to ascertain previous experience of SCVA. Having identified this specific need for training, we conducted short teaching sessions for anaesthetic and intensive care trainees at regional teaching days. This comprised a lecture followed by in vivo scanning of the subclavian vein and needling a subclavian gel phantom with real time ultrasound.

We also gathered data on participant grade, formal teaching experience in SCVA and number of times they had inserted SCVA. We compared their confidence in performing ultrasound-guided SCVA before and after the training session using a visual analogue scale (VAS).

Continuous data were checked for normality using the Shapiro-Wilk test. When this failed to detect normality, data were compared using the Mann-Whitney test.

Results: Forty-nine specialty registrars in anaesthetics or ICM responded to the SCVA training survey (response rate of 49% of deanery trainees). This revealed that 33% of respondents had sited no subclavian lines and 71% had sited 5 or fewer. Only 29% had any formal training in SCVA and mean confidence score was 35 out of 100 on the VAS. Thirty-three registrars in total were then taught over three teaching sessions. The mean confidence in SCVA prior to teaching was 25.6 and after the session this rose to 52.7 (p < 0.01) (Fig 1).

Discussion: This study demonstrates that anaesthetic/ICM registrars in our region have little experience in inserting SCVA and receive little formal training in it. Furthermore, we have shown that a short training session can improve their confidence, making them more likely to seek out further training opportunities in their routine practice. This could potentially lead to improvement in patient care by minimising complications of central venous access.

Fig 1. Confidence scores in SCVA before and after teaching.

High dose Insulin Euglycaemic Therapy (HIET) for massive calcium channel blocker overdose: A Case report

Barney Scrace¹ and Mark Sair²

¹Royal Cornwall Hospital, Truro, UK

²Derriford Hospital, Plymouth, UK

Abstract

A 69 year-old man with hypertension and type 2 diabetes mellitus presented to the emergency department having taken one month supply of his regular medication. These included (approximate dose by unit body mass) metformin 155 mg/kg, glicazide 12 mg/kg; isosorbide mononitrate 8 mg/kg, and amlodipine 3 mg/kg.

Admission clinical observations were GCS 15/15; temperature 35.9 °C; heart rate 80 bpm); blood pressure 80/40 mmHg; respiratory rate 18 bpm; blood glucose 13.5 mmol/l. Admission venous blood gas showed a metabolic acidosis with respiratory compensation (see table 1) and an acute kidney injury.

Figure 1.

Proposed treatment algorithm.¹.

Table 1.

Chronological blood gases.

Time	14:35	15:23	17:24	05:30
pH	7.349	7.338	7.308	6.98
PCO2	4.17	4.25	3.14	3.1
Base Excess	–7.7	-8.0	–13.6	–24
Lactate	7.9	8.6	9.9	>17
Glucose	10	7.1	2.6	10
HCO3−				5.6

Initial management included fluid resuscitation but despite this the patient remained persistently hypotensive with a worsening metabolic acidosis. Invasive arterial monitoring was inserted and a metraminol infusion commenced. Toxbase advice included the administration of 20 mls of 10% Calcium chloride over 20 minutes aiming to keep the ionised calcium above 2.0 mmol/l. In the ICU the patient remained hypotensive with a worsening metabolic acidosis, oliguria and increased respiratory effort. Central venous access was obtained and a dialysis catheter inserted. Isotonic sodium bicarbonate was given and a noradrenaline infusion was commenced. Further boluses of calcium chloride were administered. Continuous veno-venous haemodiafiltration (CVVHDF) was started. During the first 12 hours of ICU admission the noradrenaline infusion escalated to 1 mcg/kg/min and vasopressin was added at 2.4 units/hour with little effect on blood pressure; The day following admission the patient was intubated and ventilated for fluid overload and worsening respiratory distress. High dose intravenous insulin therapy (10 international units (IU)/Kg/Hour) with 50 ml/hour 50% dextrose was commenced. The insulin infusion rate peaked at a maximum at 900 IU per hour.

Over the next 24 hours the patient made a steady improvement in condition. Vasopressor requirements dropped and metabolic acidosis resolved. The patient was extubated 48 hours after admission and made a full recovery.

Calcium channel antagonists bind the alpha subunit of the L-type voltage gated calcium channel and prevent intracellular in flux of calcium. Refractory hypotension is mediated according to subtype of calcium channel blocker. Non-dihydropyridine sub-type (Verapamil/Diltiazem) impair myocardial contractility whereas dihydropyridine sub-type (Amlodipine/Nifedipine) cause vasodilation and vasoplegia. In severe overdose there appears to be cross over in effects of the individual subtypes. High dose insulin eugylcaemic therapy (HIET) has been recommended for treatment of severe cases of calcium channel blocker overdose.¹ The exact mechanism of action remains unclear but it is hypothesised that HIET promotes excitation–contraction coupling and contractility because increased glucose availability results in: Increased sarcoplasmic reticulum-associated calcium ATPase activity, increased cytoplasmic calcium concentrations, and enhanced calcium entrance into mitochondria and sarcolemma. Expert consensus recommendations now endorse the use of HIET as a first line treatment for cases of severe calcium channel blocker overdose and protocols for the institution of such therapy are widely available. (See figure 2)

The value of a rehabilitation programme following critical illness

Suzahn Wilson ^1,2

¹University of East Anglia, Norwich, UK

²West Suffolk NHS Foundation Trust, Bury St Edmunds, UK

Abstract

Objective: To evaluate a six-week physical rehabilitation service for critical care survivors in a district general hospital and describe the changes in physical, psychological and general health states of participants.

Background: The physical and psychological effects of recovering from critical illness are well documented with discharge from critical care and hospital being the start of an uncertain journey to recovery for many. Rehabilitation strategies vary nationally with many patients not routinely referred for follow up. More patients are surviving critical illness due to advances in medicine and technology, indicating a growing need for robust, evidence-based, patient centred rehabilitation strategies.

Methods: This service evaluation used a mixed-methods design with existing auditable outcome measures matched with interview data to give context to findings. Existing audit data was collected over 2 years with wider admission and referral data giving context to findings. 787 patients survived ICU in this period, with 55 patients referred to the programme and 16 patients completing the full six-weeks. Interviews were conducted with critical care survivors (n = 8), transcribed, and analysed using a grounded theory methodology.

Findings: Critical care survivors struggled for independence and described the physical, psychological, and social functioning challenges associated with this. Significant improvements were observed in physical function, anxiety, depression, and most of the SF-36 subsections after the six-week programme compared to before. Only 7% (55/787) of critical care survivors were referred to the rehabilitation service over 2 years, of which 29% (16/55) completed the full six-week course. There are opportunities to target behaviour change and improve engagement with physical rehabilitation.

Conclusions: This evaluation offers novel insight into the complex challenges faced by patients during recovery and explored how these may impact capability, opportunity, and motivation to engage with physical rehabilitation. The main concern for participants in this evaluation emerged as their loss of independence, which was multifaceted and posed a range of physical, social, and psychological challenges. Recruitment, accessibility and delivery of information, and using appropriate outcome measures with the rehabilitation programme all warrant collaboration with stakeholders and further research to address.

Long term compliance with a simple initiative to assist antimicrobial prescribing in intensive care: are standards maintained 5-years on?

Timothy Jones¹, Christopher Jones² and Christopher MacIsaac³

¹Northwick Park Hospital, London, UK

²King's College London, London, UK

³Intensive Care Unit Royal Melbourne Hospital, Melbourne, Australia

Abstract

Background: Appropriate prescription of antimicrobial agents is essential in all clinical environments (Leuthner and Doern, 2013) as sub-optimal antimicrobial usage is linked to the development of resistant strains. Antimicrobial stewardship in the ICU provides a particular set of barriers to appropriate prescribing, primarily related to the fact that patients are critically unwell and hence the underlying temptation is to prolong or escalate therapy, even in the absence of clear markers of sepsis (Luyt et al., 2014).

Although countless quality improvement projects are successful in the short-term, many fail to demonstrate sustainability in the long run for which they are designed. As such, efficacious and sustainable antimicrobial stewardship programmes are of vital importance the ICU.

Aim: This study aimed to re-assess the performance of a simple intervention to assist with appropriate antimicrobial prescribing established five years previously in a tertiary ICU.

Methods: We performed two snapshot audits (one week apart) of all patients receiving one or more antimicrobials while in the ICU, approximately five years after a formal antimicrobial initiative had been introduced. This initiative was shown to remain effective one year post implementation (Chaves et al., 2014). This intervention involved the introduction of a daily self-review tool in the form of a yellow-sticker to be used in the drug-chart as an aide-memoire for documenting three minimum standards of anti-microbial prescribing: start date, stop date and indication. It also involved the development of a new set of guidelines for the empirical management of sepsis in ICU and training sessions for prescribers. The evaluation aimed to assess the documentation of the three aforementioned minimum standards, the usage of the yellow sticker and the overall percentage of prescriptions that were concordant with the trust guidelines. Results were compared to data collected before the stewardship programme was introduced and data was analysed using the Chi-squared test for proportions.

Results: A total of 52 patients with 79 antimicrobial prescriptions were compared to the pre-intervention cohort of 260 patients with 306 prescriptions. 71 (90%) prescriptions were found to be utilising the yellow-sticker system. When compared to the previous audit immediately before implementation of the sticker initiative, documentation of start date improved significantly (2013 v 2018, 72% vs 86%, p = 0.01) as did documentation of indication (2013 v 2018, 58% vs 75%, p = 0.005). However, there was no improvement in documentation of stop/review date (2013 v 2018, 16% vs 10%, p = 0.4). Overall there was a significant improvement in the number of prescriptions deemed concordant with local guidelines (2013 v 2018, 74% vs 90%, p = 0.002).

Conclusions: We have demonstrated that the introduction of a simple intervention for ICU antimicrobial prescribing is effective 5 years after introduction for documentation of the start date and indications for antimicrobials but not the cease/review date.

Increasing delirium screening – the battle continues

Catherine McGow, Andrew Workman and Prashast Prashast

Wirral University Teaching Hospital NHS Foundation Trust, Liverpool, UK

Abstract

Delirium has been defined as “an acute, reversible organic mental syndrome with disorders of attention and cognitive function, increased or decreased psychomotor activity and a disordered sleep-wake cycle”.

The management of delirium is an important and challenging aspect of therapy when treating critically ill patients. The incidence has been reported in the range of 15–80%. One UK study showed that patients with delirium in an ICU had a significantly higher mortality during admission and a longer length of stay compared to those without delirium. It also showed that delirium may predispose patients to prolonged neuropsychological disturbances after they leave ICU.

There are a number of different validated scoring systems to use in critically ill patients. CAM-ICU is one of the most common. The aim of this quality improvement project was to review adherence to completing the CAM-ICU score once per 12 hour shift. An additional aim was to explore reasons for lack of compliance.

Case notes of patients were reviewed over a two week period to check whether CAM-ICU scores had been recorded in each 12 hour shift. A questionnaire was devised to assess understanding of the CAM-ICU score and identify reasons for non completion. This was administered to all nurses on the ICU.

Initial data on compliance showed that 68% of patients with risk factors for delirium did not have a CAM-ICU score completed. There was a very poor response to the survey. However, it was our view that this work would itself raise awareness and facilitate staff discussion on ways to improve compliance. We were reluctant to introduce visual reminders – being very aware that the unit was already covered in many posters.

A subsequent review of case notes identified a further drop in compliance results with up to 90% of patients having no record of delirium screening in their case notes. On reflection, it was clear that encouragement alone was insufficient to improve practice. At this point,we recruited a member of the nursing staff to work with us to raise awareness and also to design a poster as an aide-memoire.

A further review of case notes showed that 73% of patients with risk factors had no CAM score undertaken. We did not anticipate that implementing the scoring system would be so challenging. There are a number of potential explanations for this including: ICU patients are by nature complex meaning that staff already undertake many checklists and scores; there is a lack of knowledge and understanding of the negative impact of delirium on patients and a further tool increases burden on an already stretched workforce.

In conclusion, we adopted simple interventions in the hope of creating change without increasing the burden on staff. However, this was not successful and,ultimately, further trials of different interventions will be required to ensure that delirium screening is done routinely.

The influence of staff attitudes and unit organisation in the delivery of lung protective ventilation (LPV)

Inae Hwang¹, Antony Thomas², Richard McGuire³ and Vinayak Vanjari²

¹University of Manchester, Manchester, UK

²Greater Manchester Critical Care Network, Manchester, UK

³Manchester University Foundation Trust, Manchester, UK

Abstract

How patients are ventilated will affect their chances of survival [1]. How staff adjust ventilator settings will be affected by their understanding and beliefs about ventilation and how they register and follow instructions from consultant staff.

We surveyed staff in a single unit (unit one), to establish their views about what was important in the delivery of ventilation and how they felt instructions were provided by consultants.

The survey was conducted using an on-line survey (https://www.surveymonkey.co.uk/r/3YJVK6W) that staff completed while being interviewed by a medical student who clarified any ambiguity in the survey and recorded additional comments made by staff. During the study we audited the prescription and delivery of mechanical ventilation using repeated observations; we compared these observations in a second unit (unit two) that had already carried out improvement work on the delivery of lung protective ventilation. Both units were based in large teaching hospitals in Greater Manchester. We then introduced a new white board behind each bed in unit one to record the patient’s ideal body weight.

Thirty-three staff were interviewed (16 nurses, 14 doctors, 3 Physiotherapists). The median importance rating (with IQR) for survival for different factors, with a maximum score of 100 were: How patients with ARDS are ventilated, control of TV and peak pressure all: 80 (90–70), Adjusting oxygen to saturations: 90 (100–80), PEEP ladders: 75 (90–40). With respect to what staff though actually happened on unit one, the scores for agreement with the following statements were: Ventilation is prescribed: 55 (75–25), PEEP ladders are used: 35(55–25), IBW is calculated on admission: 85 (95–45). Most staff (30 responses) felt that inspired oxygen was adjusted by all staff and that all medical staff and nurses above band 5 adjusted other ventilator settings.

Observing what actually happened, Tidal volumes (TV) between 6 to 8 ml/kg was found in 27% of observations in unit one and 67% in unit two (p = 0.01). TVs were prescribed in 4% of observations in unit one and 92% in unit two. Unit one had no significant relationship between FiO2 and PEEP level. The IBW was recorded in 6% of observations before the white board intervention, increasing to 95% after (p < 0.001), this resulted in no significant change in TV within range (27% before to 20% after).

Even when staff believe ventilatory patterns are important for survival and that their unit has ways of delivering this, without set ways of prescribing modes of ventilation and recording ideal body weight these beliefs are not translated into action. LPV requires policies of prescribing modes of ventilation and auditing what is actually delivered to patients.

A Service Evaluation on the Usefulness of Procalcitonin as a Biomarker in the Critically Ill Patient

Berenice Cunningham-Walker¹ and Babu Muthuswamy²

¹Cardiff University, Cardiff, UK

²Royal Gwent Hospital, Newport, UK

Abstract

Background: Diagnosis and management of systemic bacterial infection is undergoing continued scrutiny in critical care due to the lack of a specific diagnostic biomarker. Procalcitonin is an inflammatory biomarker which has been suggested as a complementary investigation in diagnosis of bacterial infection and to guide initiation and cessation of antibiotic therapy.

Objectives: The aim of this evaluation was to see whether the use of PCT can decrease time receiving antimicrobial therapy in a population of adults in critical care. At the Royal Gwent Hospital in Newport, the Procalcitonin Guided Antimicrobial Management recommendation has been in place since 2015, which gives a PCT cut-off algorithm for use in adult intensive care.

Methods: Evaluation of PCT in 54 adult participants in critical care with suspected infection was carried out over a month. The lead consultants on critical care were asked bi-weekly whether the PCT results influenced individualised management plans at any one time during each patient’s stay. Antibiotic therapy for two weeks prior to, and across this time frame was monitored individually; non-infectious antibiotic therapy was excluded.

Results: 54 patients were included in the evaluation, with a mean age of 64.1 years and median time spent on ICU of 3 days. 42 of the 54 patients had PCT values taken, with 14 (33.3%) of these patients having their PCT values considered during their management. The most common utilisation of PCT was continuation of antibiotic management following an elevated PCT value; the second commonest decision made was to stop antibiotic therapy following a lowered PCT value. 12 patients across this time scale did not have a PCT value taken which meant that its use could not be evaluated across the entire cohort.

Conclusions: The introduction of a PCT guided algorithm in critical care has benefited consultant decision making by strengthening the choice to continue antibiotic treatment, or where there’s a lowered PCT value to encourage cessation of antibiotics.

Hypomagnesaemia on admission to the Intensive Care Unit: A retrospective review

James Schneider¹ and Callum Kaye²

¹Critical Care Department, University College London Hospital, London, UK

²Intensive Care Unit, Aberdeen Royal Infirmary, Aberdeen, UK

Abstract

Background: Hypomagnesaemia is known to be a common sequela of critical illness. Magnesium plays a role in the regulation of multiple processes affected by sepsis, thus, it may have a role in the management of sepsis. However, limited evidence exists focusing on the incidence and significance of the duration of hypomagnesaemia in septic patients.

Aims and Methods: This project aimed to determine the association between the development and duration of hypomagnesaemia with clinical outcomes in septic patients admitted to the ICU. The study population was generated using local SICSAG admission data, selecting all patients admitted to the ICU at Aberdeen Royal Infirmary between 01/04/2013–01/04/2018 with Primary Unit Diagnosis of ‘Septic Shock’. The study population was dichotomized into two independent subgroups based on admission serum magnesium: normomagnesaemia (0.7–1.0 mmol/L) and hypomagnesaemia (<0.7 mmol/L). Statistical analysis compared APACHE II Scores, duration of admission, need for mechanical ventilation, vasoactive support, readmission, and ICU mortality between the populations. Further subgroup analysis was performed to determine whether duration of hypomagnesaemia (<12 hours and >=12 hours) was a dependent factor for poor clinical outcomes. All statistical analysis was deemed significant to p < 0.05.

Results: The study population consisted of 537 patients who were admitted to the ICU with sepsis during the 5-year study period. Of these, 197 (36.7%) and 258 (40%) patients were admitted hypomagnesaemic and normomagnesaemic, respectively. In total, 43.9% of patients developed hypomagnesaemia at some stage in the ICU. APACHE II Score and duration of ICU stay did not differ between the admission serum magnesium groups, p > 0.05. With regards to clinical course, there was a reduction in need for mechanical ventilation in those with hypomagnesaemia, but no association with ICU mortality or vasoactive use. For those who had hypomagnesaemia corrected within 12 hours, there was a reduction in days mechanically ventilated & halving of 7-day mortality.

Conclusions: Hypomagnesaemia is a common presenting feature in septic patients admitted to the ICU. The severity of illness, duration of treatment, requirement for organ support, and ICU mortality between admission hypomagnesaemia and normomagnesaemia patients were not significantly different. However, correcting hypomagnesaemia <12 hours following admission resulted in significantly better 7-day survival. Findings from this study suggest that treating hypomagnesaemia aggressively may improve early and overall outcomes in septic patients.

Critical Saturation? A qualitative service evaluation of LocSSIPs checklist compliance in a critical care unit

Heather Reading^1,2 and Daniel Higgins^2,3

¹South Warwickshire NHS Foundation Trust, Warwick, UK

²University of Warwick, Coventry, UK

³University Hospitals Birmingham NHS Foundation Trust, Birmingham, UK

Abstract

Patient safety has received increasing attention in the UK during the past decade following several high-profile cases revealing avoidable harm.¹ Checklists materialise as one solution in tackling preventable errors.² Invasive procedures shape treatments in critical care necessitating application of the recently recommended Local Safety Standards for Invasive Procedures (LocSSIPs) checklists.³ Local post-implementation audits revealed poor (25%) initial uptake.

The aim of this study was to explore local multi-disciplinary team (MDT) perspectives on variant compliance of recently implemented LocSSIPs checklists.

This study utilised qualitative inquiry in a single centre via semi-structured interviews. Purposive stratified sampling was employed to include local MDT participants involved in invasive procedures from varying professional disciplines, background and critical care experience. Ten interviews were conducted (n = 10).

Content and thematic analysis identified three main themes and six interconnecting sub-themes. The main themes were: (1) Human Factors: LocSSIPs may challenge teamwork, communication and culture in critical care (2) Governance: LocSSIPs provide formal safety measures to invasive procedures, which may be perceived paradoxically by the MDT and (3) Tool Design: The prevalence and length of checklists throughout critical care raises duplication concerns and may lead to fatigue undermining the safety focus.^4,5

Personal perspectives were offered surrounding potential barriers to compliance pertaining to LocSSIPs checklists in critical care. The findings suggest ongoing modification of LocSSIPs are required together with educational opportunities in honing tool design and enhancing compliance applicable to critical care procedure, environment and personnel concerned.

There is a need for cyclic PDSA activity in future adaptations of critical care LocSSIPs. Recurrent scenario-based simulation is proposed in providing a safe setting combining checklist scrutiny and training.

Nil external funding, MSc Professional Project

This study received ethics committee approval or equivalent

Keywords: critical care, checklists, compliance, patient safety, LocSSIPs

An Audit of Ward Round Interruptions in the Regional Intensive Care Unit

Jordan McVey, Stephen McKenzie and Dominic Trainor

Belfast Health and Social Care Trust, Belfast, UK

Abstract

Background: Ward rounds play an essential role in the management of patients in the critical care environment. They are a forum for discussion with a multitude of health professionals and a vehicle for often complex decision making. Interruptions occur frequently in the critical care environment are well-recognised as having a negative effect on task performance. In an emergency situation, some interruptions may be necessary and appropriate, however this audit was devised in order to reduce the number of unnecessary interruptions.

Method: Data was collected over a two-week period during normal working hours and excluding the weekend. The ward rounds included were the morning handover round, consultant ward round, microbiology ward round and evening handover round. All interruptions were noted along with the degree of urgency, the reason for the interruption, the staff group responsible and the ward round affected. Interruptions were defined as a secondary activity unrelated to the primary task requiring diversion of attention. The data from the first phase of data collection was presented at a local audit meeting and a newsletter was distributed to nursing staff requesting interruptions be kept to a minimum. Signs were erected on the mobile workstations stating that interruptions must be kept to emergencies only. The second phase of data collection was then performed over the same timeframe and using the same parameters.

Results: During the first phase of data collection there were a total of 40 interruptions recorded, 7 of which required immediate action (17.5%). Forty-seven percent of all ward rounds were interrupted on at least one occasion. The most common reason for interruptions were queries relating to medications, accounting for 25% of the total, with 22.5% relating to referrals and admissions and 17.5% due to investigation findings. ICU nursing staff were responsible for 72.5% of all interruptions, and non-ICU medical staff account for 20%. Of the interrupted ward rounds the consultant ward round was most commonly affected at 42%.

During the second phase of data collection there were a total of 15 interruptions recorded, 2 of which required immediate action (13.3%). This represents a 62.5% decrease in the total number of interruptions and 20% of all ward rounds were interrupted on at least one occasion. ICU nursing staff were responsible for 46.6% of all interruptions and 33.3% came from non-ICU medical staff. Medication related queries (26.6%) and interruptions relating to referrals or admissions (40%) remained the most common. The consultant ward round was the most commonly affected, at 93.3% of all interrupted ward rounds.

Conclusion: Interruptions are well-described as a source of error and inefficiency in the critical care environment. While some interruptions are necessary to divert attention to tasks requiring immediate action, many are non-urgent and could be dealt with at a more appropriate time. It has been shown here that the majority of interruptions in the ICU are non-urgent and these can be reduced by creating a culture within the unit to avoid interrupting ward rounds unless necessary.

A case of survival after lethal dose Sodium Nitrite ingestion

Mary Madden, Therese McKeith and Gareth Morrison

Anaesthetics Department, Antrim Area Hospital, Antrim, UK

Abstract

Background: Sodium Nitrite (NaNO₂) has widespread use worldwide predominately as a food preservative and anti-microbial agent. Poisoning from the substance has occurred accidentally however there are now increasing reports of intentional ingestion.¹ Death usually occurs due to precipitation of Methaemoglobinaemia with resultant tissue hypoxia. Here we present a case of survival after toxic dose ingestion.

Case: A 22 year old man with a history of depression and autism arrived in the Emergency Department after collapsing at home following ingestion of 30grams of 99.5% Sodium Nitrite purchased online. Upon presentation the patient was comatose with myoclonic seizure activity and significant haemodynamic compromise. There was extensive blue grey skin discolouration and unrecordable oxygen saturations. An initial blood gas confirmed a severe metabolic acidosis with hypoxaemia and methaemoglobinaemia of >70%. Emergent management included fiberoptic intubation and DC cardioversion. Methylthioninium Chloride 250 mg (2.5 mg.kg⁻¹) was administered with a quick resolution of cyanosis alongside intravenous crystalloid, Magnesium Sulphate and Sodium Bicarbonate. After stabilisation the patient was transferred to the Intensive Care Unit and remained there for 58 days. Significant mucosal damage was sustained to both the airways and gastrointestinal tract requiring an emergency subtotal colectomy and further resection of non-viable small bowel. Pathology confirmed extensive mucosal ischaemic necrosis and ulceration. Throughout his ICU stay, the patient also displayed a persistent tachycardia, intermittent vasopressor requirements and hyperpyrexia requiring Dantrolene therapy. Despite his complicated course the patient was discharged to ward level with evidence of a good neurological recovery for ongoing physical rehabilitation, nutritional management of Short Bowel Syndrome and psychological support.

Discussion: Nitrite poisoning and methaemoglobinaemia are rare. The World Health Organisation has indicated a lethal dose of 33–250 mg.kg⁻¹ for oral nitrites³ and survival has been seen with ingestion of up to 15g.¹ This is the first case to report survival with 30 g Sodium Nitrite.

Initial toxic effects of nitrites are due to its effects as an oxidising agent and potent vasodilator. Nitrite oxidises ferrous iron (Fe²⁺) to its ferric state (Fe³⁺) forming Methaemoglobin. Ferric iron loses its ability to bind and transport oxygen. Methaemoglobin levels of >70% cause lethal tissue hypoxia.³ Methylthioninium Chloride is first line in the treatment of methaemoglobinaemia with a recommended dose of 1–2 mg.kg⁻¹.⁴ Its action is primarily as a reducing agent, dependent on the availability of reductase cofactors within erythrocytes.³

We believe this case is also the first to highlight longer-term sequelae of the toxicity of Sodium Nitrite. Although nitrite induced intestinal mucosal damage has been reported in preliminary animal studies⁵, the severity of mucosal damage to the gastrointestinal tract was an unprecedended finding to the pathology department, with resultant significant morbidity for the patient.

Lastly, the ease of discussion and accessibility of such a devastating agent on online platforms is alarming and causes concern for an increase in popularity of its use.

Conclusion: Intoxication with large amounts of Sodium Nitrite is rare and therefore information on clinical consequences is limited. This case illustrates the toxicity and subsequent morbidity associated with a lethal dose ingestion of Sodium Nitrite.

Individualized treatment strategies for patients with sub-massive pulmonary embolus: a case series from a district general hospital

Teddy Tun Win HLA¹, Colin Macleod², Gihan Hettiarachchi³ and Rahul Sarkar²

¹Intensive Care Unit, Royal Sussex County Hospital, Brighton, UK

²Critical Care Department, Medway Maritime Hospital, Gillingham, Kent, UK

³Respiratory Medicine, Medway Maritime Hospital, Gillingham, Kent, UK

Abstract

Pulmonary embolism (PE) is a common condition with an annual incidence of 60–70 cases per hundred thousand population in the West. However, the clinical course is heterogeneous with highly variable mortality and morbidity[1]. Clinical decision making for acute pulmonary embolism is clear for massive/high-risk and non-massive/low-risk pulmonary embolism -i.e., with/without haemodynamic instability. Both British Thoracic Society (BTS) guidelines and European Society of Cardiology (ESC) PE guidelines recommend systemic thrombolytic therapy as first line for massive PE, and anticoagulation with low-molecular weight heparin or direct oral anticoagulation agents as first line for non-massive PE[1,2]. Sub-Massive Pulmonary Embolism (SMPE) is a loosely defined term that defines an intermediate-risk category between these extremes that encompasses PE without overt haemodynamic instability but with evidence of right ventricular dysfunction on echocardiography or blood biomarkers (troponin/Brain natriuretic peptide). SMPE has significant mortality and morbidity, with a subset progressing to massive/high-risk PE with haemodynamic decompensation, at a median of 1.8 days after presentation[3,4]. The definitive management of patients with SMPEs remains uncertain.

BTS guideline recommend anticoagulation without systemic thrombolytic therapy for SMPE [1]. However, in light of the recent PEITHO trial results, half-dose systemic thrombolysis or catheter-directed, ultrasound-assisted, thrombolysis (CDT) using lower doses of a thrombolytic agent is recommended in ESC guidelines. ESC also recommend considering contraindications to thrombolysis as ‘relative’ to be considered case-by-case.

At our large district hospital in Kent, we have on-site, 24/7, interventional radiology available for catheter-directed thrombolysis, a treatment modality that showed equal efficacy at lower dose thrombolytic agent in the recent OPTALYSE PE [5]. Given the availability of this treatment modality locally, we believe we are in a position to offer better care for our patients with SMPE.

To illustrate our experience with SMPEs in our critical care department, we present a case series of six patients admitted from 23^rdFebruary-3^rdMarch 2018. One patient received systemic thrombolysis (recombinant tissue plasminogen activator rTPA), 3 patients were only anticoagulated (dalteparin or rivaroxaban) but not thrombolysed, 2 received CDT (one required rescue systemic thrombolysis and one had a complication of retroperitoneal haematoma). A brief summary of our sample group is tabulated below.

Patient	Age	Blood pressure/ mmHg	Heart rate/ bpm	SpO2/%	FiO2/%	Echo features	Troponin/ ng/L(0–40)	BNP/pg/ ml(0–99)	Modified PESI score	Treatment strategy
A	38	138/96	107	98%	24%	Dilated  RV	211	243	67	Anticoagulation
B	86	119/88	116	99%	21%	“	1749	434	115	Anticoagulation
C	74	99/66	105	88%	28%	“	526	259	193	Anticoagulation +  CDT + rescue  systemic  thrombolysis
D	72	114/69	110	97%	80%	“	2617	117	132	Anticoagulation +  half dose rescue  systemic  thrombolysis
E	86	76/48	109	65%	N/A	“	4287	434	145	Anticoagulation +  CDT
F	78	150/116	101	91%	21%	“	108	449	108	Anticoagulation

Whilst definitive evidence for management of SMPE is yet to emerge, we believe our ability to offer individualized treatment strategy to patients though a multimodality approach is beneficial for patients in terms of short and long term outcome.

Do Fenestrated Tracheostomy tubes still have a place in ICU? Outcomes from an audit conducted in two intensive care units at a specialised tertiary centre

Neresha Maistry and Paulo De Sousa

Royal Brompton and Harefield NHS Foundation Trust, London, UK

Abstract

Introduction: Fenestrated tracheostomy tubes are considered by some to be a routine method of facilitating respiratory weaning from a ventilator and enabling phonation.¹ Standard practice suggests that non fenestrated tracheostomy tubes are initially sited and replaced with fenestrated tubes after seven to ten days to avoid subcutaneous emphysema and to assist with weaning and that speaking valves should ideally only be used with an un-cuffed and fenestrated tubes.² Both a tracheostomy change itself and fenestrated tubes have associated risks.³

Objectives: The aim of this audit was to assess whether fenestrated or non-fenestrated tracheostomy tubes impact Speech and Language Therapy outcomes, by comparing two cardiothoracic intensive care units (ICU’s). One unit (“unit B”) utilises fenestrated tracheostomy tubes (FTT) routinely, to facilitate weaning, whilst the other unit (“unit A”) only uses non-fenestrated (NTT) tubes from acute through weaning stages.

Methods: We have conducted a retrospective audit over five months. Data was collected from patients electronic records. Categorical data was summarised as frequencies (%) and continuous data as median (IQR) or mean (SD) as appropriate; groups were compared using the Wilcoxon’s rank-sum (Mann-Whitney) test.

Results: From December 2017 to April 2018, a total of 44 patients were tracheostomised, of which 25 were male (56%) with a mean age of 61 years (14). NTT group was predominantly female (62%) while the FTT one was predominantly male (74%). There was no age difference between groups (mean age of 61(14) years compared with 61(15)). The non-fenestrated group took less time to phonation, compared with the fenestrated group [6 (3–11) vs 10 (6–38); (p = 0.003)] and had lower median time in days to oral intake [14 (4–21) vs 18 (12–52); (p = 0.029)]. There was no significant statistical difference regarding time to decannulation [20 (13–25) vs 21 (16–81); (p = 0.098)].

Conclusion: The results from our audit suggest that non fenestrated tubes may be more beneficial than fenestrated tubes, in terms of time to phonation and time to oral intake. There may therefore be quality of life as well cost implications.

External transfer of invasively ventilated patients by Advanced Critical Care Practitioners

Gavin Denton

University Hospitals Birmingham NHS Foundation Trust, Birmingham, UK

Abstract

The Heartlands hospital ACCP team is integral to the provision of critical care services across several hospitals and has transferred mechanically ventilated patients without direct medical supervision for several years. This audit considers the activity and safety profile of ACCP external transfer of mechanically ventilated patients.

A web based anonymised electronic form was devised. Data were submitted shortly after transfer of mechanically ventilated critical care patients. The aim was to attempt to capture all internal and external transfers, with a focus on invasively ventilated patients.

Between December 2016 and June 2018, 131 external transfers of mechanically ventilated patients were identified by the audit. 93% (n = 123) of these events occurred on two of three sites, where ACCPs provide a dedicated transfer role. A total of 125 transfers were provided by ACCPs without direct medical supervision (95.4%). Lack of a critical care bed in the primary hospital was the chief reason for transfer (81.1%, n = 69). The second most common reason for transfer was for specialist care in a tertiary hospital (27.2%, n = 34), this was mainly for neurosurgical intervention. Transfer for specialist treatment within our trust consisted of 8% cases (n = 10). A further 12 cases fell into a miscellaneous group.

Intropic support was required by 68% of patients (n = 85), with 4% needing more than one inotrope (n = 5). Arrhythmia was evident in 4.8% of cases. Two patients were in receipt of Intra-arortic balloon pump support (n = 2, 1.6%). In terms of the acuity of ventilatory support, 52% received more than 50% oxygen (n = 65), and 12.8% needed a PEEP greater than 10 cmH2O, 4 patient required airway pressure release ventilation.

In 88.8% of transfers, no adverse events occurred. Of the 15 adverse events, five were related to logistical issues with the ambulance service (delay in attendance, refusal to repatriate equipment). Hypotension was the most common clinical adverse event with an incidence of 2.4% (n = 3). There were three equipment failures (2.4%). There were no extubations or airway events.

The NICE recently evaluated internal and external hospital transfers1. The type and frequency of complications varied between studies due to differences in study design and definition of complications (10–24%)1. Complications associated with transfers can be broadly categorised in to those relating to patient factors (severity of illness, comorbidities, interventions), system factors (transfer distance, equipment failure, absence of checklists) and staff factors (communication, insufficient training, inadequate resuscitation, unsecured devices)2,3. NICE reported very low quality evidence supporting the use of specialist transfer teams and checklists, two systems already adopted as part of the ACCP led transfer service1. This may in part explain the relatively low rate of adverse events (<10%) reported in this audit.

Within a accredited training curriculum, governance structure and consultant oversight, ACCPs can safely and effectively transfer mechanically ventilated critically ill patients externally. Further study is required to show how ACCPs perform on a national level and to provide benchmarking of standards and identify best practice.

Changing the Culture in ICU – Introduction of a Framework for Quality Improvement

Clare Brennan, Stephanie Frearson

NHS Ayrshire and Arran, Kilmarnock, UK

Abstract

Background: According to the Health Foundation (2013) improving quality is about making healthcare safe, effective, patient-centred, timely, efficient and equitable. (1) Within our Intensive Care Unit (ICU) we felt there was a compelling case to introduce quality improvement methodology as a framework for change. We aimed to make quality improvement part of our ‘daily core business’ in Crosshouse ICU.

Objectives: ICU is dedicated to supporting the delivery of the Quality Strategy (2011) through improvements in clinical practice. (2) Our objective was to develop an ongoing framework for improvement, with full staff engagement, that would allow a systematic approach to implementing change and measuring progress. By doing this we hoped to:

• Optimise patient care

• Optimise staff experience

• Change culture

• Continuous Clinical Improvement

Methodology:

Initially we asked staff to complete a questionnaire to identify learning needs. The results of this allowed us to plan our improvement journey which included:

• Setting up MDT Improvement Group

• Introducing ICU Yearly Improvement Plan

• Introducing bi-annual ‘away days’

• Dedicated improvement advisor support

• Involving ALL staff

• Disseminating our work

Thereafter using Improvement Methodology/Model for Improvement we devised a driver diagram defining our improvement plans

Results/Outcomes: It is the belief that we have developed a framework whereby staff have a better understanding of the benefits of a quality improvement approach. This has included:

• Culture change – made improvement part of the daily ICU business

• Suggestions for improvement now come from a variety of sources and are no longer seen as ‘top-down’

• Raised the profile of the unit

• Embraced a collaborative ‘MDT’ approach

• Changed recruitment process

• Financially – achievable

This has been reflected anecdotally in our i-matters survey and interestingly early data is suggestive of improvement in staff sickness levels.

Discussion: Moving forward our plan is to increase staff participation in new QI projects and encourage a continous, ongoing quality improvement culture in ICU. Furthermore we plan to look at the benefits/improvements these changes may have on actual clinical outcome measures for ICU patients such as average length of stay and ventilator days.

The Introduction of a Personal Care Plan for last hours and days of life within ICU

Catherine Morison, Diane Bowler, Alison Spiers and Stephanie Frearson

NHS Ayrshire and Arran, Kilmarnock, UK

Abstract

Background: Despite end-of-life (EOL) care being recognised as an important component of intensive care there appears to be compelling evidence of the need to improve standards of care provided. Locally it was acknowledged that a structured multi-disciplinary approach was required to improve the care offered and delivered to patients and families.

Aims/Objectives:

The aims/objectives of our improvement project were:

• Phase 1 Staff learning needs/education

• Phase 2 Development/testing combined nursing and medical care plan

• Phase 3 Measurement plan in order to collect data.

• Phase 4 Offer opportunity for family involvement in patient care before and after death

Our project supports the Scottish Government vision and our journey towards 2020 and beyond by allowing us to provide care to the highest standards of quality and safety, with the person and their families at the centre.

Methodology: Our Quality Improvement project aimed to develop a multidisciplinary holistic care plan to ensure patients who die within the ICU do so with dignity. Using the Model for Improvement we devised a driver diagram and employed PDSA methodology to test ideas for change.

Results/Outcomes:

Discussion: The last hours and days of life care plan was implemented within our ICU in Oct 2016. To promote relationship centred care, we are currently offering families the opportunity to be involved with the care of their loved ones before and after death. The next steps in this project will be capturing relative feedback and development of a follow up service to ensure relative satisfaction and engagement

New-onset atrial fibrillation is independently associated with increased ICU length of stay and mortality

Jonathan Bedford, Robert Hatch, Duncan Young and Peter Watkinson

Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK

Abstract

Background: Atrial fibrillation (AF) is a common arrhythmia in critically ill patients (1). Observational data suggest new-onset AF (NOAF) occurs in 4.5–11% of all patients admitted to the intensive care unit (ICU) rising to 23% of those patients with septic shock (2,3).

New onset AF in critically ill patients is associated with increased length of both ICU and hospital stay (4). It is also associated with increased mortality (5), though whether it is an independent risk factor or merely a marker of disease severity is uncertain.

Materials and methods: We performed a retrospective analysis of patient data from the PICRAM (Post Intensive Care Risk-adjusted Alerting and Monitoring) database. This contains data associated with over 14,000 adult patients admitted to three general adult intensive care units in Oxford and Reading from 2006 to 2015. We identified adult (aged 16 years or above) patients who were admitted for the first time to one of the three general intensive care units in the PICRAM database. We excluded patients with a history of AF and those with a documented arrhythmia within the first four hours of admission.

We adjusted for admission covariates to investigate the association between NOAF and ICU length of stay and ICU mortality. Using stepwise backward regression we identified 7 independent predictors of in-hospital mortality from our dataset (sex, age, admission type, OASIS score, vasopressor use, and urea and lactate on admission). The OASIS score consists of seven physiological measurements, elective surgery, age, and pre-ICU length of stay. Physiological measurements were therefore not included as individual covariates.

Results: Of the 8012 patients that fulfilled our inclusion criteria, 948 (11.8%) developed NOAF during their admission.

Impact on Length of stay: Multiple linear regression analysis revealed that after correction for confounding variables, new-onset AF significantly prolonged length of stay (β = 1.81 (95% CI 1.71–1.92, p < 0.001)).

Impact on mortality: We analysed the impact of NOAF on ICU mortality. Crude ICU mortality was increased in the NOAF cohort (26% vs 10%, p < 0.001).

We adjusted for 7 variables associated with ICU mortality using multivariate logistic regression. New-onset AF was independently associated with increased ICU mortality (adjusted OR 1.47 (95% CI 1.21–1.78), p < 0.001). Analysis of the cause-specific hazard ratios (CSHR) revealed that NOAF both reduces the daily chance of being discharged from the ICU (CSHR 0.86 (95% CI 0.79–0.94), p < 0.001) and increases the daily chance of dying on the ICU (CSHR 1.33 (95% CI 1.13–1.57), p < 0.001).

Conclusions: This study adds to recent growing evidence that new-onset atrial AF in critically ill patients independently contributes to increased ICU length of stay and ICU mortality. We showed that NOAF has a compound effect by prolonging exposure to an increased daily risk of death. These results highlight the need for further research into the prevention and management of new-onset AF on the ICU.

Anaesthetists’ knowledge of laryngoceles and its management in acute respiratory failure. Questionnaire study

Huw Griffiths, Robert Coe, Azin Aghili, Parakkal Petson, Jeff Phillips and Rajnish Saha

Princess Alexandra Hospital, Harlow, UK

Abstract

Background: A laryngocele is characterised by dilatation of the laryngeal saccule resulting in a supra-glottic lesion.¹ Its incidence is estimated at 1 per 2.5 million populations per year and the risk factors for its development include any activities which endure repeated episodes of raised intra-laryngeal pressure such as coughing.¹ As laryngoceles do not extend below the glottis, adherence to the Difficult Airway Society’s (DAS) algorithm may require cricothyroidotomy to avoid an airway disaster. A patient with type 2 respiratory failure, worsened by non-invasive ventilation (NIV), who was subsequently found to have an unanticipated difficult airway due to a laryngocele, presented to our hospital. This clinical case prompted the authors of this study to review anaesthetists’ knowledge of this rare condition.

Objectives: The primary aim of this study was to investigate anaesthetists’ knowledge of laryngoceles and their understanding of DAS guidelines and the UK’s National Audit Project 4 (NAP4) findings. The secondary aim was to assess anaesthetists’ management of type 2 respiratory failure in an undiagnosed upper airway obstruction.

Methods: Anaesthetists working in a district general hospital in the UK were asked to complete a questionnaire with themes pertaining to the objectives of the study. Standardised questionnaire methodology, including anonymity and a five point Likert scale was used.

Results: 30 completed questionnaires were received. 13 respondents had greater than 15yrs anaesthetic or critical care experience, 10 had 2–15yrs experience and 7 had less than 2yrs experience. When asked how they rated their knowledge of laryngoceles 83% of the respondents rated it as limited or non-existent. When asked how they rated their knowledge of DAS unanticipated difficult airway guidelines only 17% of respondents rated it as extensive. A quarter of the respondents were unaware of the NAP4 recommendations for airway management in critical care. 87% of those surveyed would use NIV for a patient in type 2 respiratory failure with an unknown diagnosis of laryngocele.

Conclusion: There is limited knowledge of laryngoceles amongst anaesthetists and intensivists at our hospital. This may also be representative of clinicians at other hospitals, considering the rarity of the condition. Positive pressure ventilation via NIV may increase the size of a laryngocele which was the case in our clinical scenario. There are no guidelines on upper airway obstruction assessment prior to starting NIV which may need to be readdressed in light of our clinical case. Knowledge of difficult airway management in the Critical Care setting, may aid clinicians who have not encountered this condition, to manage it safely.

A Pilot Study on Feasibility of Intensive Care Unit Bereavement Follow-up Service

Nikki Yeo^1,2, Benjamin Reddi¹ and Stewart Moodie¹

¹Royal Adelaide Hospital, Adelaide, South Australia, Australia

²Addenbrooke's Hospital, Cambridge, UK

Abstract

Introduction: Death in intensive care units (ICU) is common, frequently unexpected and traumatic, and can be associated with marked psychopathology among bereaved family members. Good end-of-life care around and beyond death is crucial.

Objective: The primary aim was to determine the feasibility of providing bereavement follow-up (BFU) in the ICU. Secondary aims were to collate and evaluate the feedback received about end-of-life care and assess how the BFU service was viewed by respondents.

Design, setting and participants: This was a prospective, single-centre observational study conducted in a tertiary adult ICU in Australia. All patients who died in the ICU were included. Patients were referred to a 24-hour follow-up service at medical consensus of end-of-life. After the patient’s death, the nominated family member received information about bereavement and details of the 30-minute follow-up telephone interview. Calls were made at 4 weeks by trained clinicians, with questions based on a modified CAESAR Tool evaluating the experience of patient and family around the end of life. Responses were recorded as Likert scores (1, least satisfied to 5, most satisfied) with the option to provide qualitative feedback.

Results: From February to September 2018, 129 out of 151 (85%) ICU deaths were enrolled. Eighty-five (77%) calls were finalised and out of these calls, 52 (61%) families completed the follow-up calls between March and September 2018. The median score for questions relating to care of patient and family, and communication with health professionals was 5 (most satisfied); whether families were clearly informed if their loved ones were dying had a median score of 4 (IQR 3–5). Qualitative feedback indicated that the patient being acknowledged as a person was viewed positively, whilst lack of clarity around the severity of illness and being kept apart from loved ones were recurrent sources of dissatisfaction. Majority (80%) of families were comfortable having organ donation raised and were positive about the opportunity to discuss it. Majority of the responders also found the follow-up telephone interview a positive experience.

Conclusion: Comprehensive bereavement follow-up after death in ICU is feasible but resource intensive. Calls were well received and identified areas that were done well during the end-of-life care but also highlighted areas for improvement.

Reading Between the “Lines”: Improving Renal Replacement Therapy in a District General Hospital

Naomi French, Radha Sundaram, Lisa Gemmell and John Hunter

NHS Greater Glasgow and Clyde, Glasgow, UK

Abstract

Background: The Kidney Disease Improving Global Outcomes (KDIGO) guidelines recommend using regional citrate anticoagulation for continuous renal replacement therapy (RRT) in critically ill patients with acute kidney injury. We introduced citrate RRT in May 2017 in our intensive care unit.

Aim: To assess the effect of citrate RRT on filter usage, RRT days, and filter downtime in critically ill patients in a district general hospital.

Methods: We performed a retrospective observational study looking at patients admitted to our 7 bedded Level 3 critical care unit (370 patients per annum) for one year after the introduction of citrate RRT. We included patients over 16 years of age who had received RRT in our unit. Data was collected from WardWatcher, a Scottish Intensive Care Society Audit Group (SICSAG) database, electronic patient records (EPR) and PACS (a radiology information system). 53 patients received RRT on the unit during the year. A total of 8 patients were excluded; five of these patients had no RRT prescription chart on the EPR and three of them had received heparin.

Results: There were a total of 53 patients, median age 58 years with 60% males. Of the 45 patients whose data were available, the average time on citrate RRT was 3.1 days vs 4.6 in 2015 giving a p value of 0.157. Average number of set changes was unchanged with 1.8 set changes per patient after introduction of citrate vs 1.8 in 2015, giving a p value of 0.362. We achieved the prescribed blood pump speed 80% of the time with citrate RRT as opposed to only 50% of the time in 2015. In 40% of the patients who received citrate RRT, the filter ran 100% uninterrupted. Filter clotting was still the commonest reason for filter set changes. A subgroup analysis based on line site (i.e. right internal jugular (RIJ), left internal jugular (LIJ), femoral veins (FV), or subclavian veins (SV)) was carried out due to the high proportion of LIJ lines in our unit. This revealed an average of 1.2 set changes for RIJ lines, 1.4 for LIJ, 0.67 for FV and 0 for the single SV line inserted. These included changes due to routine filter set changes at 72 hours.

Discussion: It was predicted that the introduction of citrate RRT would reduce number of filter set changes per patient and filter downtime, ultimately reducing cost. The reasons for not being able to demonstrate reduction in filter set usage may include small numbers, line position (LIJ), and not recommencing filter after short planned interruptions such as CT transfers. It has not been possible to do a composite cost analysis at this stage to demonstrate cost reduction. We hope to feedback results of this study to the multidisciplinary team, improve awareness regarding choice of site for dialysis lines, and re-examine our outcomes in a year's time.

A survey examining Mechanical Insufflation-Exsufflation use in UK, adult intensive care units

Ema Swingwood¹, Fiona Cramp² and Lyvonne Tume²

¹University Hospitals Bristol NHS Foundation Trust, Bristol, UK

²University of the West of England, Bristol, UK

Abstract

Background: The use specifically of mechanical insufflation-exsufflation (MI-E) to aid weaning and extubation has been investigated (Goncalves et al., 2012; Bach et al., 2015) with a recent Cochrane review (Rose et al., 2017) concluding that MI-E inclusion may increase successful extubation rates. There are a number of devices available for use within the UK, however, to date it is not known if, and how widely MI-E is used by Physiotherapists on intensive care units (ICU) across the UK.

Aim: To describe the use of MI-E devices in adult ICU across the UK.

Specific objectives include:

• To quantify use of MI-E in an ICU setting across the UK

• To explore perceived enablers/barriers to using the device

• To examine ICU patient groups that clinicians feel are/are not appropriate for MI-E use

Methods: Survey development included the peer review of questions by an expert clinician familiar and experienced with MI-E but with a neuromuscular population. A final survey draft was then trialled and tested for ease of understanding and flow.

Physiotherapists working in a permanent post on UK, adult ICU were invited to participate in an electronic survey.

Ethical approval was granted by the Faculty Research Ethics Committee, University of the West of England, Bristol. Survey return was considered as clinicians providing consent to participate.

Results: 166 complete surveys were returned, with good geographical UK representation. Over half (72% 119/166) respondents were band 7 and above, with a mean (±SD) of 13 (±7) years since qualification.

Nearly all (98% 163/166) clinicians had access to MI-E on ICU, with a range of devices reported. Estimated frequency of use varied, with the majority reporting weekly or monthly use (n = 52/163, 32% and n = 50/163, 31% respectively). Just over half of respondents (n = 86/163, 53%) stated they use MI-E with intubated patients. In contrast, 99% of clinicians stated MI-E use with extubated patients. Of those clinicians who did not use MI-E in intubated patients (n = 74/163, 45%), a range of perceived barriers was reported including a perceived lack of clinical need, and a lack of confidence and perceved competence in using the device.

Conclusions: MI-E is widely available on adult ICUs across the UK. Clinicians reported greater use with extubated patients. Barriers to MI-E use in the intubated population warrants further investigation.

High-flow nasal cannula therapy for type 1 respiratory failure outside of critical care: Is it safe?

Susan Jackson and Edward McKeown

Royal Berkshire Hospital, Reading, UK

Abstract

Introduction: High-flow nasal cannula therapy (HFNC) reduces mortality in type 1 respiratory failure (T1RF) compared to conventional oxygen therapy and non-invasive ventilation strategies.¹ There remain concerns that its use may delay intubation leading to worse clinical outcomes.² We reviewed outcomes of patients started on HFNC for T1RF on ITU and a non-critically care staffed level 1 unit at a district general hospital.

Method: Consecutive adult patients, who were candidates for ITU admission and started on HFNC for T1RF over a 3-month period were identified and followed up to death or discharge. All patients were reviewed by a senior clinician to determine suitability for initial treatment site and escalation. 41 patients were identified and 9 excluded as they were deemed unsuitable for escalation to ITU on initial senior review. Patients were either given HFNC on ITU or a non critically care staffed level 1 unit (Higher Medical Unit) but with input from the Critical Care Outreach Team. Patients were assessed for the following outcomes: mortality and intubation rates. This work was approved by the local audit committee

Results: 10 (31%) patients started HFNC on ICU whilst 31 (69%) started HFNC on a level 1 higher medical unit (HMU). 8 patients were subsequently escalated to ITU from HMU (26%). There was no significant difference in intubation rates, mortality or Length of stay (LoS) between the patients who were admitted immediately to ITU and those who were escalated (Table 1).

Table 1.

Rates of intubation, Mortality and LOS for patients receiving HFNC for T1RF.

Patient group		Intubation	Mortality	LoS (days; median/IQR)
Admitted direct to ITU		6/10 (60%)	1/10 (10%)	31 (13–45)
Admitted to HMU	Escalated to ITU	3/8 (37.5%)	1/8 (12.5%)	13.5 (7–39)
	Remained on HMU	N/A	0/22 (0%)	12.5 (10–23)

Conclusion: Our data suggests that patients escalated to ITU from HMU have comparable outcomes, compared with patients admitted directly to ICU. With appropriate staffing and input from critical care and Critical care outreach, this provides a safe management environment without adverse effect on mortality or length of stay. The reduced rate of intubation in those escalated, compared with direct admission, suggests effective assessment of patients and treatment location. This suggests that at this district general hospital we are safely providing HFNC outside of the ICU setting which arguably has both a capacity and cost benefit. Guidelines are currently in development to ensure continued safe practice.

Lung-derived soluble mediators cause ventilator-induced lung injury in a porcine ex-vivo lung perfusion model

Andrew Boyle^1,2, Michael McKelvey¹, John Conlon¹, James McNamee², Danny McAuley^1,2 and Cecilia O'Kane¹

¹Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

²Regional Intensive Care Unit, Royal Victoria Hospital, Belfast, UK

Abstract

Background: Mechanical ventilation with high tidal volumes and alveolar pressure drives ventilator-induced lung injury (VILI). The mechanisms of VILI, and the relative importance of cellular inflammation and alveolar epithelial barrier function remain unclear. Furthermore, the biotrauma hypothesis proposes that mediators produced in the setting of injurious mechanical ventilation drive systemic organ dysfunction. Data from murine studies support the presence of a biological mediator entering the systemic circulation and driving lung injury in VILI [1].

Methods: We hypothesised that the perfusate recovered from porcine lungs receiving injurious ventilation could induce lung injury in porcine lungs receiving non-injurious ventilation.

In the ex-vivo lung perfusion model, isolated lungs are ventilated, and receive perfusion with an albumin-containing solution supplemented with 10% whole blood. After gentle re-warming to 36°C, porcine lungs were assigned to either control (10 cmH₂O CPAP), VILI (12 ml/kg tidal volume, zero end expiratory pressure), or perfusate transfer (control ventilation, perfused with perfusate obtained from VILI lungs). All lungs were ventilated for 4 hours. Bronchoalveolar lavage (BAL) and weight measurements were performed at baseline and 4 hours. The primary outcome measure was absolute weight gain, and secondary outcome measures included 4-hour BAL fluid total cell count and differential cell count.

Results: VILI and perfusate transfer lungs gained more weight than control (control 629 [399, 693], VILI 1096 [821, 1345] and perfusate transfer (1197 [753, 1922] g; p = 0.05) (Figure 1A). BAL total cell count at 4 hours was similar between groups (control 16.3 [15.0, 23.8], VILI 26.3 [15.8, 56.4] and perfusate transfer 14.7 [9.4, 23.8] x10⁷ cells) (Figure 1B). Similarly, BAL macrophage (control 16.1 [14.7, 23.2], VILI 25.6 [15.2, 53.4], and perfusate transfer 14.2 [9.2, 20.8] x10⁷ cells) (Figure 1C) and neutrophil (control 1.1 [1.0, 3.1], VILI 3.7, [1.3, 7.0], perfusate transfer 1.2 [0.2, 28.5] x10⁶ cells) counts showed no difference between groups (Figure 1D).

Conclusion: In an EVLP model of VILI, perfusate transfer alone caused lung injury characterised by weight gain, a marker of pulmonary oedema. This injury was independent of alveolar cellular infiltration, a finding that is in keeping with previous studies that suggest regulation of alveolar barrier integrity may be independent of cellular infiltration.

Are there jobs for consultants with a single CCT in intensive care medicine?

Thomas Jones

Kings College Hospital, London, UK

Abstract

Introduction: The single CCT in Intensive Care Medicine was launched in 2011. Since then over 600 doctors have joined the new ICM curriculum. With the new single CCT came the option to dual train in anaesthesia, acute internal medicine, renal medicine, respiratory medicine or emergency medicine. Recruitment data from the 2016 national recruitment to ICM ST3 showed that 5.5% intended to complete the single CCT. The rest intended to dual with anaesthetics (65.5%), medicine (21.5%) or emergency (7.5%). A frequently asked question of the Faculty of Intensive Care Medicine is “will there be single ICM consultants jobs?” The FICM highlight they do not have control over the creation of consultant jobs. As a trainee from a medical background considering single CCT ICM training I wanted to know if there will be jobs for consultants with a single CCT in ICM?

Methodology: Between 18^th October 2017 and 17^th January 2018 consultant jobs in adult intensive care medicine were identified using the website “www.jobs.nhs.uk”. Individual job descriptions and person specifications were reviewed and jobs were classified as either: 1) Anaesthesia with a special interest in ICM 2) Dual ICM and anaesthesia 3) Single ICM 4) Dual ICM (not limited to anaesthesia) 5) Single ICM or dual (flexibility within recrutiment).

Results: Over the three month period 60 jobs were identified. There were 41 jobs that required a CCT (or equivalent) in ICM. Table 1 shows the distribution of jobs available as per the above defined categories. It also shows the distribution of whether such jobs were in regional hospitals or district general hospitals.

Of the four “dual ICM” jobs all accepted anaesthetics as the dual speciality, two accepted acute medicine, and one accepted respiratory or emergency, none accepted renal. Of the ten “single ICM or dual” all accepted anaesthetics, five did not specify (possibly implying any dual CCT), three accepted acute medicine and two accepted emergency medicine. Two mentioned “medicine” as the second CCT, whether this meant any medical speciality was not clear. None specifically accepted respiratory or renal.

Table 1.

	Regional/Tertiary Hospital (24 jobs)	District General Hospital (36 jobs)	Total (60 jobs)	Total of jobs requiring ICM CCT (41 jobs)
Anaesthesia with a  special interest in ICM	1	18	19 (32%)	N/A
Dual ICM and anaesthesia	6	7	13 (22%)	13 (32%)
Single ICM	9	5	14 (23%)	14 (34%)
Dual ICM	1	3	4 (7%)	4 (10%)
Single ICM or dual	7	3	10 (17%)	10 (24%)

Summary: Many consultant intensive care jobs offer flexibility in recruitment. There is a significant proportion of jobs available to single CCT ICM consultants. Anaesthetics is the largest dual CCT speciality at consultant level, this correlates with data of training intentions from 2016. There is however a potential gap in the proportion of consultant jobs available and trainees’ intentions to dual CCT in ICM and emergency and potentially a similar gap with dual CCT in ICM and medicine. This may require consultants who dual trained to practise in a single speciality.

The state of focused intensive care echocardiography training amongst non-consultant grade intensive care doctors in a large tertiary centre

Thomas Jones, Richard Fisher and Shahana Uddin

Kings College Hospital, London, UK

Abstract

Introduction: Focused Intensive Care Echocardiography (FICE) is point of care echocardiography by non-experts to assist in the management of patients with undifferentiated shock. Accreditation requires the completion of a course, 10 supervised scans and mentored completion of a log book of cases. Within the critical care department of our large tertiary hospital we aimed to identify the interest, requirements and barriers to accrediting in FICE.

Methodology: Using “www.surveymonkey.com” a survey was sent to all non-consultant grade medical staff in our critical care department. The survey was open for three weeks from the 15th of February 2018. The survey explored how much echocardiography training they had received and what barriers they had faced in completing FICE training? They were also given the option to write any free text comments.

Results: 32 non-consultant grade doctors answered the survey. None had completed British Society of Echocardiography comprehensive transthoracic echocardiography accreditation. Two had completed a focused echocardiography accreditation (FICE or equivalent). Four had attended a course as well as completing 10 supervised scans and were in the process of completing a logbook. 13 had attended a course but had been unable to complete their supervised scans. 13 were planning to attend a course. The most common barriers to completion were a lack of mentor (21/32; 66%) inability to achieve 10 directly supervised scans (18/32; 56%) and lack of suitable ultrasound equipment (16/32; 50%).

Summary: Within our department there is a high level of demand for accreditation in focused echocardiography. However – having attended a FICE course – many non-consultant grade doctors struggle to progress with the accreditation process, with the most common stumbling block being a lack of available mentors. This situation is not unique, with one small study reporting that most individuals who had attended a FICE course in Reading had not gone on to achieve 10 supervised scans. If focused echocardiography is to form part of the normal working practises for intensivists in the UK we believe that training in performing point of care ultrasound should become a mandatory component of the FICM curriculum, which will encourage departments to invest in the necessary ultrasound equipment and adequately trained personnel to act as mentors.

Too much of a good thing? Auditing the prescription and titration of Oxygen on a general intensive care unit at a university teaching hospital

Christopher Hebbes, Rebecca Perry, Nathan Sloane and Alex Keeshan

University Hospitals of Leicester NHS Trust, Leicester, UK

Abstract

Background: Oxygen is the most commonly used drug in anaesthesia and intensive care, and both hypoxia and hyperoxia are associated with complications of organ dysfunction. In the acute setting, high arterial Oxygen tension is associated with harm in the perioperative patient, traumatic brain injury and myocardial infarction.

In surgical practice, WHO recommendations are for surgical patients to receive FiO₂ 0.8 for the immediate perioperative period and 2–6 hours postoperatively to reduce surgical site infection (SSI).¹ However, these recommendations are controversial. The recent OXYGEN-ICU trial demonstrated a significant mortality and morbidity reduction in a mixed ICU population when Oxygen was targeted conservatively (94–98%), compared with conventional practice (97–100%).²

The PaO₂ at which the harms of Oxygen exceed the benefits is unknown at present, and a guideline cut-off has not been defined. Therefore, this audit seeks to define current practice in a mixed surgical and medical intensive care unit in the prescription, titration and administration of Oxygen therapy, using the British Thoracic Society guidelines for the administration of Oxygen in emergency patients as a standard.³

Standards: 95% of patients should have a target saturation range specified during their admission

95% patients not at-risk of hypercapnic respiratory failure should have a saturation target of 94–98% documented

95% of patients at-risk of hypercapnic respiratory failure should have a saturation target of 88–92% documented

Methods: A retrospective audit was conducted during March 2017 of all admissions to the intensive care unit at the Leicester Royal Infirmary. Reason and time of admission, and any risk of hypercapnic respiratory failure were extracted from the notes. Hourly Oxygen saturations were recorded from bedside nursing charts. A record of all arterial blood gases taken was downloaded from the gas analyser, discarding any non-arterial sample, and samples from patients not on the intensive care unit.

Data were coded and analysed using SPSS and R for Windows, and permission gained from the Clinical Audit Department and Privacy Officer.

Results: During March 2017, there were 107 admissions, 42 (39.3%) medical and 65 (60.7%) surgical. Of these, 5 (4.7%) were at risk of hypercapnic respiratory failure. 48 (44.9%) of patients had a saturation target specified at any time during their admission. Of the 102 patients not at risk of hypercapnic respiratory failure, 44 (43.1%) had a target specified, although only 27 (26.5%) of these patients had an appropriate target specified with an upper and lower limit. The corresponding figures for the hypercapnic group were 4 (80%) and 3 (60%).

Of the 58 patients that were not initially normoxic, 3 did not achieve normoxia. For those that did achieve normoxia, median time to normoxia was 290 minutes. Normoxia was achieved within 6 hours for (34) 61.8% of cases.

Conclusions and Recommendations: In agreement with British Thoracic Society findings, these data highlight that prescription and titration of Oxygen is poor. Improved awareness of appropriate saturation targets with an initial goal set at admission may result in a more rapid down-titration of Oxygen, avoiding hyperoxia.

Derivation of a screen for identifying severe sepsis and septic shock in the Emergency Department: BOMBARD vs SIRS and qSOFA

Steven Rothrock, David Cassidy, Drew Bienvenu, Erich Heine, Brian Guetschow, Joshua Briscoe, Sean Isaak, Kenneth Chang, Mikaela Devaux and BOMBARD Study Group

Orlando Health, Orlando, FL, USA

Abstract

Background: Experts recommend screening adults for sepsis using systemic inflammatory response syndrome (SIRS) criteria or the quick sequential sepsis-related organ failure assessment (qSOFA) score. However, these screens may be inaccurate at identifying early sepsis. Our study was performed to develop an emergency department (ED) screen to identify patients who have or develop severe sepsis/septic shock in the hospital.

Methods: We conducted a retrospective case-control study of patients ≥18 years old admitted to two urban hospitals with a combined annual ED census of 162,000. Study cases comprised patients admitted via the ED with a final hospital discharge diagnosis of severe sepsis or septic shock. Controls consisted of patients ≥18 years old admitted to the same study hospital via the ED without severe sepsis/septic shock. Clinical features were compared between severe sepsis/septic shock and non-sepsis groups using the Mann Whitney U test for continuous and ordinal variables and χ² analysis or Fisher’s exact test for categorical variables. Continuous and ordinal variables were dichotomized using ROC curves. Using multivariate logistic regression, a prediction rule was constructed. The model’s AUROC was internally validated using 1,000 bootstrap samples.

Results: 143 study and 286 control patients were evaluated. Features predictive of sepsis included: SBP ≤ 110 mm Hg, shock index/SI ≥ 0.86, Mental status abnormality or GCS < 15, Breathing ≥ 22 per minute, Assisted living facility residency, Rise in temperature ≥38°C, and Disabled immunity (mnemonic BOMBARD). Two points were assigned to SI and temperature with other features assigned one point. BOMBARD was superior to SIRS criteria (AUROC 0.860 vs. 0.798, 0.062 difference, 95% CI 0.022–0.102) and qSOFA scores (0.860 vs 0.742, 0.118 difference, 95% CI 0.081–0.155) at predicting severe sepsis/septic shock. An initial BOMBARD score ≥3 was more sensitive than SIRS criteria ≥2 (74.8% vs. 49%, 25.9% difference, 95% CI 18.7–33.1) and a qSOFA score ≥2 (74.8% vs. 33.6%, 41.2% difference, 95% CI 33.2–49.3) at predicting sepsis. A BOMBARD score ≥3 was superior to SIRS criteria ≥2 (76% vs. 45%, 32% difference, 95% CI 10–50) and a qSOFA ≥2 (76% vs. 29%, 47% difference, 95% CI 25–63) at predicting sepsis mortality.

Discussion: The U.S. Centers for Medicare and Medicaid currently monitors hospitals for outcome of patients with sepsis using older Sepsis-2 definitions (severe sepsis, septic shock). Thus, we chose these definitions to identify study patients.

Our prediction rule was derived by comparing admitted sepsis to non-sepsis patients. Subsequently, BOMBARD might be less useful in a general ED population with the majority discharged from the ED. It is more likely to be useful in identifying ED patients with a high probability of being admitted who are at risk for severe sepsis/septic shock.

Conclusion: A clinical prediction rule was developed to identify patients who have or develop severe sepsis/septic shock in the hospital. This rule was more accurate and more sensitive than SIRS criteria and qSOFA scores at predicting severe sepsis/septic shock and sepsis mortality. Prospective validation in larger, more diverse settings is needed before clinical use.

Can Neuron Specific Enolase levels predict death in ITU patients post out of hospital cardiac arrest?

Frances Palmer¹, Oliver Clifford-Mobley¹, Graham Bayly¹, Agnieszka Skorko² and Kieron Rooney²

¹Department of Clinical Biochemistry, University Hospitals Bristol NHS Foundation Trust, Bristol, UK

²Department of Intensive Care Medicine, University Hospitals Bristol NHS Foundation Trust, Bristol, UK

Abstract

Neuron specific enolase (NSE) is found in neural and neuroendocrine tissue. Elevated concentrations of NSE in serum are often found in diseases following rapid neuronal death. Previous studies have proposed using NSE as a marker of prognosis post cardiac arrest (1,2,3). Our laboratory introduced an NSE assay in 2017 for ITU patients admitted following out of hospital cardiac arrest (OOHCA); levels are taken on ITU admission and 48 hours later and the change (delta) in NSE concentration (ΔNSE) calculated. NSE is measured in serum using an automated immunoassay (Roche Diagnostics) with a measuring range of 0.05–370.0 ng/mL. The aim of this study was to determine the best Δ NSE cut-off to predict death in ITU.

Inclusion criteria were patients admitted to ITU between 19/10/2017 and 18/08/2018 following OOHCA who had NSE samples collected at baseline and 48 ± 6 hours post ITU admission. Samples were excluded if they were not separated and frozen within 1 hour of collection or if the haemolysis index was >20 (studies in our laboratory have shown that haemolysis causes an increase in serum NSE due to its release from erythrocytes). There were 72 patients who met these criteria, of whom 33 died in ITU. The NSE results are summarised in the table below.

	Survived to ITU discharge (n = 39)	Died in ITU (n = 33)
Baseline NSE- Median	28.9	34.5
Baseline NSE- Range	16.0 – 125.0	17.0 – 125.8
Baseline NSE- IQR	22.7 – 37.7	24.7 – 43.3
48 hour NSE-Median	18.0	93.6
48 hour NSE-Range	9.9 – 79.1	17.5 – 370.0
48 hour NSE-IQR	15.2 – 23.2	40.1 – 271.9
ΔNSE –Median	–10.5	51.7
ΔNSE –Range	–45.9 to +1.7	–80.4 to +345.0
ΔNSE –IQR	–16.3 to −6.3	2.9 – 235

Units of measurement ng/mL

The baseline NSE results overlapped significantly between those who died and those who survived to ITU discharge. In contrast, any positive ΔNSE was an excellent predictor of death in ITU. All the patients who survived to discharge from ITU had ΔNSE <1.7 ng/mL and in 37 of the 39 survivors ΔNSE was actually negative. Therefore using the cut-off ΔNSE >1.7 ng/mL gave a specificity of 100% for predicting death in ITU, with a sensitivity of 79%. Other authors have suggested higher ΔNSE cut-offs to predict poor outcome: Rundgren et al. >28 ng/mL (1) and Stammet et al. >33 ng/mL (2). The data presented here are in keeping with this previous literature, however it should be noted that our primary outcome was death in ITU, rather than cerebral performance score (CPC) at 6 month follow up. Applying these higher ΔNSE cut-offs to our data keeps the specificity of 100% but results in a lower sensitivity of 58%.

In conclusion, our data show that NSE results from a single time point do not aid prognostication. In contrast the change in NSE over 48 hours is an excellent rule out test because almost any positive ΔNSE means there is a very low chance of survival.

Video-assisted nasojejunal insertion: evaluation of a novel bedside technique to aid enteral nutrition

M. Asim T. Khan and Stephen J. Shepherd

St Bartholomew's Hospital, Barts NHS trust, London, UK

Abstract

Introduction: Malnutrition remains a significant concern on intensive care yet may be multifactorial in nature.¹ Surgical stress, acute illness and hypermetabolism coupled with medication effects and delayed gastric emptying all hinder nutrient delivery. Inadequate nutrition may prolong ICU length of stay. Post-pyloric feeding via nasojejunal tubes (NJT) increase nutrient delivery and may reduce the risk of pneumonia.² NJT insertion usually requires endoscopic support but logistical factors may bar their placement. Non-endoscopic methods of insertion using fluoroscopy, electro-magnetic guidance devices, blind insertion with weighted tubes or pH sensitive tips offer varying degrees of success.³ St. Bartholomew’s Hospital is a teaching hospital providing care predominantly for cardiothoracic and oncology patients without elective on-site gastroenterology or interventional radiology services. Placement of NJT requires coordination with colleagues from our sister hospital and specific arrangements. Due to these logistical difficulties our default practice for patients with evidence of delayed gastric emptying (high gastric residuals, vomiting, etc), has been to initiate parenteral nutrition after failure of prokinetic therapy. This is costly with risks including the need for central venous access, frequent insulin administration and potential infectious complications. We piloted the use of a novel video-assisted technology allowing bedside NJT insertion. The KangarooTM IRIS device developed by Covidien enables NJT insertion under direct visual guidance using real-time playback of images taken from a 3 mm camera proximal to the tip.

Methods: Patients were identified from within St. Bartholomew’s. The project was deemed a service evaluation and ethics approval waived. A single operator familiar with a number of methods of NJT insertion acted as primary operator. Indications for post-pyloric feeding were assessed by dietetics who initiated referral but all included recurrent gastric residual volumes >250 ml every four hours despite appropriate prokinetic therapy. 80 cm 10Fr tubes were placed using the IRIS technology using a gastric insufflation technique with advancement under video surveillance. Tubes were bridled to prevent movement and placement confirmed using X-Ray as per current Trust recommendations.

Results:We attempted NJT insertion in eight patients. Siting was successful in seven (87.5%), comparable to endoscopic rates of 90%.^4,5 One patient required later endoscopic insertion. Five out of seven patients had a significant reduction in gastric residual volumes. One patient required supplemental parenteral nutrition due to bowel ileus and there was one episode of tube failure. No other complications were observed.

Discussion: Endoscopic insertions cost up to £700 per patient; each NJT costs £180, but total expenses exceed £1000 including transportation costs and staff time. The Kangaroo console costs £1100 and each disposable tube is priced from £140. Costs for parenteral nutrition may exceed this many times over. This technology offers the potential to avoids such costs and the logistical problems associated with endoscopic insertion.

Performance analysis of the VILIAlert clinical decision support tool for protective lung ventilation

Aleksandar Novakovic¹, Charles Gillan², Adele Marshall¹, Murali Shyamsundar³ and Tonny Veenith⁴

¹School for mathematics and physics, Queen's University Belfast, Belfast, UK

²Centre for Data Science and Scalable Computing, Queen's University BElfast, Belfast, UK

³Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

⁴School of Infection and Immunity, University of Birmingham, Birmingham, UK

Abstract

Performance analysis of the VILIAlert clinical decision support tool for protective lung ventilation

Background: Lung protective ventilation (LPV) significantly reduces mortality in acute respiratory distress syndrome (ARDS) and reduces ARDS development in critically ill patients. Lack of translation of evidence to practice puts patients at risk and is a global issue. Clinical decision support (CDS) driven by real time data analytics has the potential to overcome the barriers to LPV delivery and lead to improved patient outcomes.

Objective: To investigate the performance of the VILIAlert clinical decision support tool that is being developed to improve the quality LPV delivery.

Study Design: The VILIALert project is an ongoing quality improvement project (QIP) in the Regional Intensive Care Unit, Royal Victoria Hospital, Belfast to improve compliance to LPV. Tidal volume (TV)/kg ideal body weight (IBW) received every minute and height of patients admitted between 31.08.2016–26.10.2017 was used to develop and test VILIAlert. LPV was defined as TV ≤8 ml/kg IBW and the average TV/kg IBW was calculated every fifteen minutes. Injurious ventilation was defined as a breach of LPV threshold for four successive fifteen-minute intervals.

Methods: VILIAlert performance measured by Shewart p chart based on the assumption that if an alert is generated for breaching the 8 ml/kg IBW tidal volume for four consecutive fifteen-minute intervals, the tidal volumes in the near future will also violate the threshold. If future fifteen-minute averaged values do not breach the threshold, it will be considered a “defect”. We consider alert as genuine if thenumber of defects do not exceed the upper control limits of the p-chart and define accuracy as the proportion between number of genuine and total number of alerts. Defects map relies on Ward’s hierarchical clustering algorithm to display alerts with similar defects patterns grouped together, and hence provide better insight into the patients’ tidal volume oscillations during the 6 hours after the alert has been triggered.

Results: Routinely collected anonymised tidal volume, gender and height for mechanically ventilated patients in both “control” and “support” modes of ventilation were utilised. A total of 1450 alerts were triggered. Accuracy assessment of the VILIAlert system has been made based on the number of defects that occurred during the period of 24 fifteen-minute blocks after triggering the alert. The upper control limit for the p-chart was 52% suggesting the alert is genuine if the number of defects is no greater than 12 which implied the overall accuracy of the VILIAlert system of 76.7%. We sought to identify patterns in the TV trends by using a novel defects map.

Conclusions: The use of Shewart p chart in conjunction with defects map to analyse the performance of CDS is novel and provides an opportunity improve performance prior to real world testing. Previous studies have demonstrated potential uses of a CDS to improve LPV but with a positive predictive value of only 59%. It is essential to introduce effective noise filtering techniques to capture true events and improve the accuracy of VILIAlert or similar CDS.

Survey of the use of the landmark subclavian approach for trauma line access in code red trauma – A national and regional perspective

Barney Scrace, Thomas Geliot and Ben Warrick

Royal Cornwall Hospital, Truro, UK

Abstract

In line with national guidelines and consensus the drive to optimise resuscitation in major trauma is well established. NICE recommendations state that peripheral access should be obtained and if this fails intra-osseous access should be considered whilst central access is being achieved.² A key component of this is the rapid establishment of secure wide bore intravenous access.¹ Major trauma centres have developed their own ‘code red’ trauma guidelines taking into consideration the failings of the IO route when attempting to achieve the high flow rates required for effective massive transfusion protocols. Many of these include the use of landmark subclavian ‘trauma lines’. NICE recommendations for the insertion of central venous access under direct ultrasound scanner (USS) guidance is widely adopted.³ As a result of this there is a preference for internal jugular vein cannulation and the subclavian route is seldom used. Additionally, few trainees are comfortable inserting central venous access without the use of USS.

We undertook two surveys to define these issues. Survey 1 was sent to all major trauma units in the UK asking them to define their practice in terms of preference of anatomical site and access device in code red trauma when attempts at peripheral access have failed. Survey 2 was sent to all trainees above ST3 level in emergency medicine (EM), Anaesthetics, Intensive care medicine (ICM) and Pre-hospital emergency medicine (PHEM) in the South West/Peninsula deanery. This survey aimed to assess training, familiarity and confidence in the use of the landmark subclavian approach.

Figure 1.

MTC preferences for site of Trauma line insertion.

Figure 2.

Training and performance of subclavian CVC by trainees in South West and Peninsula deanery.

Survey 1

10/25 MTCs contacted replied giving a 40% response rate. 50% listed landmark subclavian as their preferred site of trauma line insertion. The remainder listed operator discretion, peripheral venous access, or internal jugular/femoral CVC under USS guidance. If attempts at peripheral venous access failed 60% listed IO access whilst central access was obtained. 60% centres stated that no formal training was provided for the insertion of trauma lines.

Survey 2

68/138 registrars contacted replied giving a response rate of 49%. 48% of respondents had not attempted the procedure. 56% stated they had received no formal training in insertion of landmark subclavian approach. 25% of respondents stated they had attempted the procedure without formal training and 26% had received formal training and attempted the procedure. 38% of respondents stated they would not attempt to insert subclavian approach trauma line if required. Of the reasons listed for not attempting the procedure responses included: No confidence, lack of experience and lack of training.

Our surveys demonstrate that the landmark subclavian approach to secure trauma line access is used in major trauma centres in the UK. We have highlighted the training gap that exists for the insertion of landmark approach subclavian venous access. We now intend to embark on a project to increase familiarity and confidence in the insertion of subclavian CVCs amongst trainees in our region.

Clinician and patient perceptions of the barriers and facilitators of mobilisation of intensive care patients in a UK NHS Trust

Huw R Woodbridge^1,2, Christine Norton³, Mandy Jones⁴, Caroline M Alexander^1,2 and Anthony C Gordon^1,2

¹Imperial College London, London, UK

²Imperial College Healthcare NHS Trust, London, UK

³King's College London, London, UK

⁴Brunel University London, Uxbridge, UK

Abstract

Introduction: Mobilisation of intensive care unit (ICU) patients has been shown to improve short-term functional outcomes. However, studies have demonstrated relatively little implementation of ICU mobilisation, therefore greater knowledge is required about the barriers and how they can be overcome. In particular there is a lack of in-depth consideration of the perceptions of clinicians from different professions and patients in a UK setting. This study therefore aims to explore the views of ICU clinicians, former ICU patients and their relatives/carers on the barriers and facilitators of ICU mobilisation. This abstract presents the results of the initial data analysis.

Methods: This study was given Research Ethics Committee approval. Purposive sampling was then used to select clinicians from different professions either involved in the decision to initiate ICU mobilisation or in carrying out the mobility treatment itself. Moreover, former ICU patients and their relatives/carers were recruited. Following informed consent, semi structured interviews were completed with each participant which were recorded, then transcribed verbatim. Finally, thematic framework analysis was used to analyse the interview transcriptions.

Results: Twenty five participants were recruited, including 15 clinicians (four doctors, four nurses, two occupational therapists and five physiotherapists), eight former ICU patients and two relatives.

Analysis produced an initial thematic framework of barriers and facilitators. Barriers comprised concerns over safety, resource and logistical issues, ICU culture and team working factors and staff experience levels. Furthermore, instances were reported where patients did not want to mobilise. Finally, patients described how clinician communication and the ICU environment could be a hindrance to mobilisation.

“…well I’m in ICU… you’re having intensive care, don’t rock the boat by making things worse by trying to get out of bed.” (Patient.)

Facilitators included suggested modifications to the mobilisation treatment sessions and strategies to improve team working and culture towards mobilisation. In addition, an increase in resources was suggested and patient characteristics that made mobilisation easier were identified. Finally, the importance of patient motivation was discussed.

“allow them to see… “I can actually do something and not only [be] here as a passive recipient of lots of treatment and being turned”” (Clinician)

Discussion: This study provides insight to the implementation of ICU mobilisation at an NHS Trust from the perspectives of ICU clinicians from a range of professions and levels of seniority and the experiences and opinions of former ICU patients and their relatives. Before an in-depth exploration of themes is undertaken, further interviews are needed to achieve apparent thematic saturation.

Funding acknowledgements: This report is independent research arising from a Clinical Doctoral Research Fellowship, awarded to HRW, (ICA-CDRF-2015-01-026), supported by the National Institute for Health Research (NIHR) and Health Education England. The views expressed in this publication are those of the authors and not necessarily those of the NHS, the National Institute for Health Research, Health Education England or the Department of Health. The authors also acknowledge the support of the NIHR Clinical Research Network and NIHR Imperial BRC. ACG is an NIHR Research Professor (RP-2015-06-018).

Condition	Sub-condition	Initial SUDS	End SUDS	Result
Within all  sessions	All sessions	M = 5.47 SD = 2.089	M = 4.67 SD = 1.905	T(88) = 8.8 p < 0.001
	Depression only	M = 6.41 SD = 2.146	M = 5.55 SD = 2.097	T(28) = 5.304 p < 0.001
	Anxiety only	M = 5.25 SD = 1.799	M = 4.33 SD = 1.493	T(23) = 4.609 p < 0.001
	Stress only	M = 4.86 SD = 2.002	M = 4.19 SD = 1.785	T(35) = 5.292 p < 0.001
Between first  and last  measure of  different sessions	All sessions	M = 6.62 SD = 1.521	M = 3.90 SD = 1.852	T(52) = 9.825 p < 0.001
	Depression only	M = 7.52 SD = 1.545	M = 3.40 SD = 2.222	T(16) = 4.366 p < 0.001
	Anxiety only	M = 6.5 SD = 1.019	M = 3.64 SD = 1.598	T(13) = 7.601 p < 0.001
	Stress only	M = 6.0 SD = 1.480	M = 3.4 SD = 1.501	T(21) = 6.417 p < 0.001

Developing and testing specific psychological input for critical care – how well does it work?

Christopher Lambert¹, Zoe Martin¹ and Jacobus Prellar²

¹Addenbrooke’s Hospital, Cambridge, UK

²John V Farman Intensive Care Unit, Addenbrooke’s Hospital, Cambridge, UK

Abstract

This abstract describes the initial results of an innovative project which has developed and used specific psychological techniques, matched to specific psychological problems, with critical care patients. This took place over an 18-month period during the launch and development of a psychological well-being service which covered the ICU and Neuro Critical Care Unit of Addenbrooke’s hospital. In order to test the validity of this process, problem specific Subjective Units of Distress (SUDS) on a scale of 1–10 were collected with a sample of fifty-seven patients before and after each intervention session in order to measure progress. Further, the Intensive Care Psychological Assessment Tool (IPAT) was used to identify initial psychological distress with all of the sample, and the Hospital Anxiety and Depression Scale (HADS) was undertaken early on and also at the end of the sessions with many of the patients in order to provide validated concurrent measurement of change.

Pearson’s correlation coefficient, and matched pair T-tests were conducted on the data in order to investigate: (1) the initial and developing relationship between the standardised IPAT measure and the SUDS measures; (2) whether a significant difference in SUDS can be detected both overall and for specific problems, within and between sessions; and (3) whether a significant difference in HADS for both depression and anxiety can be detected from beginning to end of the interventions.

Pearson’s correlation results are displayed in the table below:

Condition	r-value	P-value
IPAT and 1st SUDS  initial session	(57) = 0.39	P = 0.001
IPAT and 2nd SUDS  initial session	(57) = 0.321	P = 0.008
IPAT and 1st SUDS  final session	(39) = 0.032	P = 0.039
IPAT and 2nd SUDS  final session	(39) = 0.245	P = 0.066

Matched-pair T-test results:

The initial finding showing a moderate correlation between IPAT and SUDS scores indicates that the initial SUDS measures are positively connected to the standardised measure. Further the gradual reduction in this correlation may suggest that treatment effects are weakening this connection.

The finding that a strong statically significant difference exists between all the within and between SUDS sessions analysed suggest that overall the psychological interventions undertaken are effective at reducing distress in each area at the time of the respective sessions, and that this effect maintains between sessions. The finding that HADS measures for both depression and anxiety is also shown to reduce (HADS-Depression: t(16) = 3.742, p = 0.002; HADS-Anxiety: t(16) = 3.495, p = 0.003) demonstrates concurrent reduction amongst a sample of patients on a standardised measure, suggesting that the SUDS reductions may have a greater degree of objectivity than simply their immediate context

Time for ECPR?

Ian Tyrrell-Marsh¹ and Julian Barker²

¹North West Air Ambulance, Manchester, UK

²Wythenshawe Hospital, Manchester, UK

Abstract

Out Of Hospital Cardiac Arrest (OOHCA) carries a mortality approximating 90% in the UK.¹ Advanced Life Support algorithms provide a basis of care for the majority of OOHCA cases. However, certain subsets within this cohort are unlikely to respond; massive PE, shock resistant VF and hypothermic patients. These patients are often young, with few, if any co-morbidities.

Veno-Arterial (VA) ECMO carried out during an arrest situation offers patients resistant to standard ALS techniques the opportunity for continued support capable of providing near normal oxygenation and flow to critical organs. Primarily it provides the opportunity to transfer the arrested patient to specialist areas capable of treating the underlying pathology such as the catheter laboratory or interventional radiology.

Wythenshawe hospital have developed such a service locally, producing results in line with published ELSO data of approximately 30% survival to good neurological recovery.² Close collaboration between The North West Air Ambulance (NWAA) and the Wythenshawe ECMO team enabled patients from a much greater catchment area to be offered this opportunity. Strict admission criteria were adopted from current practice, an SOP developed, and rehearsals carried out during moulage training. To date NWAA have transferred 7 patients for assessment, of which 4 have been successfully established on ECMO. Average time from arrest to established flow on ECMO was 79 mins.

Discussion: ECPR offers a last chance to those patients unresponsive to standard ALS techniques. It is widely accepted that once the decision to go onto ECMO has been made there should be minimal delay to establishing flow, with positive outcomes diminishing with time.

Approaching jobs with an awareness of time of arrest, potential for ECMO admission and clearly defined task allocations scene times have reduced to a mean of 29 mins. To date none of the patients placed on ECMO have survived and it would seem logical that to maximise the benefit of this intervention we need to minimise the low flow state time. Whilst there undoubtedly is potential for further gains in terms of reducing scene and transit times to hospital these gains are likely to be marginal at best and provide no benefit for patients in difficult locations, for example upstairs rooms or remote from aircraft landing sites. A potential solution is the introduction of a service capable of providing ECMO at scene. Arrest to on scene time averaged at 22 minutes, allowing a further 20 mins for cannulation would give an arrest to ECMO time of approximately 45 mins; a potential reduction in low flow time of 30 mins.

Conclusion: We have demonstrated that identifying, reaching and transporting patients to an ECMO centre in a timely fashion is possible, if ECMO centres form close working partnerships with their local pre-hospital teams. Even given this, time to hospital is likely to be close to 60 minutes and outcomes may be adversely affected by this. Other strategies, for example pre-hospital ECMO may need to be considered in order to maximise the chance for good neurological outcomes.

A Tracheostomy Passport: safety on the journey through critical care

Emma Bryden and David Garry

Oxford University Hospital NHS Foundation Trust, Oxford, UK

Abstract

Introduction: Prolonged ventilation in intensive care may require patients to have a tracheostomy to facilitate respiratory weaning. The procedure has its own risks as well as potential long-term sequelae which can impact on the patient both physically and psychologically.

A national recommendation for a tracheostomy passport for each patient, available at the bedside, was made in 2014.¹ Guidelines for tracheostomy care in intensive care units from France² have highlighted the need for a standardised approach to pre-procedure assessment, ongoing care, follow up and ultimately decannulation.

Aim: To produce a tracheostomy pathway to aid clinician decision making pre-procedure, through ongoing critical care, discharge to ward care and decannulation using a format.

Results:

Conclusions: Introduction of this tracheostomy passport is a step towards standardisation of care for critical care patients in our institution. Audit of critical incidents involving tracheostomy patients before and after its introduction will allow us to assess its impact.

Adherence to non-injurious tidal volume ventilation in thoracic trauma: a tertiary trauma centre retrospective cohort analysis

Vasileios Zochios^1,2, Joht Singh Chandan³, Éimhín Dunne¹ and Tomasz Torlinski¹

¹University Hospitals Birmingham National Health Service Foundation Trust, Department of Anaesthesia and Intensive Care Medicine, Queen Elizabeth Hospital Birmingham, Birmingham, UK

²Birmingham Acute Care Research Group, Institute of Inflammation and Ageing, Centre of Translational Inflammation Research, University of Birmingham, Birmingham, UK

³Institute of Applied Health Research, University of Birmingham, Birmingham, UK

Abstract

Background: Pulmonary contusions due to thoracic trauma can potentially lead to acute respiratory distress syndrome.¹ In patients with ‘injured lungs’, low tidal volume ventilation (LTV) [<6-8 ml/kg predicted body weight (PBW)] has been shown to improve outcomes.^2–4 It is therefore paramount that a ‘non-injurious’ ventilatory strategy is adopted in patients with pulmonary contusions. We sought to determine the rate of adherence to LTV, and association between tidal volume delivered and patient-relevant outcomes in a cohort of invasively ventilated patients with chest trauma.

Methods: Following institutional approval, a cohort of consecutive patients with pulmonary contusions (between August 2016 and August 2017) requiring invasive mechanical ventilation (IMV) on intensive care unit (ICU) were retrospectively reviewed. The primary end-point was rate of adherence to LTV for the duration of IMV. Simple descriptive comparison between groups was undertaken either by chi-squared tests (categorical data comparison) or Wilcoxon Rank Sum (non-parametric continuous data). In order to test for association between groups for length of stay (LoS), generalised linear regression was used to ascertain a regression coefficient to describe a unit of change per dependent variable. For binary independent outcomes such as mortality, logistic regression was used to determine an odds ratio. Univariable and multivariable logistic regression was used to construct a predictive model for non-adherence to LTV.

Results: Ninety-eight patients were included in the analysis. The median (IQR [range]) age (years) was 33 (24–48 [17–92]) with 85% of the population studied being male. The median tidal volume (ml) delivered in this cohort was 541 (504–590 [366–820]) and tidal volume (ml/kg PBW) was 8.0 (7.2–8.6 [5.2–13.4]). Fifty patients (51%) received tidal volume greater than 8 ml/kg PBW and only six patients (6.1%) received tidal volumes between 6 and 6.5 ml/kg PBW. Apart from height and blood transfusion during IMV, there were no statistically significant differences in baseline characteristics or clinically important outcomes between the high (>8 ml/kg PBW) and low (<8 ml/kg PBW) tidal volume groups. The between-group difference in LoS [regression co-efficient (95% CI)] and ICU mortality [adjusted odds ratio (AOR), 95% CI] remained non-statistically significant: ICU LoS −1.80 (-8.44–4.83 p = 0.595), hospital LoS 10.30 (-5.87–26.48; p = 0.212) (Figure 1), and mortality 0.09 (0.00–3.59; p = 0.198). Following univariate analysis five baseline characteristics (sex, height, body mass index, pre-existing lung disease and blood transfusion) as well as two respiratory parameters [having a partial pressure of arterial carbon dioxide (P_aCO₂) >6kpa, and positive end-expiratory pressure] were identified to be associated with ‘injurious’ ventilation (p value <0.2). However, following multivariate analysis only two of these factors [height (AOR 0.94; 0.89–1.00 p = 0.038) and blood transfusion (AOR 2.81, 1.07–7.40 p = 0.036)] remained significant (p < 0.05). The c-statistic for the primary end-point was 0.7847 (95% CI, 0.69–0.88) (Figure 2).

Conclusion: In chest trauma patients, LTV ventilation is underutilised within our institution. Non-adherence was predicted with moderate discriminative ability. There is a need for continuous education and further investigation focusing on identifying barriers to adherence to best practice and strategies to overcome them. The performance of our model needs to be externally validated in a larger cohort.

An Innovative messenger-based virtual journal club for Critical Care Departments

Timen Swartbol^1,2, Dagmar Holmquist¹ and Rajnish Saha¹

¹Princess Alexandra Hospital, Harlow, UK

²Reinier de Graaf Hospital, Delft, The Netherlands

Abstract

Journal clubs have developed throughout the ages, from lounges where physicians socialised and discussed articles, to recent social media based platforms reaching out to a wide audience.¹ Departmental journal clubs are still the most prevalent, with each hospital servicing its own unique demographic with its own resources, especially in critical care. Therefore it is becoming increasingly relevant to reflect on local practices in a contemporary and user-friendly manner.²

Our aim was to create a messenger-based virtual journal club for our critical care department. The aim was for this journal club to be inclusive, multi-disciplinary and not to negatively impact on staff’s work-life balance.

The Princess Alexandra Hospital Critical Care Journal Club (PAH-CCJC) was created on 13th November 2017. The CCJC is based on WhatsApp, an end-to-end encrypted messenger application. Our group has grown to have multi-disciplinary membership, from junior and senior physicians to nursing staff, physiotherapists, pharmacists and a clinical librarian. A relevant article is chosen weekly by the CCJC’s Lead, working in tandem with our clinical librarian. Members can suggest articles based on clinical events that week and are encouraged to answer specific questions given by the Lead or engage in discussion with each other via the app. Rules ensure patient confidentiality, social posting times and professional language. At the end of the week, a critical appraisal is recorded detailing the salient points of the discussion. Articles and appraisals are stored online, which can be accessed and reflected on to guide our practice together with local guidelines.

Over 5 months we have appraised 16 articles from 10 journals. We recruited 51 members who wrote 353 unique discussion posts. Article topics and number: Cardiology 2, GI 1, Haematology 1, Neurology 2, Organisational 1, Physiotherapy 2, Radiology 1, Renal 1, Respiratory 4, Sepsis 1.

This format is easy to set up and allows participation regardless of location and time-constraints. It generates an accessible record of discussions that is easily audited, both in the actual chat and appraisals. The club format will continue to be updated based on feedback from our members.

We have found a virtual journal club to be the most pragmatic format for the challenges of maintaining an inclusive multi-disciplinary sustainable journal club for a department of our size.

Temperature Management Post Cardiac Arrest – Time to standardize?

Tom Syratt and Paul McLindon

Cheshire and Mersey Critical Care Network, Liverpool, UK

Abstract

There are an estimated 30 000 out of hospital cardiac arrests (OOHCA) in the UK annually and whilst cardiac arrest centres have become prominent in other European countries, this is not yet the case in the UK. Mersey is a relatively small region but includes ten individual hospitals with intensive care units responsible for the care of post arrest patients, though only one of these hospitals performs primary percutaneous coronary intervention.

A 2013 national survey of post cardiac arrest temperature management found a large variation in practice in both duration (12–72 hours) and target temperature (<32–37°). This survey was performed prior to the release of the Targeted Temperature Management trial which showed no difference in outcomes between a target temperature of 33° & 36°.

As intensive care trainees rotate around the region, discrepancies between protocols have been highlighted with only 30% of trusts using identical protocols. Of the hospital 60% would consider using temperature control in all arrest rhythms with the remainder stipulating specific rhythms excluded. Target temperature varied from 33° to 37° and the duration from 24 to 72 hours. The mechanism to control temperature was stipulated by 50% of the units and of these there were four different mechanisms used.

Getting it right first time (GIRFT) is a national programme aiming to improve the quality of care delivered by reducing unwarranted variations in practice. Our practice in OOHCA temperature management is based on the limited evidence available and the ILCOR guidelines, yet interpretation of this varies at even regional levels. Whilst tailoring care to individual patients is vital in intensive care, if temperature management is to be implemented then a standardized approach in duration and target temperature would allow trainees to implement and manage temperature control and reduce error in the care delivered. Whilst we await the arrival of the cardiac arrest centre in UK, ensuring equality in treatment where we can is of upmost importance.

Re-admissions in ICU- Does seniority of the receiving doctor matter?

Sonia Mason, Jagdish Sokhi and James Armstrong

West Middlesex University Hospital NHS Trust, London, UK

Abstract

Background: The 2015 GPICS guidelines state that ICU patients should be discharged between the hours of 0700 and 2200 using a “standardised handover procedure,” (1). These guidelines do not specify who should receive handover and manage the patient’s care, prior to a consultant review on the ward.

ITU patients are often complex, with significant co-morbidities requiring a daily consultant led multi-disciplinary team approach. To facilitate on-going management and prevent re-admissions, defined within 48 hours of discharge, effective handover is essential.

We propose that the rate of re-admissions will be reduced, by handing over to more senior doctors, (i.e. SpRs and above).

Methods: Retrospective data was collected over a period of 6 months in a busy district general hospital. Discharge summaries, patient's notes and ICNARC data were reviewed. Incomplete discharges were excluded. The re-admission rates, time of day and grade of doctor were then analysed.

Results:

47% (114/245) of discharges were appropriately completed with details of the handover including the time, destination and receiving doctor. Of these discharges, 94% (107/114) occurred within the specified guideline hours and 6% (7/114) were out of hours. There were 5 re-admissions, 4 of which had completed data.

Medical patients were discharged primarily to SHO/FY doctors 72% (25/61) in hours and out of hours 25% (1/4), with 4 re-admissions from in hours discharges. Patients discharged to surgical specialities were to SHO/FY doctors in 81% (42/52) of cases. There were no re-admissions for the surgical discharges.

Those discharged to medical SpRs had no re-admissions and there was a progressive increase in re-admission percentage with declining seniority.

Discussion: 47% of discharge summaries were appropriately completed. Our re-admission rate of 2% was the same as the national standard (1). From our data, there is a trend towards re-admission and declining seniority. This is likely to be heavily cofounded and underpowered, but efforts should be made to ensure the receiving doctor is competent to manage the patient, until senior input is available.

None of the surgical patients were re-admitted despite discharging to more junior doctors, compared to medicine. A possible explanation for this is that all surgical patients are reviewed daily by a SpR or consultant, on the ICU, and prior to discharge.

Medical patients were discharged primarily to SHO/FY doctors and had a higher re-admission rate. This is likely to be explained by the omission of daily senior medical reviews by the parent teams, prior to discharge. Therefore the medical junior may be the only doctor managing the patient, until the next SpR/consultant ward round.

Conclusions:

1. All patient discharged from ICU should have clearly documented handovers.

2. If a SpR from the receiving team is not available, the SHO/FY doctor should see the patient on ICU where handover can occur and concerns can be raised, prior to step down.

3. Handovers out of hours should be to the relevant SpR.

Single centre experience of High Frequency Percussive Ventilation (HFPV) in patients over 75 years of age

Astrid Van Lantschoot, Martijn Rongen, David Schramme, Margot Vander Laenen and Willem Boer

Ziekenhuis Oost Limburg, Genk, Belgium

Abstract

During HFPV (VDRâ-4, Percussionaire, Sandpoint, Idaho), high frequency oscillatory diffusive ventilation is superimposed on conventional tidal volume convective ventilation. Benefits include enhanced oxygenation and hemodynamics, as well as alveolar recruitment, while supposedly providing lung-protective ventilation. HFPV, as distinct from HFOV, which has fallen out of favor in the light number of recent negative studies, has been used in our centre as a rescue therapy in patients with respiratory (and ventilatory) failure for a number of years.

From 2009 until January 2018, 27 patients over the age of 75 (male 21, female 6) on conventional mechanical ventilation were treated with HFPV (total = 140, 19%). Median age was 79 (IQR4). On admission to ICU, mean SOFA score was 8.6 (±3.7), median APACHE II score was 18 (IQR 12). 14 patients suffered from chronic lung disease (52.

Two of 27 patients (7%) were switched to ECMO, one survived, one died on ICU. Twenty-eight day mortality was 48% (13/27), survival to discharge from hospital was 37% (10/27).

In survivors, conventional ventilation took place for a median of 39 hrs (IQR 105) before HFPV was initiated. HFPV was performed for a mean of 115 hrs (±74) and total invasive ventilation was performed for a mean 35 days (±23). Total admission time on ICU in survivors was a mean of 56 days (±28) and mean time of admission in hospital was 99 days (±43 days). There was no difference between survivors and non-survivors for age, presence of chronic lung disease pre-admission to ICU, severity of illness score at admission and ventilation and blood gas parameters on initiation of HFPV. Follow-up parameters such as pH, PO₂, PCO₂ and P/F at 2, 6, 24 and 48 hrs after initiation of HFPV did not differ significantly. Both SOFA (expected mortality 15–20%) and APACHE II scores (expected mortality approx. 30%) at admission underestimated subsequent mortality in this subpopulation. Median P/F ratio at initiation of HFPV was 73 (IQR 39), with only 7 of 27 (26%) having a value >100. Though no studies exist with a similar population, a comparison of ECMO vs conventional ventilation for ARDS (with muscle relaxants, prone positioning, recruitment maneuvers and NO) and similar P/F values, but a younger population (mean 56), demonstrated a mortality at 60 days of 35% vs 46% and at 90 days of 37 vs 47%. Another study comparing prone vs conventional, with P/F < 150 demonstrated a mortality 23 vs 41%, with a benefit for prone.

Conclusion: Mortality was high (63%) and survivors subsequently underwent protracted periods ventilation, ICU and hospital stay. The question remains if, in this subpopulation, a standardized ventilation protocol should favour prone positioning and ECMO, either in combination with, or ahead of HFPV.

This study was conducted after permission from the local ethics committee of Ziekenhuis

Oost-Limburg, Genk, Belgium.

When the Body is Out of Resources- Anorexia and Critical Care

Bethany Foster, Raluca Ene, Venkat Sundaram

Ysbyty Glan Clwyd, Rhyl, UK

Abstract

An interesting case is hereby presented, encompassing a critical illness secondary to a severe mental health disorder. The prevalence of Anorexia Nervosa is increasing with limited literature available for its management in Intensive Care.

Description: A 44-year-old lady was admitted to our hospital following a fall. She had a previous medical history of smoking, intermittent alcohol excess and anorexia nervosa since the age of 14. Her weight was 39 kg, with a Body Mass Index(BMI) of 16. Despite this she was living an independent life, working and driving just a few weeks prior to the incident. On arrival to A&E she was found to be in cardiogenic shock, have profound metabolic derangements and be hypothermic. Gross pitting oedema and multiple pressure wounds were noted on her extremities. She was admitted to the Intensive Care Unit (ICU) for vasopressor support and metabolic derangement correction. The following day she developed respiratory failure and was intubated and ventilated. After multiple failed attempts at extubation a tracheostomy was performed. Her stay in ICU was complicated by multiple blood product transfusion requirements and multi-drug resistant infections. Two weeks since admission decreased mobility was noticed in the left arm. A CT head revealed an acute infarct in the right mid parietal region and a subacute infarct in the left occipito-parietal region. It was also noted that she had advanced atrophy for her age and multiple chronic subdural haematomas. After a month in ITU she was discharged to the ward and is making slow recovery with ongoing nutritional issues.

Discussion: Anorexia is a mental health issue with multisystem physiological consequences. Its incidence is increasing, and more patients are requiring hospitalisation for treatment. The MARSIPAN guideline is a useful tool to manage critically ill patients suffering with anorexia nervosa. It highlights the concerns and management of refeeding and underfeeding and the importance of multidisciplinary involvement with psychiatry and nutritional teams. Management in a Specialised Eating Disorder Unit (SEDU) is preferable, however these are limited, with none in Wales.

There are many anaesthetic implications for managing anorexic patients with an increased risk of aspiration, cardiovascular instability and a high risk of drug errors. This group of patients are at of risk of pressure related necrosis and fractures. They often benefit from HDU/ICU monitoring.

Aknowledgements: Presented with written consent from patient’s next-of-kin.

Significant lactataemia following low dose Salbutamol nebulisation

Tim Martindale and Sonia Akrimi

St Richard's Hospital, Chichester, UK

Abstract

We report a case of a non-asthmatic intubated patient who, following a bronchospasm received three 5 mg nebules of Salbutamol which triggered a fulminant metabolic acidosis, generating a lactate of 7.1 mmol/l and causing a fall in her pH to 7.17. This is, we believe the first time this has been reported following such a low dose of Salbutamol. This side effect of salbutamol is under reported and clinically very important to recognise to avoid unnecessary and dangerous treatment escalations and iatrogenic patient harm.

The South West Point of Care Ultrasound (POCUS) Governance Survey A Regional Review of POCUS Utilisation and Application, Governance and Education with comparison between Specialties across Multiple Centres within the Peninsula Trauma Network

Jonathon Lowe and Colin Bigham

Derriford Hospital, Plymouth, UK

Abstract

Introduction: With the increasing application and utilisation of ultrasound throughout medical and surgical specialties, it is widely recognised as a direct extension of the bedside examination.¹ With increasing use of Point of Care Ultrasound (POCUS) comes an expectation of enhanced theoretical and clinical training, supervision and continuing professional development as would be expected with providing an ultrasound service.² The Royal College of Radiologists (RCR) has produced guidance for clinicians utilising POCUS and the requirement for robust quality assurances practices, echoed by the British Medical Ultrasound Society (BMUS)^3,4.

As such, the authors set out to review POCUS practices across the South West Peninsula NHS trusts. The aim, to review current POCUS governance as well as the provision of training, education and barriers to its use.

Method: A questionnaire was designed and distributed to the five NHS trusts looking to gather information from four different specialties; Anaesthetics, Intensive Care Medicine, Acute Medicine and Emergency Medicine.

Those clinicians who had access to a machine submitted the following; use as part of clinical practice, level of training, use of images for decision making, image review and storage. Those who did not have access to a machine declared if they felt there was a role for POCUS in their practice and if they had an interest in learning more.

Results: The questionnaire was received by 323 clinicians with 298 responses (response rate 92%).

97.3% of all respondents had access to POCUS with 83.9% overall stating regular use. The remaining 16.1% admitted to not using ultrasound but could see a role for POCUS if better training opportunities were available locally.

Of those using POCUS, only 13.1% had any form of recognised accreditation or qualification. The remaining 86.9% had variable experience that may include formal training but no formal accreditation. 50.3% of users made management decisions based on findings but only 47.6% would record clinical findings in the patient’s notes.

Of all POCUS users, 13.1% had recognised supervisors to review their images (in line with RCR recommendations) but only 9.4% saved images for review to their local imaging system.

Discussion: Although encouraging to see such high use, the adherence to RCR guidance is less so. Of those that do follow recommendations, the majority hold recognised ultrasound qualifications from an Emergency Medicine or Intensive Care background, which may give some explanation to the more robust governance given the level of training.

With an increasing role of POCUS, the need for increasing governance is evident from a patient safety perspective, but also an educational one. Clinicians want the opportunity to learn how to use and include POCUS in their clinical practice. For this to become a reality, it must be done safely.

This abstract suggests that there is significant work to be done in creating a robust governance system for users of POCUS but does show encouraging uptake in its utilisation in the bedside assessment. The idea of a national governance model or structure may support directorates in establishing and maintaining safe practice for both patients and practitioners.

Clearance of inflammatory mediators using EMIC®-2 in a patient with sepsis and AKI

Jennifer Reynolds¹, David Ochiltree¹, Kieran Donnelly² and Neil Crooks¹

¹Heartlands Hospital, University Hospitals Birmingham NHS Foundation Trust, Birmingham, UK

²University Hospitals Birmingham NHS Foundation Trust, Birmingham, UK

Abstract

Septic shock is a significant cause of mortality in critical care patients.¹ The release of pro-inflammatory cytokines during an episode of sepsis can lead to significant end-organ dysfunction such as Acute Kidney Injury (AKI). Modifications to haemofiltration techniques aimed at reducing the burden of inflammatory mediators remains an area of active research. We present the first case report in the UK, to our knowledge, describing the use of the EMIC®2 (Enhanced Middle Molecule Clearance)² filter in a patient with severe sepsis secondary to necrotising pancreatitis.

The EMIC®2 filter utilises a modified pore size (40 kDa) vs the 30 kDa pore size used in standard continuous veno-venous haemofiltration (CVVH) circuits, in order to increase the clearance of “middle molecules” such as cytokines, interleukins, myoglobin and serum free light chains (SFLC), whilst retaining essential plasma proteins, such as albumin. This enhanced mechanism is based on the concept of emulating the function of the glomerular basement membrane in normal physiological renal function.²

We report a case of a twenty-eight-year-old female with severe acute pancreatitis, who subsequently developed intra-abdominal sepsis and multi-organ failure secondary to pancreatic necrosis during her 39 day stay on critical care. Due to refractory septic shock, the standard CVVH filter was changed to a super high flux membrane, EMIC®2. Following this intervention, inotropic and vasopressor requirements significantly reduced (adrenaline and vasopressin stopped, 50% reduction in noradrenaline) and ventilatory parameters improved, with sequential drop in inflammatory markers over the subsequent 4-day period. There was no significant change in the serum albumin level (16–18 g/L) during the time the EMIC®2 membrane was in use.

Following this interval improvement in the patient’s condition, 5 days after the introduction of EMIC®2 membrane, escalating vasopressor requirement and C-reactive protein prompted a computed tomography scan of the abdomen which demonstrated further retroperitoneal collections. Despite percutaneous drainage of these collections, continued deterioration in the clinical course lead to laparoscopic examination, which revealed widespread bowel ischemia and necrosis. In view of this finding treatment was withdrawn, palliative care was instituted, and the patient passed away.

We discuss the use of a modification to standard renal replacement therapy which may have a role in future practice, and examine the data surrounding the ability of these membranes to clear molecules implicated in the systemic inflammatory response syndrome. Our case report highlights the use of the EMIC®2 membrane in a patient with refractory septic shock. Although many factors may have influenced the clinical course, the use of the membrane was associated with a reduction in vasopressor requirements and ventilatory support. Serum albumin levels were stable during this period.

By enhancing middle molecule clearance whilst preserving plasma albumin levels in patients with cytokine dysregulation and systemic inflammatory syndromes, it may be possible to demonstrate improvement in haemodynamic status, oxygenation indices and organ dysfunction. Larger scale studies into super high flux haemodialysis membranes are required to establish if they have a potential role in the management of the critically ill patient with multi-organ failure, and if they confer any outcome benefit in this patient group.

A six-month quality improvement project to improve adherence to optimal physiological parameters for patients in Aintree University Hospital Critical Care Unit

Jun Jack Wong, Eleanor Houghton, Phoebe D'Orton-Gibson, Emma Fadden, David Whitmore and Ben Morton

Aintree University Hospital, Liverpool, UK

Abstract

Introduction: Continuous monitoring of critically ill patients enables intensive care unit staff to titrate therapy to maintain physiology within optimal parameters. While the optimisation of patients’ physiology is fundamental in maximising potential recovery, strategies to reduce the risk of overtreatment and iatrogenic harm are also required.

Methods: We completed a six-month quality improvement project to evaluate adherence of ICU patients’ physiological parameters to evidence-based targets. Aintree University Hospital has a combined HDU (12 beds) and ITU (11 beds) and admits approximately 1400 patients per year. Standard physiological parameter targets were agreed by the critical care consultant group (deviations permitted for specific patient conditions at medical discretion and instruction):

1. Oxygen saturations 94–96%¹

2. Tidal volume (TV) of <450 ml for male and <400 ml for female while on pressure controlled ventilation.²

3. Mean arterial pressure (MAP) of 65–70 mmHg.³

4. Richmond Agitation Score for sedation of 0 to −2.⁴

5. Urine output >0.5 ml/kg/hour.⁵

6. Blood glucose level 6–10 mmol/L.

We conducted a snapshot audit for at least 20 patient episodes during each ‘Plan, Do, Study, Act’ cycle. Data was collected electronically using the KoboTool Box data collection platform (https://www.kobotoolbox.org/). Subsequently, data was analysed using STATA V13.1 (StataCorp, USA).

Results: Data was collected from January 2018 through to August 2018 (Table 1). Baseline data obtained in January 2018 reflects unit performance prior to any intervention followed by monthly cycle of interventions implementation and further data collections to reassess unit performance over time.

The interventions (Table 1) are:

A. Introduction of nurse champions to advocate for achieving standard parameters to ensure patient safety.

B. Education of nursing staff during the ward round.

C. Reminder memo with standard parameters attached to each nursing work station.

D. Display audit results in staff room to increase awareness.

E. Education reinforcement of nursing staff by nurse champions.

F. Incorporation of standard parameters target into patient observation chart.

Table 1. 6 month data on compliance of standard physiology parameters

Overall, the interventions employed did not lead to any meaningful statistically significant improvement.

Discussion: Through this quality improvement project, our team recognise that it is extremely challenging to influence change in unit culture and human behaviour especially when it involves large number of stakeholders (nurses and doctors). Therefore, further measures need to be devised in the near future. We will continue our project for the coming months with the aim of maintaining and improving the compliance of standard physiology parameters.

NEWS score of 10 or more is a Medical Emergency: Analysis of Peri – arrest calls after the introduction of a Detiriorating Patient Pathway

Meenakshi Agarwal, David King and Nishit Sud

Southend University Hospital, Southend on sea, UK

Abstract

Objective: Following introduction of a Deteriorating Patient Pathway recommending periarrest call if NEWS 10 or above, we audited number and outcome of periarrest calls

OUTCOME OF PERI-ARREST CALLS	NUMBER OF PATIENTS	PERCENTAGE OF CALLS
ITU ADMISSIONS	16	27%
STEP - UP  (HDU, ARCU)	13	22%
TEP/DNAR	25	42%
NO CHANGE	10	9%

ITU - Intensive Therapy Unit; HDU - High Dependency Unit; TEP - Treatment Escalation Plan; DNAR - Do Not Attempt Resuscitation.

Method: Following introduction of a Deteriorating Patient Pathway in October 2017 we audited all adult cardiac arrest and periarrest calls in one calendar month (January 2018), a time of high demand for critical care. We used our resuscitation audit database to identify cardiac and periarrest calls and searched for Intensive Care Unit admission or step up in care to another location. We also looked for any treatment limitation decisions made such as Do Not Resuscitate order (DNAR) or a Treatment Escalation Plan prohibiting Intensive Care admission. We calculated 30 day mortality for patients with periarrest calls. We also compared volume of both cardiac arrest and periarrest calls with a historical control group from January 2017.

Results: In January 2018 there were 115 calls identified. These consisted of 32 cardiac arrest and 64 periarrest cals, with 2 paediatric arrest calls and 17 cardiac arrest cancelled calls. This is an increase in both cardiac and periarrest calls compared with January 2017 which had 63 calls in total (17 cardiac arrest and 46 periarrst).

There were 64 periarrest calls in 60 patients, with some patients having multiple calls.

OUTCOME OF PERI-ARREST CALLS (TOTAL NUMBER = 64)

30 day mortality for periarrest calls was 35 out of 60 patients (58%)

Conclusion: Introduction of a Deteriorating Patient Pathway recommending periarrest call for patients with NEWS of 10 or above was associated with an increased number of periarrest calls compared with historical controls. Most calls resulted in a stepup in care or treatment limitation decision.

We believe that these results support the introduction of our Deteriorating Patient Pathway. This pathway effectively converts the cardiac arrest team into a Medical Emergency Team with NEWS 10 or above as calling criteria, and may lead to more timely Intensive Care admission and help prevent cardiac arrests.

Developing a Pan-London Intravenous Medication Competency for Critical Care Nurses

Chris Hill¹, Gezz Van Zwanenberg²

¹North East North Central London Critical Care Network, London, UK

²North West London Critical Care Network, London, United Kingdom

Abstract

Administering Intravenous (IV) medication to critically ill patients has significant potential for causing harm. To prevent this clinical staff undertake rigorous education and assessment.

Critical Care Services in London employ a large clinical workforce, with significant turnover requiring repeated allocation of clinical educators for education and assessment of staff to enable administration of IV medication (alongside other essential training). Units have their own processes for doing this and as staff move between units their competencies require reassessment each time. These practices present opportunities to reduce duplication of process and resources. By standardising competencies, education and assessment this will eliminate duplication, release time to care and improve patient safety.

The North West London Critical Care Network (NWLCCN) ‘Clinical Applied Skills Passport’ (CLASP) project began in 2016 aiming to reduce: i) variation in standards; ii) duplication of effort and resource utilisation in delivering skills training and competency assessment. CLASP initially focussed on transfer equipment competencies.

The NWLCCN established a CLASP task and finish group with colleagues from the Pan-London Critical Care Nurse Education group. The aim is a Pan-London IV medication Critical Care Competency (PLIVCCC), using the ‘Capital Nurse’ Systemic Anti-Cancer Therapy Competency Passport ¹ as a replicable model.

Principal concepts agreed were:

• Staff would meet Trust requirements for giving IV medication prior to accessing the PLIVCCC.

• Medical device training would remain separate.

• Duplication of work should be minimised.

Initially the group looked to merge existing IV medication competencies from critical care units across London, however the variation was significant, and it was decided to gain consensus using a modified Delphi approach (modelled on the process NWLCCN successfully used to deliver a standardised transfer bag ²).

Stage 1 (The long list): Was created using the existing competencies and respondents from across London confirming items they felt should be included. These were merged and grouped to form the long list.

Stage 2: The resulting long list was distributed widely as a survey for respondents to rate the important of each item. Pilot survey return rates were disappointing, and the survey mechanism was changed from Excel ® spreadsheets to a Typeform ® survey. This elicited an excellent return rate from across all critical care units Pan-London.

Stage 3: These survey results were used to inform agreed learning outcomes.

A key aim is to reduce duplication of work, so the learning outcomes were aligned with the medication administration section of the National Step 1 Practice Assessment Document ³ and duplicated items removed. We added the prerequisite for completion of this section of the Step 1 PAD before accessing the PLIVCCC.

A draft of the PLIVCCC was produced for piloting and further review by members of the Pan-London nurse group.

This project has demonstrated successful large scale collaboration from a working group spread across London with the use of a modified Delphi process to gain consensus. The PLIVCCC aims to reduce duplication, improve resource utilisation and patient safety. The competency and Passport are being launched Pan-London on 23^rd January 2019.

Haemorrhage, transfusion and coagulation monitoring in elective and emergency abdominal aortic aneurysm repair – the experience of a large district general hospital in Cornwall

Christopher Pritchett¹, Jefferson Wong², Zulai Ibrahim², Gary Matthews¹ and Catherine Ralph¹

¹Royal Cornwall Hospital, Truro, UK

²Peninsula School of Medicine and Dentistry, Exeter, UK

Abstract

Introduction: Open repairs of abdominal aortic aneurysm (AAA) are at high risk of blood loss and allogeneic blood transfusion. Stoneham et al. have demonstrated exceptionally low allogeneic blood product use in elective AAA repair through effective use of intra operative cell salvage (ICS) and thromboelastography (TEG).¹ In 2015 Mallett and Armstrong proposed monitoring coagulation throughout surgical procedures at high risk of blood loss as opposed to waiting for obvious bleeding problems to become apparent.² In principle, standard laboratory tests (SLTs) can be utilised. Prolonged turn around times often limits their usefulness in making transfusion decisions during acute haemorrhage.³ In this retrospective audit of 80 open and endovascular AAA repairs, we established transfusion practice and use of intraoperative SLTs to assess their use in guiding blood transfusion.

Methods: Using our electronic theatre database we identified all patients undergoing AAA repair. We included endovascular and open procedures, elective and emergency. Patient details were used to retrospectively extract data from the hospital blood bank, cell salvage database and pathology results system. Data collected included: Type of procedure, volume of intra operative cell salvage (ICS) blood re-infused, use of and number of blood products transfused. SLT data collected included FBC and coagulation results. Timing of these results were noted to establish their use pre, intra and post operatively.

Results: 80 patients underwent AAA repair over a 21 month period. 43% were open repairs. ICS was used in 69% and 95% of open emergency and elective repairs respectively. Re-infusion of ICS blood averaged 807mls in emergency and 778mls in elective cases. 77% of emergency and 9.5% of elective repairs received allogeneic blood products. Of those transfused, RBC use averaged 4.1 and 1.5 units for emergency and elective repairs respectively. FFP use averaged 2.67 and 4 units for emergency and elective respectively. Intra operative SLTs use (FBC and coagulation) was 11% and 6% respectively in emergency cases and 5% and 10% respectively in elective cases.

Conclusions: ICS use in emergency AAA repair was lower than expected. This may explain why 56% of this cohort received donor blood products. Use of SLTs intra-operatively was low. This suggests FFP use was transfused on clinical grounds alone. Following the impressive results of Stoneham et al. looking at elective open AAA repair, it seems that increased use of ICS a and point of care viscoelastic tests may be part of a solution to improve transfusion practice.

The Impact of Bundle Care in Emergency Laparotomy patients in the Intensive Care setting: a Quality Improvement Project

Harry Wadman, Emma Saunsbury

Royal United Hospital, Bath, UK

Abstract

Patients undergoing emergency laparotomy have the second highest mortality rate in England and Wales. [1] One of the key standards resulting from the National Emergency Laparotomy Audit (NELA) to facilitate improvement in their post-operative care was direct critical care admission in higher risk cases. However, no formal protocol exists to indicate specific care goals in this setting.[2]

The objective of this quality improvement project was therefore to develop, implement and evaluate a care bundle for these patients, with a primary focus on maximising enteral nutrition delivery. The finalised ‘Emergency Laparotomy Pack’ included compulsory use of a SACETT® endotracheal tube to reduce ventilator-associated pneumonia risk (as part of bundle care), a five lumen central venous catheter (CVC) to provide appropriate critical care access, a nasojejunal (NJ) tube to facilitate sufficient nutrition delivery and a nasogastric (NG) tube for aspiration monitoring. Specific outcome measures included compliance with bundle components and impact on nutrition in terms of protein and carbohydrate delivery. The Emergency Laparotomy Pack components were combined into pre-formed sealed packs in dedicated emergency theatre trolleys, and their introduction was supported by an education programme and promotional materials.

Data was collected over 4 months and the sample size included 43 patients. Compliance with each Emergency Laparotomy Pack component increased from 0% to 71% for NJ tubes, 40% to 93% for NG tubes, 33% to 100% for endotracheal tubes and 80% to 93% for CVCs. The ratio of NG:NJ patient fed days increased as NJ tube compliance increased (from 6.33:1 to 1:2), however protein and carbohydrate delivery remained higher in NG fed patients, but this may have been due to a small sample size.

This quality improvement project demonstrates that education surrounding post-laparotomy care can result in increased compliance with bundle intervention. Our primary objective of increasing NJ tube use and post pyloric feeding was met. Although our project did not show improved carbohydrate and protein delivery, it has been well documented that post pyloric feeding can improve the delivery of calories in the critically unwell compared to standard NG feeding. [3] The “Emergency Laparotomy Packs” have now been included on the emergency theatre trollies on a long term basis as a result of this project.

Obstetric admissions to critical care: a retrospective audit

Sophie Lane, Angus Royal, Christopher Lockie and Behrad Baharlo

Chelsea and Westminster Hospital NHS Foundation Trust, London, UK

Abstract

Objective: To review the incidence, characteristics and outcomes of obstetric patients presenting to an eleven bed critical care unit in a teaching hospital. Retrospective review of critical care admissions over a three year period. Data was compared against an Intensive Care National Audit & Research Centre (ICNARC) report commissioned by The Obstetric Anaesthetist’s Association (OAA) and a standard outlined in the Royal College of Anaesthestists (RCoA) report ‘Care of the critically ill woman in childbirth; enhanced maternal care’.

Methods: Data was retrieved for all currently or recently pregnant (within 42 days) critical care admissions between April 2015 and August 2018 from the Acubase database software. April 2015 represented the first date that the unit participated in ICNARC data collection.

Results: During the period between April 2015 and August 2018 there were 37 obstetric admissions to the ICU. This represented 1.87% of all general admissions and 12.29% of female admissions aged between 16–50 years. The mean age was 35.4 (6.3 SD) years.

Examining source of admission, 94.6% were from the obstetric unit, one from the ward and one from the emergency department. The median gestation of patients currently pregnant was 26 (IQR 25.3, 28.3). The median gestation of recently pregnant patients at the time of delivery was 36 (IQR 33, 38.5). Of these patients, 58.1% underwent a caesarean section (n = 18), 35.5% assisted delivery (n = 11) and two underwent a termination of pregnancy. Four patients underwent a hysterectomy at the time of or since delivery.

The most common reasons for maternal admission to ICU were antepartum and postpartum haemorrhage. Level 3 care was required in 48.6% of admissions (n = 18). The mean number of days of level 2 care provided was 2.06 and level 3 care 2.17. The mean APACHE score at admission was 9.62 (SD 5.22). This is comparison to the non-obstetric mean for the same period for all patients of 13.62 (SD 7.6) and for all female patients aged between 16–50 years of 9.02 (SD 5.8) The mean length of stay in the unit was 2.29 days (median 1.49, IQR 0.94,2.79). There were no maternal deaths over this period.

Conclusions: The RCoA noted increasing rates of illness surrounding childbirth, as a consequence of increasing maternal age, obesity and other co-morbidities. ICNARC demonstrated a critical care admission rate of 2.4 per 1000 maternities. Therefore, critical care providers must be able to participate in the recognition and management of acute obstetric emergencies. The RCoA report highlights standards expected for critical care services. They stated that there should be a lead consultant for maternal critical care and that all women should be offered a postnatal review within input from a clinician with experience in critical care follow-up. These measures are currently not in place in our critical care unit. The current audit identifies that a significant minority of female admissions aged between 16–50 years are obstetric, most presenting with specialist problems. The current audit emphasises the need for a clear pathway to critical care services and ongoing multi-disciplinary working between obstetrics and critical care.

aSSI-SST: Critical care assisted Specialty Specific In-Situ Simulation Training – A novel model for the evolution of critical care outreach services and training

Sampath Weerakkody, Samatha Petty and Monica Trivedi

John V Farman Intensive Care Unit, Addenbrooke’s Hospital, Cambridge, UK

Abstract

Background: The last two decades has seen a shift from resuscitation at the point of cardiorespiratory arrest to early identification and adaptation of resuscitation strategies to prevent cardiac arrest.¹ The advocacy for rapid response teams and the various iterations of early warning scores have driven this shift in focus.² The genesis of this study began with a number of critical incidents that occurred on specialist wards despite the presence of a well established rapid response team. The success of such teams has lead to these services being overstretched with the possibility of non-critical care staff becoming de-skilled in the recognition and management of the critically ill patient. Most UK hospitals have increasingly specialist wards, and the majority of resuscitation courses available offer only generic resuscitative training, with a recognised attrition of knowledge and skills between courses.³

Aim: To create a novel specialty specific in-situ simulation training program (aSSI-SST) to address the needs of individual speciality teams, that can be delivered frequently, with a focus on high yield training in clinical scenarios relevant to the specialty.

Methods: A group of 29 respiratory nurses at Addenbrooke's University Hospital were enrolled on the aSSI-SST course, divided into 3 cohorts. In consultation with Intensive Care and Respiratory physicians, physiotherapists and nursing teams, a list of key scenarios were constructed including the management of type 1 and type 2 respiratory failure, with a particular focus on non-invasive ventilation, a blocked chest tube, a displaced tracheostomy tube, obstructive sleep apnoea and opiate toxicity. Prior to the course, all candidates were initially given a scenario-based multiple choice test (MCQ) for objective knowledge assessment, and a self-assessment questionnaire evaluating perceived knowledge in key areas and its application. Pre-course testing was followed by a teaching program consisting of presentations and practical workshops. Following this, candidates performed four to six simulations and then completed the post course assessments.

Results: The average pre-course MCQ scores were high (81%), with no post course improvement. Subjective scores of knowledge and application improved by 14–17% after aSSI-SST.

Summary of results of aSSI-SST course

Abstract

Discussion: These results suggest that despite excellent knowledge, practical application in crisis settings was poor, as ascertained by trainer feedback. The candidates confidence seemed to improve dramatically after the course. Further objective video assessment and analysis of simulator performance would help to support the self-assessment results. This provides a basis for expanding this model to other specialties with targeted simulation training. We feel that a multidisciplinary approach and engagement of parent teams to eventually deliver this training is a key aspect to its potential success, thus reducing the burden on critical care services.

Conclusion: We believe the aSSI-SST concept is a novel approach and an evolution of successful rapid response teams, providing strengthening of critical care service delivery throughout hospitals, tailored to individual specialties. Its aim is to contribute to reducing the severity and frequency of critical incidents, as well as improving staff performance, confidence and competence.

	Objective Knowledge	Self- assessment of knowledge	Self- assessment of application
Improvement/  Deterioration  (%)	-2	+17	+14

Packed Red Cell transfusions in the Intensive Care Unit: a quality improvement project on prescribing and administering practices

Nida Kalyal¹, Chris Ward¹, Akshay Ankuli¹, Tatyana Bolonenkova¹, Dan Harding¹ and Ashraf Molokhia²

¹Department of Intensive Care Medicine, Queen Elizabeth Hospital, London, UK

²Department of Intensive Care Medicine, University Hospital Lewisham, London, UK

Abstract

Introduction: Transfusion of packed red cells (PRC) is an important treatment option for patients requiring intensive care but, like all treatments, it is not without risk. These patients, although more sensitive to anaemia¹, are also at increased risk of transfusion-related complications. Since the 1999 TRICC trial², restrictive strategies have gained popularity and shape current guidelines³ on transfusion. As a result, the FICM endorses guidelines that advise a transfusion trigger of 70 g/L or less in most critically unwell patients.

Methods: We audited blood prescribing and administering practices in our intensive care unit over a 31-day period, using proformas placed in blood prescribing forms. All transfusions of PRCs were logged over this time and transfusion trigger, post-transfusion Haemoglobin (Hb), and whether Hb was checked between units was recorded, in addition to other supplementary information. We re-audited over a 31-day period after the introduction of a clinical guideline.

Results: 25 transfusion events were recorded in the initial period. We identified that only 12% of patients transfused for a low Hb had a transfusion trigger of 70 g/L or less and 12% who were transfused for a low Hb had a Hb of 80 g/L or more.

Following our initial data collection, we developed a guideline for blood transfusion in adult intensive care patients. This was based on best practice guidelines from NICE and BCSH, endorsed by FICM, and was reviewed and approved by our intensive care medical and nursing teams. We re-audited after the introduction of the guideline, with the results summarised below:

	Pre- Guideline (Cycle 1) n = 25	Post- Guideline (Cycle 2) n = 33
Mean Hb g/L (All transfusions)	76.1	70.0
Mean Hb g/L (Low Hb only)	74.4	67.6
Median Hb g/L (Low Hb only)	74	69
Mode Hb g/L (Low Hb only)	74	69

We analysed our data using the unpaired t test and found the difference between pre and post-guideline trigger Hb to be statistically significant (p = 0.0017).

Conclusion: We identified a mismatch between best practice and our current practice, but this was easily rectified with a guideline and teaching sessions for nursing and medical staff, with a statistically significant improvement in our practice. We hope that this achieves an improvement in patient outcomes, as it has been demonstrated in the literature that restrictive transfusion strategies are safer than liberal ones.² It should also lead to financial savings, as a single unit of PRC costs £121.85.⁴

Decanting the Salt out of the (Data) Swamp

Romit Samanta¹, Emma Rocheteau², Olaolu Oloyede³ and Charlotte Summers¹

¹Division of Anaesthesia, Department of Medicine, University of Cambridge, Cambridge, UK

²University of Cambridge, Cambridge, UK

³Public Health England, London, UK

Abstract

Introduction: Raised chloride levels are a well-recognised cause of metabolic acidosis due to a reduction in plasma strong ion difference. The cause is almost always due to the administration of 0.9% sodium chloride solutions with only a few exceptions. The literature is conflicting on the implications of hyperchloraemia on patient outcomes, with well-conducted randomised controlled trials showing both harm¹ and no harm² when comparing saline with balanced solution administration.

Methods: We investigated the outcomes of patients admitted to ICU with hyperchloraemia using the open access database eICU, which is an open resource developed and maintained by the MIT Laboratory for Computational Physiology containing data on over 195,000 patients.³ Patients were stratified by the diagnostic groups sepsis, trauma and abdominal surgery (prior to ICU admission). We excluded unlikely extreme and erroneous values using quantiles of the data based on their respective distributions, chronic dialysis patients and surgical procedures after the first 24 hours in ICU. Chloride levels were corrected for volume status to eliminate hypovolaemic causes (Cl_corrected = Cl_measured x (Na_normal/Na_measured). We looked at the outcomes: death, renal replacement therapy, change in creatinine from baseline and KDIGO stage, using APACHE IV score as a covariate. Statistical analysis with linear and logistic regression, odds ratios with p < 0.05 significance using R software. Details of our SQL queries and methods in R can be viewed on Github.⁴

Results: The lab results of 160,759 patients contained values for chloride levels during the first 24 hours of admission to ICU. 33,824 patients (20.0%) had an admission chloride level of >110 mmol/L. There were 26,815 patients with sepsis, 10,957 with trauma and 2,816 post abdominal surgery.

Patient's admitted to ICU with a hyperchloraemia had an increase in KDIGO acute kidney injury staging score (OR 1.09, 1.06–1.13, p < 0.001) over the first five days of their admission. There was an association with increased ICU mortality amongst patients admitted with trauma (OR 4.1, 1.99–8.5, p < 0.001) and abdominal surgery (3.75, 1.07–13.1, p = 0.04) but not sepsis (1.02, 0.73–1.41, p = 0.93).

Discussion: Raised admissions levels of chloride in ICU patients is a common finding and is associated with poor patient outcomes, especially following abdominal surgery or trauma. The lack of a significant difference amongst septic patients likely reflects their inherent heterogeneity, diagnostic uncertainty and how the database has been coded for a diagnosis of sepsis. Large ICU databases provide an opportunity to answer research questions by combining a pragmatic approach and real-world data, once allowances have been made for data quality.

Figure 1.

Forrest plot of odds ratios of outcomes for patient groups.

Morale with a tail – Initiating pet therapy in Critical care

Kate Tantam, Emily Putt, Richard Freeman and Moira Ford

Plymouth Hospitals NHS Trust, Plymouth, UK

Abstract

Therapy dogs have long been welcome in hospitals to support patients and their loved ones. These visitors are often used with less unwell hospitalised patients, but their presence in Critical Care is new for many hospitals. University Hospital Plymouth NHS trust is one of a handful within the UK that is embracing this innovative approach to supporting recovery and rehabilitation.

Treatment in Critical Care is extremely stressful both physically and psychologically. This is an issue for both the patient being treated and the relatives and carers observing and supporting the patient. The collation of symptoms that affect both patients and their carers after Critical Care have been collated into a syndrome – Post Intensive Care Syndrome (PICS) (Needham et al 2012).

The team here in Plymouth are exploring methods to support and treat the psychological distress that some patients can experience after critical illness. Therapy dogs offer the patients an alternative route to engage with rehabilitation and recovery and they are challenging the view that recovery is just a physical process. Mood and pain scores are improved when patients are exposed to Pet Therapy and engagement in rehabilitation has also been found to improve (Hosey et al 2018). Moira and Hovis are key to the psychological recovery of patients and their loved ones in Critical care. They offer non-medical support that for many is hugely important.

The Critical Care team have only just welcomed Hovis and Moira to the Rehabilitation team. They started working with us in the autumn of 2017. This has made analysis of their impact challenging and much of the evidence of their work is qualitative and reflective. Feedback from relatives has been incredibly positive, so positive that Moira and Hovis are increasing their time in the department from twice weekly to three times a week.

Patients and their loved ones report that they love seeing the team, and that Hovis and Moira are able to help distract and engage patients in a way that clinical staff may not. Animal therapy in some cases includes a physiotherapist, pairing physical rehabilitation with the dog visit to achieve specific functional goals. This was evident recently with a patient who was refusing to engage in physiotherapy. When the patient learned that the therapy involved taking Hovis for a walk, they jumped at the chance and walked further than they had ever walked since admission.

NB As this is a novel and innovative concept in Critical care, there is no formal evidence available in the UK regarding the use of PAT therapy teams in ICU. We are currently collecting formal qualitative feedback to be able to actively support PAT teams to integrate into other Critical care areas in the UK. We would like the opportunity to present this work as a poster.

Improving Patient Safety through an Emergency Call Safety Huddle

Shuaib Quraishi and Claire Rowley

Surrey and Sussex Healthcare NHS Trust, Redhill, UK

Abstract

Communication amongst members of the emergency team is integrally linked to patient safety. ¹ The need to promptly identify and manage the acutely unwell patients is key towards preventing harm to patients. A short daily meeting can help save lives by helping emergency teams to work together more effectively.

At Surrey & Sussex NHS Trust we have approximately five to eight emergency calls in a 24-hour period. These are composed of Medical Emergency Team (MET), which is composed of the medical registrar, senior house officer, foundation doctor and a critical care outreach nurse (CCOT). A cardiac arrest team is composed of the MET team as well as an anaesthetist. In the past the emergency team would initially meet over an unwell patient unaware of who each other was, what role they played and what was expected of them. It is known that there is a 1 in 400 million chance of the same team working together again. ²

The Safety Huddle has been a part of the culture of improving patient care at SASH since October 2016. Members of the cardiac arrest and medical emergency teams meet each other, roles are allocated every morning and learning from previous emergencies is discussed. Roles and training needs are documented through a standardised checklist on a daily basis. This is in order to create effective team working and improve patient safety.

At SASH we wanted to elicit whether the safety huddles was actually intending to serve its purpose. A qualitative SurveyMonkey was sent to medical and nursing staff who had attended the safety huddles. We had 29 responses from Nurses (CCU and CCOT), medical registrars, and junior doctors (SHO and Foundation). A thematic analysis of free text comments was undertaken and the following themes were identified.

1. Structure of the team. 100% of respondents found the huddle to be a useful. It identified and allocated roles and created familiarity between team members

2. Improve team working. 100% felt the huddle improved team working.

3. Patient safety. 91% of respondents felt patient safety was improved. This is through increased efficiency during emergencies, effective team working, better organisation and early involvement with critical care.

4. Identification of learning needs. 87% felt learning needs that were identified at the safety huddle had been addressed.

The questionnaire also asked where improvements could be made and these were as follows.

1. Night safety huddle. 72% would like to introduce an emergency huddle for the night team.

2. Debriefing. A debrief session would be useful for feedback on learning from emergencies.

Our Medical Emergency Team audit for 2018 has demonstrated that we have made an improvement in patient outcomes by an increase in patients who made an immediate improvement (79% in 2018 from 61% in 2017) versus the patients who made no improvement immediately post MET call (4% in 2018, 21% in 2017).

This suggests that by implementing the safety huddle we may be working more effectively as a team resulting in improved patient outcomes.

Investigating the decision-making process surrounding treatment escalation to intensive care and the effect of a decision-making tool using OSCE-style scenarios

Adam Boulton, Hisham Omer and Christopher Bassford

University Hospital Coventry, Coventry, UK

Abstract

The decision to admit a patient to the intensive care unit (ICU) is complex and involves consideration of numerous factors. The use of admission algorithms has been problematic for clinicians and may not be appropriate for this complex decision.¹ As such, there is no standardised process for considering ICU admission and few groups have, to date, analysed this decision-making process.² There are multiple barriers to investigating the decision-making process thoroughly and robustly in a clinical setting.³ Furthermore, in-depth analysis into the efficacy of decision-making tools is also troublesome in clinical practice. Simulated objective structured clinical examination (OSCE) scenarios may allow assessments of real-time thought processes, including participant analytical and evaluative skills.^4,5 Structuring the decision-making process for ICU admission may help to improve trainee confidence, comfort, and facilitate communication between clinical teams. This group have previously developed an evidence-based decision-making tool/support framework. This study sought to investigate the ICU admission decision-making process using OSCE-style scenarios and to evaluate the impact of a decision-making tool.

Two OSCE-style scenarios were used; each of these involved a patient whose suitability for admission to the ICU could be considered borderline. A family member of each patient was simulated. Participants were intensive care medicine and anaesthetic registrars with experience of making ICU admission decisions. Participants completed one scenario, followed by training with the decision-making tool and subsequently a second scenario. Data was collected by a mixed methods design using clinical note entries, post scenario questionnaires, and semi-structured interviews.

On visual analogue scales of 0–10, participants identified OSCE-style scenarios as being realistic (mean 7.6, SD, 1.2, N = 23) and they felt they had learnt from the experience (mean 7.4, SD 1.2, N = 22). Many of the participants volunteered that the decision-making tool training had provided a useful structure to their decision-making process and one participant felt that the training should be delivered regionally.

Initial results report that OSCE-style scenarios were viewed as realistic and a positive learning opportunity by trainees. Analysis of data from semi-structured interviews, post scenario questionnaires, and clinical note entries is ongoing and is allowing detailed investigation of the decision-making process and the influence of the decision-making tool training. Our experience and initial results support the use of OSCE-style scenarios as a feasible method of investigating decision-making processes for ICU admissions and examining the impact of a decision-making tool. Future studies of decision-making interventions, such as educational tools or training, should consider this type of methodology to evaluate their impact.

An audit of the administration of rFVIIa (NovoSeven) as an emergency haemostatic agent for intractable bleeding in cardiac surgery at a cardiothoracic quaternary centre

Lucy Miller¹, Martin Besser² and Stephen Webb³

¹West Suffolk NHS Foundation Trust, Bury St Edmunds, UK

²Addenbrooke's Hospital, Cambridge, UK

³Royal Papworth Hospital, Papworth Everard, UK

Abstract

Activated recombinant factor VII (rFVIIa) is increasingly used off-license as an effective haemostatic agent for refractive bleeding from cardiac surgery. Its use has also been linked to increased prevalence of thromboembolic events.

Moderate to severe haemorrhage is a common post-operative complication in cardiothoracic surgery. Treatment for intractable bleeding from cardiothoracic procedures is usually re-exploration surgery and large quantities of resuscitation fluids (including packed red blood cells, platelets, fibrinogen, and cryoprecipitate) and licensed haemostatic agents.

Activated recombinant factor VII (rFVIIa) (NovoSeven), a synthetic haemostatic agent licensed for treatment of haemophilia, has increasingly been used off-licence to control refractory haemorrhage post cardiothoracic surgery. Despite reports of success, a link between use of NovoSeven and increased morbidity has also been raised. Adverse outcomes include increased mortality, renal failure and thromboembolic events.

Meta analyses focusing on use of NovoSeven for treatment of bleeding in cardiac surgery have found a significantly increased rated of stroke following administration of rFVIIa. A proposed link between cardiothoracic surgery and the development of DIC has been hypothesised, leading to increased risk of thromboembolic events following administration of NovoSeven. Given this hypothesis, it is essential that NovoSeven is administered according to clinical guidelines.

This paper presents the results of an audit of the administration of NovoSeven compared to the stated trust guidelines. 20 patients were given NovoSeven at a leading cardiothoracic quaternary centre due to intractable bleeding during cardiac surgery. Systematic analysis of patient pre-administration prothrombin time, activated partial thromboplastin time and platelet numbers were undertaken and compared to the guidelines.

Use of a 1:1 Ratio of Fresh Frozen Plasma to Packed Red Blood Cells During Early Haemorrhagic Resuscitation in Severe Trauma: A Literature Review

Lucy Miller

West Suffolk NHS Foundation Trust, Bury St Edmunds, UK

Abstract

Trauma induced coagulopathy (TIC) has been identified in over 30% of severe haemorrhage cases presenting to hospital, with a similar value predicted for military casualties. Initial stages of TIC have been linked to increased morbidity and mortality, highlighting an area of vital importance in both civilian and military trauma responses.

The condition is exacerbated by dilution of coagulation factors during early fluid resuscitation. Fluid absent in clotting proteins is administered with plasma-poor packed red blood cells (PRBC), without additional coagulation products.

A growing body of evidence supports the use fresh frozen plasma (FFP) to deliver coagulation factors, at a 1:1 FFP/PRBC ratio in early fluid resuscitation following severe haemorrhage. Various studies have indicated that a 1:1 ratio is significantly beneficial compared to protocols using a lower ratio of FFP/RBC. This practice is already adopted in UK military practice, but not yet in civilian pre-hospital treatment of trauma.

Watts et al, 2016 (DSTL Ltd, Porton Down) present a comparison of fluid resuscitation protocols in a porcine model of severe trauma. Three groups received different pre-hospital care conditions; 0.9% saline only, 1:1 FFP/PRBC, and PRBC only. Clotting time and dynamics were measured using thromboeslastography. Significant attenuation in trauma induced coagulopathy was seen in both blood products groups compared to the 0.9% saline control. Use of a porcine model has both limitations and benefits. Porcine liver vessel injury is recognised as a good approximation for injuries sustained in the military.

A review of the literature shows FFP is required in addition to packed red blood cells to counter the dilution effect of adding plasma-poor PRBCs. Reviews of clinical practice for massive blood transfusion conclude a need for either a 1:1 or 2:3 FFP/PRBC ratio. Despite clear evidence for its role in resuscitation in severe haemorrhage, there are a number of recognised limitations to FFP use, such as complex refrigeration and transportation requirements.

Retrospective analysis of military casualties suggested a significant reduction in mortality with a higher FFP/PRBC ratio. Mortality rate decreased from 65% with a FFP/PRBC ratio of 1:8 to 19% with a ratio of 1:1.14. Variations in injury severity is one of several noteworthy limitations to the wider application of military research. Age, health and physical capacity of casualties care also likely vary compared with the general population.

Prevalence of Trauma Induced Coagulopathy in haemorrhagic trauma patients is significant. A 1:1 FFP/PRBC ratio in early fluid resuscitation improves outcomes. This practice is already adopted in UK military practice, but not yet throughout civilian pre-hospital treatment of trauma. Research is being conducted to consider the addition of specific coagulation factors, which unlike FFP wouldn’t require refrigeration.

A closed loop audit of the incidence and management of hypoglycaemia in the West Suffolk Hospital Intensive Care Unit, Bury St Edmunds

Carolyn Dales¹, Sarah Chetcuti² and Ayush Sinha¹

¹West Suffolk NHS Foundation Trust, Bury St Edmunds, UK

²Addenbrooke's Hospital, Cambridge, UK

Abstract

Introduction: Dysglycaemia is common in intensive care patients. Numerous studies have shown that tight glycaemic control can positively impact clinical outcomes in patients in intensive care units (ICUs), however this came at the expense of an increased rate of hypoglycaemia. Research has shown that severe hypoglycaemia is independently associated with an increase in mortality, length of hospital stay and is also associated with poor neurological outcomes. Early recognition and management is therefore imperative.

The most widely accepted definition of hypoglycaemia is a blood glucose of less than 4.0 mmol/L and is used in the most recent guidelines by the Joint British Diabetes Societies for Inpatient Care Group.

Aim: To assess the incidence of and treatment rate of episodes of hypoglycaemia among patients admitted to our ICU.

Method: Data from 1st November 2016 to 30th June 2017 was retrospectively collected using the software programme ‘MetaVision ICU’ (iMDsoft, Needham, MA, USA). We identified all patients admitted between this period who had an arterial blood glucose level of <4.0 mmol/L. Electronic patient records were examined to determine whether each episode was treated. Once a patient was on end-of-life care data were excluded. Following this, we organised training of intensive care staff to improve recognition, prevention and management of hypoglycaemia. This was done verbally and by displaying posters as an aide-memoire around the unit. We also distributed a questionnaire amongst ICU staff to identify knowledge gaps to tailor future training. Repeat data collection was conducted between 1st July 2018–31st August 2018.

Results: A total of 284 patients were admitted to the ITU at West Suffolk Hospital from 1st November 2016 to 30th June 2017, 65 of which had at least one episode of hypoglycaemia. Of these patients, 52% (34) were female and 66% (43) were 65 years or older. Of all patients, 31% (20) were previously diagnosed with diabetes, the majority of which (85%, 17) had type 2 diabetes. During this period, there were 187 separate instances of hypoglycaemia recorded. The most common cause of hypoglycaemia was an abrupt cessation or reduction in oral intake. Importantly, treatment was only administered in 110 cases (59%).

Following the training intervention, 94 patients were admitted (July-August 2018), 15 of which had at least one episode of hypoglycaemia. Of these patients, 13.3% (2) were known diabetic (Type 1). There were 33 separate instances of hypoglycaemia recorded, 25 of which received treatment (75.8%).

Conclusion: Hypoglycaemia has been associated with an increase in adverse outcomes including an increase in mortality. Even though an improvement was seen in the recognition and management of hypoglycaemia following our intervention, the incidence remains unacceptably high and its treatment unacceptably low.^1–3 Future work will focus on education of critical care staff, specifically on transitioning strategies when patients with diabetes are admitted and the importance of ongoing nutritional intake. A modification in the ‘MetaVision ICU’ programme will highlight all arterial blood glucose level of ≤4.0 mmol/L allowing easy identification.

Cardiac Ischaemia Detected by Pre-operative CPET Prior to Major Non-Cardiac Surgery at Manchester Royal Infirmary – a Single Centre Experience

Sam Forryan^1,2, Steve Benington¹ and Angella Bryan¹

¹Manchester University Foundation Trust, Manchester, UK

²Southport and Ormskirk Hospital NHS Trust, Southport, UK

Abstract

Cardiopulmonary Exercise Testing (CPET) is a central part of pre-operative assessment for patients approaching major surgery in our centre. The presence of ischaemia on CPET has been shown to be associated with excess peri-operative mortality in many studies. Anecdotally many patients with an ischaemic CPET in our centre are referred for further investigations, although this does not always change management and risks delaying surgery. We examined a sample of our own patient cohort to characterise the investigation and management of CPET ischemia detected pre-operatively, and its relationship with outcome.

110 consecutive patients with cardiac ischaemia planned for major non-cardiac surgery were identified from our CPET database. Patient electronic records were examined to determine what further investigations were performed and whether the patient was referred to a cardiologist or had medication changes or other interventions pre-operatively. Length of stay (LOS) and mortality data were collected from the hospital PAS system.

Patients were predominantly male (75/110) with a median age of 70 years. The commonest surgical groups were hepatobiliary(45%) and upper GI (20%). On CPET testing 52% were assessed as having mild, 41% moderate and 7% severe ischaemia by ECG criteria. 45% of patients were referred for further cardiac investigations, and 25% were referred to a cardiologist. These interventions were more likely in patients with moderate/severe ischaemia. Myoview™ was the commonest investigation, requested in 43 patients and performed in 39. 18/39 scans were reported as abnormal to some degree, with 11/39 (28%) showing a high burden of ischaemia. 13/18 patients with an abnormal Myoview™ were reviewed by a cardiologist, with medication changes alone in 6 and coronary angiography performed in 6 patients, 2 of whom proceeded to coronary artery bypass grafting. 40/110 patients with an ischaemic CPET ultimately did not proceed to operation (36%). The commonest reasons for this were unacceptable risk based on CPET/cardiology assessment (17/40), and disease progression (12/40). Referral to a cardiologist identified a high-risk group; 16/27 patients referred to a cardiologist ultimately did not proceed to surgery compared with 24/83 who were not referred (59% vs 29%). Post-operative median hospital LOS was 9 days and median critical care LOS 3 days. Three patients died within 90 days; 1 stroke, 1 anastomotic leak, 1 post-operative pulmonary complications. Only one patient was diagnosed as having a perioperative myocardial infarction.

Ischaemia is a common finding on CPET which frequently leads to further non-invasive testing in our centre. The incidence of major perioperative cardiovascular events appears low; this may be due to a combination of thorough pre-operative optimisation and selecting out unsuitable patients for surgery. There are inconsistencies in our approach to further investigation and referral for these patients and this may benefit from a more protocolised approach.

Nurse led fluid balance management and the use of a fluid balance calculator tool whilst delivering continuous renal replacement therapy

Max Richardson, David Burtonwood and Ajay Raithatha

Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK

Abstract

Introduction: The provision of continuous renal replacement therapy (CRRT) is recognised to significantly add to the work load of critical care nursing staff as reflected in Nursing Activates Score and Therapeutic Intervention Scoring System 28.^1,2 It is also established that a positive fluid balance in patients with acute kidney injury is associated with an increased 60 day mortality.³ We have recently transitioned greater responsibility in fluid management for patients on CRRT to the bed side nurses. In order to reduce nursing workload and improve the accuracy in reaching the prescribed fluid targets, we have also introduced a fluid balance calculator tool to aid nursing staff. We have subsequently assessed the implementation of this tool using a staff satisfaction survey.

Method: The fluid balance calculator tool is built in a read-only Microsoft Excel spread sheet. The tool is stored on the hospital intranet and is accessed by a hyperlink from our electronic bedside patient monitoring and records system. The tool takes into account the patient’s current fluid balance, target fluid balance, time until the target balance, hourly inputs from intravenous and enteral routes, average urine output over the last 2 hours and planned fluid infusions or additional losses. The tool calculates the projected fluid balance and the required hourly ultra filtration rate to achieve the desired goal.

12 months following the launch of the tool, all CRRT trained nursing staff were invited to complete a staff satisfaction survey. The survey was structured as a 5-point Likert scale addressing uptake, usability, speed and accuracy of the tool in addition to the individuals comfort in the transition to nurse led fluid management.

Results: 103 staff members were identified as trained to deliver CRRT. Trained staff members were invited by Emails to complete an online survey or to complete a paper based version. There were 52 responses giving a response rate of 50.5%.

73% of staff reported using the tool at least once per shift while managing patients on CRRT. 84.5% of those surveyed found the tool either easy or very easy to use. 86.4% of nursing staff reported that the tool was faster than manually calculating the ultra filtration requirements. 71% felt that the tool allowed somewhat, more or definitely more accurate fluid balance management compared to manually calculating the balance. The transition to nurse led fluid management has been well received with 76.8% either comfortable or very comfortable, 17% neutral and 5% reporting been uncomfortable with the change.

Conclusion: Our experience of transitioning greater responsibility in fluid management to the bed side nurse has been overwhelmingly positive. The fluid balance calculator tool has proven to be simple to use and has had a high uptake with our nursing staff. Nursing staff have also reported that they feel the fluid balance calculator tool is not only quicker but more accurate than manually calculating the fluid balance and ultra filtration rates. Furthermore the implementation of the fluid balance calculator tool using pre existing software has incurred no additional costs to our unit.

Identifying barriers and enablers to performing a Damage Control Surgery protocol amongst non-medical theatre staff at a Major Trauma Centre

Neil Roberts¹, Kate Holmes¹, Sam McAleer², Ed Brittain¹ and Paul Moor¹

¹Derriford Hospital, Plymouth, UK

²Royal Cornwall Hospital, Truro, UK

Abstract

Introduction: Severely injured patients sometimes require immediate transfer to theatre for life-saving Damage Control Surgery (DCS). Achieving this in a timely fashion requires a complex pathway of behaviours amongst theatre staff, ranging from obtaining and communicating critical patient details, to sourcing an appropriate theatre, trained staff and resources like blood and surgical kit. A ‘DCS protocol’ is in use to facilitate this at our hospital. However, case review reveals variable performance, with informal discussion suggesting multiple factors affecting this. Implementation science frameworks may be used to systematically analyse factors affecting clinical behaviour. This study uses the Theoretical Domains Framework (TDF) to identify barriers and enablers towards performance of the hospital’s DCS protocol and facilitate more effective performance improvement.

Methods: Paper questionnaire distributed to a purposive sample of all non-medical theatre staff (administrative, non-registered, band 5, band 6, band 7) for one week. Two questions for each of 14 TDF domains, addressing perceived importance and agreements with the domain as either a barrier or an enabler. Participants completed a 5-point Likert scale (1 = Very unimportant/strong barrier, 5 = Very important, strong enabler). Importance was compared using mean average. Agreement was described using a novel 2 x 2 graph to display concordance/discordance within participant groups.

Results: 69 questionnaires were returned out of 200 distributed (35% response rate). Out of those completing demographic questions, 4 were administrative staff, 11 were non-registered nursing, 39 were Band 5/equivalent nursing, 10 were Band 6 and 4 were Band 7. 20 responded as anaesthetic staff, 36 as surgical.

Domains perceived as most important for performing the DCS protocol were Skills and Resources (mean 4.1), closely followed by Beliefs about Consequences, Intentions, Social and Professional Role, and Knowledge(4.0). Domains perceived as less important were Behavioural Regulation (BR) (3.7), Emotion (3.4) and Reinforcement (3.0). There was some variation between staff groups, for example feedback-based domains (Reinforcement and BR) were more important for leadership groups (Band 6/7).

There was a wide spectrum of both concordant and discordant barrier and enabler domains. For example, Knowledge, Skills and Belief in Capabilities were important barriers for non-registered staff, along with Memory/Decisions/Attention – an enabler domain for other staff groups. Resources was a particularly discordant domain amongst non-registered and Band 5s. Audit and feedback domains (eg BR) were strong enablers for Band 6/7 but barriers for others.

Conclusions: This study systematically identifies targets and conducts a behavioural diagnosis for improving performance of the DCS protocol at this trust, for example particular need to focus on capability-based domains for non-registered staff, a need to further investigate variation in resource availability amongst Band 5s, and a need to maintain good feedback about DCS cases for those in leadership roles. Other trusts could use this method to identify their own barriers and enablers towards effective DCS. The method may easily be applied to guide other performance improvement processes beyond DCS.

Delivering Prehospital Enhanced Care in the Thames Valley region from BASICS: a volunteer solo responder’s first year from inception to service development

Shah Mizanur Rahman

Emergency Department, Wexham Park Hospital, Slough, UK

Abstract

This abstract will discuss the provision of prehospital enhanced care (PHEC) to patients within the Thames Valley region, UK. Specifically, the ongoing need for care beyond NHS care, becoming a ‘BASICS doctor’ from scratch, associated costs, use, benefits and the process of developing and improving a prehospital care service which provides critical care at the roadside. With more incidents than resources available, a small group of volunteer (BASICS) doctors provide ad-hoc voluntary responding around their full time jobs.

PHEC includes analgesia e.g. fentanyl, ketamine, methoxyflurane and regional anaesthesia, skills such as surgical airways, thoracostomy and thoracotomy and decision support with prehospital ultrasound imaging, special circumstance resuscitation and trauma management. Advanced drugs include critical care treatments such as ALS drugs beyond standard ambulance care such as vasoactive drugs, Doctors provide prehospital emergency anaesthesia (PHEA) as level 6–8 (Faculty of Prehospital Care) practitioners

The intervention described is the development of a solo responder, the intended outcome, an ambulance physician able to provide independent PHEC with indirect supervision and support from the on-call Medical Incident Advisor. This required the specification, design and procurement of most equipment and documentation needed to provide this service.

After sign off, an iterative process for service development started. This involved electronic post incident reflections with subsequent face to face discussions. This was key to identifying good practice, areas for development and root cause analysis of issues identified through near-peer and consultant supervision.

The following outcomes were assessed for since the inception of solo responding with contemporaneous logs and reflections being kept.

1. Incidents attended

2. Incidents attended requiring enhanced care (in any form)

3. Use of enhanced care skills

4. Time from activation to mobilisation

5. Time from at scene to ready to move ‘RTM’

Any appropriate activation was taken as an improvement, in that the incident, patients and emergency staff would otherwise not have access to enhanced care other than conveying to hospital.

Results are incomplete at present, however the current running numbers are below.

104 incidents involving 124 patients

70 ‘assists’ – attended to, and left care of the ambulance crew

27 ‘escorts’ – patients subsequently escorted to hospital

4 paediatric and 5 traumatic arrests

3 sedations

2 combat tourniquet deployments

1 thoracotomy

Database is still under completion, but will detail:

• jobs attended during the hours HEMS available

• average times from activation to attendance

• compared against time to HEMS attendance

• incidents attended outside of HEMS hours

• full breakdown of incident and intervention types/numbers

Conclusions:

1. The process from inception to service delivery requires significant financial, time and personal investment from the individual responder, with ongoing obstacles including skillset limitations such as PHEA, ongoing expenditure including fuel, enhanced care consumables, kit maintenance and CPD.

2. There is a considerable return on investment for the ambulance service from volunteer medical responders

3. Need for ongoing support and investment from NHS Ambulance Services to ensure continuing provision of enhanced care to patients who are critically unwell both for primary missions and secondary time-critical transfers.

Steroids in sepsis – prescribing practices of the West Suffolk Hospital Intensive Care Unit, Bury St Edmunds

Carolyn Dales and Wasim Huda

West Suffolk NHS Foundation Trust, Bury St Edmunds, UK

Abstract

Introduction: The current evidence base for the use of steroids to treat septic shock patients after fluid resuscitation and vasopressor therapy is weak. Two recent trials on its use in sepsis have produced varying results on 90-day survival but showed a beneficial effect on shock reversal. The different guidelines recommend a dose of less than 400 mg intravenous hydrocortisone for at least 3 days in patients with the above clinical scenario.

Aims: To review the prescribing practices of corticosteroid therapy in septic patients requiring a noradrenaline infusion (NI) in the West Suffolk Hospital Intensive Care Unit.

Method: Data from 01/01/2018–20/04/2018 was retrospectively reviewed using the software programme ‘MetaVision ICU’ (iMDsoft, Needham, MA, USA). We identified all patients admitted between this period who were prescribed a NI. Those who had a Sepsis-related Organ Failure Assessment score (SOFA) of ≥2 were defined as septic.

Within the NI group, we identified those who received hydrocortisone therapy, excluding those who received steroid therapy for reasons other than sepsis. We reviewed each NI prescription, identifying for each patient the strength of noradrenaline (mcg/kg/minute) at which hydrocortisone was commenced.

Using this data, we looked specifically at whether patients fulfilling sepsis criteria were given adjunctive hydrocortisone therapy. We determined what hydrocortisone prescription they received and at what noradrenaline strength the hydrocortisone was commenced.

Results: In the above time period, 171 patients were admitted to the West Suffolk ICU. Of these, 35 (20.46%, aged 19–88, mean age 66.46, M:F 1:0.75) patients were prescribed a NI.

Of the 35 patients, 26 (74.2%, aged 19–87, mean age 66.38, M:F 1:0.85) fulfilled sepsis criteria. Of these septic patients, 8 (30.7%, aged 59–78 mean age 68.75, M:F 1:0.6) received hydrocortisone specifically as an adjunct to their noradrenaline therapy. The NI strength when hydrocortisone was added ranged from 0.092–0.39 mcg/kg/minute (mean 0.33). Of those on a noradrenaline infusion who did not receive hydrocortisone, the highest infusion strength ranged from 0.04–0.533 mcg/kg/minute (mean 0.135).

The mean duration of hydrocortisone therapy was 2.9 days (median 3, range 1–4). The mean dose of hydrocortisone per day was 171.4 mg (median 150, range 150–300). The mean starting dose of hydrocortisone was 87.5 mg (median 100, range 50–100). There was wide variation in prescribing practices of hydrocortisone, though 50 mg TDS was the most common prescription.

Conclusion: Only 30.7% of patients on NI received adjunctive steroid therapy for treatment of their sepsis. Of those, the prescribing practices varied widely and did not conform to available recommendations. Whilst the choice not to commence steroid therapy is acceptable given current evidence, if adjunctive hydrocortisone therapy is deemed appropriate, it may be of use to follow the standard guidelines. Further work is planned trialling changes to the department’s MetaVision system whereby NIs of >0.5 mcg/kg/minute are flagged up as an opportunity to consider adjunctive therapy. Guidance of appropriate hydrocortisone prescriptions will be incorporated into local guidelines.

Intervention fidelity in trials of physical rehabilitation in critical illness across the recovery continuum: a systematic review

Kirsty Jerrard¹, Matthew Barclay², Linda Denehy³, Julie Bernhardt⁴, Sue Berney⁵, Bronwen Connolly^2,6,7,8

¹Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

²Lane Fox Clinical Respiratory Physiology Research Centre, Guy’s and St. Thomas’ NHS Foundation Trust, London, UK

³Melbourne School of Health Sciences, The University of Melbourne, Melbourne, VIC, Australia

⁴The Florey Institute of Neurosciences and Mental Health, Melbourne, VIC, Australia

⁵Physiotherapy Department, Austin Health, Melbourne, VIC, Australia

⁶National Institute for Health Research Biomedical Research Centre based at Guy’s and St. Thomas’ NHS Foundation and King’s College London, London, UK

⁷Centre for Human and Applied Physiological Sciences, King’s College London, London, UK

⁸Department of Physiotherapy, The University of Melbourne, Melbourne, VIC, Australia

Abstract

Intervention fidelity in clinical trials is imperative for ensuring interventions are implemented as intended, thereby enabling accurate interpretation of findings and consideration of the true therapeutic effect. Existing trials of physical rehabilitation in critical illness where no difference has been evident between experimental and standard care interventions, may have been influenced by varying levels of fidelity which may not have been accounted for. The aim of this systematic review was to evaluate intervention fidelity in trials of physical rehabilitation for adult patients following critical illness across the recovery continuum.

Figure 1.

Adherence to intervention fidelity recommendations.

Four electronic databases (CINAHL, EMBASE, Medline and CENTRAL) were searched. Eligible studies included randomised controlled trials evaluating effectiveness of physical rehabilitation interventions in adult critically ill patients, at any stage across the recovery continuum. Examples of physical rehabilitation interventions included functional training, early mobilisation, and exercise programmes. Two review authors independently screened articles for eligibility and completed data extraction. A third review author arbitrated where necessary, and cross-checked data. Data extraction mirrored the Stroke Recovery and Rehabilitation Roundtable (SRRR), and included adherence to recommendations within domains of intervention i) development, ii) monitoring, and iii) reporting of experimental and standard care intervention. Data were descriptively summarised as number and proportions of trials meeting recommendations fully, partly, not reported, or not applicable. The protocol for this review was registered a priori (PROSPERO CRD42017081286, www.crd.york.ac.uk).

17338 search results were identified, from which 28 trials were included for qualitative synthesis. Thirteen trials involved physical rehabilitation interventions delivered in the intensive care unit, 2 studies on the ward, and 6 studies post-hospital discharge only. Seven trials delivered interventions across multiple stages of the recovery pathway. Of the 28 included trials, 13 adopted an intention-to-treat (ITT) analysis approach, 10 analysed per-protocol (PP), and 5 reported both methods. The degree to which recommendations were fully reported in included trials ranged from 0% to 96% (Figure 1). Within development, the recommendation for clearly describing the core components was the most frequently fully reported (n = 22, 79%). In monitoring, reporting of assessment and measurement of fidelity were fully reported most frequently (n = 5, 18% each). Reporting of experimental interventions was fully adhered to the most (n = 27, 96%) in the recommendation for trials to provide a brief name descriptor. For reporting of usual care interventions, providing a description of the intervention components was the most fully adhered to recommendation (n = 12, 42%).

This systematic review highlights the lack of standardised and complete reporting of intervention fidelity in studies of physical rehabilitation during critical illness, with many key areas under-addressed. Adopting a more consistent approach, as per the SRRR, as well as reporting results of both ITT and PP analyses would enable greater understanding of the extent to which trial findings represent experimental intervention efficacy or the effect of poor fidelity.

Individual bars represent total number of trials (/28) that fully adhered to fidelity recommendations across domains of intervention development, monitoring, and reporting. Recommendations abridged from the Stroke Recovery and Rehabilitation Roundtable

‘Using Muscle Ultrasound to Establish Optimum Probe Placement for Rectus Femoris (RF) Image Acquisition'

Ashley Thomas^1,2, Mark Griffiths^1,2 and Julie Sanders^1,2

¹Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

²St Bartholomew's Hospital, Barts NHS trust, London, UK

Abstract

Background: Intensive care unit-acquired weakness (ICUAW) entails muscle wasting associated with an imbalance in muscle synthesis and breakdown, caused by immobility, inflammation and deranged bioenergetics. Serial ultrasound (US) measurement of the rectus femoris cross-sectional area (RFcsa) is a well-recognised means of quantifying wasting. B-mode ultrasound and echogenicity correlate well with athletic performance and skeletal muscle function in chronic disease. Muscle US echogenicity is low in healthy tissue, and increased values correlate with reduced strength and function with age.

Figure 1.

Comparison of Healthy volunteer (HV) and patient (ICUAW) in B-mode ultrasound and rectus femoris (RF) echogenicity. (VL, vastus lateralis; VI, Vastus intermedius; VM, vastus medialis).

Aim: To optimise the US technique for measuring RFcsa and muscle quality, and to test intra-observer reproducibility.

Methods: RFcsa was measured using B-mode US (13 MHz bandwidth, 6 cm linear transducer array). Two groups of volunteers were assessed: healthy (HV: n = 20) and patients recovering from critical illness (ICUAW: n = 10). HV was split into two groups: in the first, RFcsa was measured at intervals (50–90% total distance) from the anterior superior iliac spine (ASIS) to the superior patellar border (SPB) in both legs (n = 10). Using the optimised technique, the second group of HV (n = 10) and ICUAW group (n = 10) had RFcsa measured at two time points and muscle quality determined by histogram analysis of pixel intensity (pixel distribution having been removed by grey-scaling images) following established techniques (Adobe Photoshop).

HV compared with patients, had larger, better-defined muscle (fig 1): RFcsa 3.43 (0.80) versus 1.77 (0.54) cm (mean [SD], n = 20, p < 0.01 Student’s t-test). Pixel density for HV and patients were 49.08 (55.65–43.00) and 76.39 (86.9–65.8) (median [IQR], n = 20, p < 0.01 Mann-Whitney U test). Optimum probe placement was at 70% ASIS to SPB distance, based on image clarity and traceability (table 1).

Intra-observer variability was assessed using the optimised technique. Measurements taken at different times from HV and ICUAW groups (both n = 10) correlated well (r = 0.998). The Bland-Altman analysis confirmed adequate intra-observer agreement as >95% observations were within the limits of agreement.

Conclusion: Healthy volunteers had larger RFcsa and better muscle quality than patients, as expected. Acquiring US images at 70% from the ASIS to the SPB was optimal for RFcsa quantification. Using this method, the intra-observer variation was within acceptable limits as evidenced by a Bland Altman analysis.

Table 1.

Rectus femoris cross-sectional area (RFcsa) at progressive distances from 50% - 90% in healthy volunteers measured by ultrasound. Values are given by mean [standard deviation] n = 10. (ASIS, anterior superior iliac spine; SPB, superior patellar border; U/A, unable to achieve.

Distance ASIS to SPB	50%	60%	66%	70%	80%	90%
RFcsa (Left)	U/A	4.63 [1.00]	3.92 [1.37]	3.47 [1.22]	0.97 [0.43]	U/A
RFcsa (Right)	U/A	3.62 [0.60]	2.96 [0.85]	3.56 [1.61]	1.16 [0.64]	U/A
Accessible views  (Left and Right)	(L) 0/10 (R) 0/10	(L) 4/10 (R) 3/10	(L) 6/10 (R) 4/10	(L) 9/10 (R) 8/10	(L) 9/10 (R) 9/10	(L) 0/10 (R) 0/10

#End PJ paralysis can it be implemented within Critical Care?

Alex Hemsley

Newcastle upon Tyne NHS Foundation Trust, Newcastle upon Tyne, UK

Abstract

#EndPJparalysis is a social movement focused on getting patients up dressed and moving led by the director of NHS 360 Professor Brian Dolan. A recent pilot gave patients back 91,728 days or 250 years' worth of time across nine trusts in the East of England, as result of getting patients up and dressed. Studies suggest three-in-five immobile, older hospital patients have no medical reason requiring bed rest and that doubling their walking levels cuts the length of stay. Therefore on 17.04.18 a 70 day campaign was launched to get as many patients dressed and in their own clothes as possible. This project was to review if it is possible to engage with this project within a critical care environment.

Between 17.04.18 and 26.07.18 data was collected to identify how many patients on W38 critical care unit were dressed in their own clothes and mobilised out of bed.

In a 70 day period a total of 173 patients out of a total of 294 (59%) were dressed and mobilised out of bed (see figure 1). Of the 173 patients 59 were within ITU and 114 within HDU(see figure 2). The mean ICNARC mortality probability % 14.66% (range 0.4–97.2) and mean length of stay 3.49 days (range 0.1–33).

This piece of work demonstrates that it is possible for patients within critical care environment to be dressed in their own clothes and out of bed. A limitation of this study is that data regarding number of patients mobilised and in their own clothes was not collected at a weekend. However, all patients admitted and discharged over the weekends were included into the total number of patients. A further limitation was that patients admitted post-operatively, still had their belongings based on the ward they were initially admitted to. Additionally, nursing staff would historically ask family members, upon admission to critical care, to take patients clothes home.

Following on from this project after proving that it is possible to get critical care patients up and dressed. We are now looking into critical care purchasing a supply of clothes for patients to get dressed when they do not have their own. We are labelling this #GetoutofGowns so join us in getting critical care patients out of hospital gowns

The association between HbA1c and long-term neurologic outcome in cardiac arrest patients underwent therapeutic hypothermia

Dong Hun Lee and Byung Kook Lee

Chonnam National University Hospital, Gwangju, The Republic Of Korea

Abstract

Introduction: We examined the association between HbA1c and six-month neurologic outcome in cardiac arrest survivors treated with therapeutic hypothermia (TH) and the association between glycaemic status and neurologic outcome according HbA1c.

Method: This retrospective study included comatose cardiac arrest patients underwent TH from September 2011 to December 2017. HbA1c and glucose were measured after restoration of spontaneous circulation (ROSC) and normal or high HbA1c was defined using cut-off value of 6.4% of HbA1c. Mean glucose during induction, maintenance, and rewarming phase were calculated. Neurologic outcome was assessed using the cerebral performance category scale scale six months after cardiac arrest. The primary outcome was a poor neurologic outcome, defined as CPC 3–5.

Results: Of 384 included patients, 247 (64.3%) had an poor neurologic outcome. Poor neurologic outcome group had significantly higher HbA1c level and the HbA1c level was independently associated with neurologic outcome (odds ratio [OR], 1.408; 95% confidence interval [CI], 1.041–1.904). Patients with high HbA1c were dominant in poor group than in good group (27.5% versus 9.5%, p < 0.001). Mean glucose during induction phase remained a significant predictor for neurologic outcome in whole group (OR, 1.004; 95% CI, 1.001–1.007) and normal HbA1c group (OR, 1.004; 95% CI, 1.000–1.008) after adjustment, while all glucose variables were not associated with neurologic outcome in high HbA1c group.

Conclusion: Higher HbA1c was independently associated with poor neurologic outcome. Higher glucose level during induction was independently associated with poor neurologic outcome.

Performance of three score systems in predicting massive transfusion in patients with unstable gastrointestinal hemorrhage

Dong Hun Lee ¹

¹Chonnam National University Hospital, Gwangju, The Republic Of Korea

Abstract

Background: Upper gastrointestinal bleeding (UGIB) is the most common disease that may require massive transfusion in medical conditions, except trauma and surgery. The present study aimed to analyze and compare the prognostic performances of the Glasgow Blatchford (GB) score, pre-endoscopic rockall (PER) score, and Modified early warning (MEW) score in patients with UGIB.

Methods: This retrospective observational study included patients with UGIB from March 2016 to February 2018. Receiver operating characteristics analysis was performed to examine the prognostic performance of GB, PER, and MEW scoring systems. Logistic regression analysis was used to identify independent risk factors for massive transfusion, after adjusting for relevant covariates. The primary outcome was massive transfusion.

Results: Of the 484 included patients with unstable UGIB, 3.9% (n = 19) of patients received massive transfusion. The area under the curves (AUC) of GB, PER, and MEW scores for massive transfusion were 0.577 (95% confidence interval [CI], 0.531–0.621), 0.570 (95% CI, 0.525–0.615), and 0.767 (95% CI, 0.727–0.804), respectively. The AUC of MEW score was significantly different from those of GB and PER scores. In multivariate analysis, MEW score was independently associated with massive transfusion in patients with unstable UGIB (Odds ratio, 1.495; 95% CI, 1.100–2.033; P = 0.010).

Conclusion: MEW score has the highest prognostic performance for massive transfusion among score systems in unstable UGIB.

EQ-5D-5L for critical care patients receiving rehabilitation

Tanya Usher, Aron James and Gita Marais

St Bartholomew's Hospital, Barts NHS Trust, London, UK

Abstract

The self-reported EQ-5D-5L questionnaire is a popular Quality of Life (QoL) measure in the inpatient environment and is easy to complete. It has been used within the critical care environment (1) and been linked with economic considerations of intervention (2).

AHPs into Action (2017) states AHPs should evaluate improve and evidence the impact of their contribution. In this day and age of challenging NHS funding, therapists must be able to prove their worth to commissioners, and demonstrate a positive impact of interventions both from a clinical and economical perspective. The aim of this study was to assess the outcomes using the EQ-5D-5L for post-critical care patients in order to implement a functional assessment for patients after critical care in accordance with NICE CG83 guidelines.

Patients who received rehabilitation on critical care for over 48 hours and subsequently discharged home were included. Assessments were completed at discharge, 6-weeks and 3-months follow-up via telephone call. Those repatriated to a local hospital or following an integrated care pathway were excluded.

Fourteen patients were included in the study. There was a generalised trend of improvement in QoL and overall perception of individual health states with time. At discharge, all functional and non-functional domains highlighted perceptions spanning from ‘No Problems’ or ‘Slight Problems’ to ‘Severe’ and ‘Extreme Problems’. Subsequently, at 6-weeks follow-up, extreme problems were not expressed, with the least favoured response being ‘Slight problems’ in the EQ-5D Domain ‘Usual Activities’. After 3-months, the Domain ‘Usual activities’ again demonstrated the least favourable results with ‘Moderate Problems’. At this point, all physical and non-physical domains had respondents that felt they had ‘No Problems’. Thus, ‘Usual Activities’ may represent an area of focus for future rehabilitation needs.

Limitations included: a small sample size; and reduced data collection period meaning linear data could not be collated. Additionally it is recommended that physical, cognitive and psychological concerns are considered post critical care. The EQ-5D-5L does not address cognitive impairments which could directly influence a patient’s perceptions of their physical and non-physical morbidities.

However, the EQ-5D-5L is easy to complete and implement with appropriate organisation of follow-up assessments. Established critical care follow-up clinics could facilitate 6-week and 3-month data collection by sending EQ-5D-5L questionnaires with the appointment letter to minimise the time taken to make phone calls to the discharged patients and therefore clinician time. Furthermore, domains requiring further rehabilitation focus could be identified, this study revealed ‘usual activities’ to have the least favourable perceptions at each time-frame. Attention on rehabilitation goals related to usual activities could therefore be the focus of the rehabilitation team covering this patient population in the future.

Further data collection is required to generate linear relationships between results. However, once established this also could be extrapolated to consider Quality Adjusted Life Years and cost-benefit analysis.

Audit of appropriateness of Stress Ulcer Prophylaxis (SUP) prescribing in Critical Care

Christopher Couzens¹ and Robert McCormick²

¹University Hospitals Southampton NHSFT, Southampton, UK

²Royal Bournemouth Hospital, Bournemouth, UK

Abstract

Background: Patients on Critical Care are at higher risk of stress ulceration, with over 75% developing mucosal erosions and subepithelial haemorrhage within 24 hours of admission (Mutlu et al 2001). This risk is reduced by gastric acid suppression with histamine-2-receptor antagonists (H2RA) or proton pump inhibitors (PPI). These drugs are not without risks; including drug interactions, increased rates of Clostridium Difficile infection, hyponatremia and Ventilator Associated Pneumonia, and are currently being investigated in the pan-European SUP-ICU Trial.

Our Critical Care Unit uses an electronic notes system allowing for ‘prescribing bundles’, were groups of common drug and fluid therapies can be charted on admission including our default SUP bundle as Esomeprazole 40 mg IV or Lansoprazole 30 mg PO. In view of potential harm from inappropriate prescribing we wanted to audit current prescribing.

Methodology: This was a retrospective audit, in which 100 consecutive patients admitted after 1st July 2017 were identified from our admission database with notes and drug prescription reviewed.

At the time of audit, we had no published SUP policy, therefore we used criteria from the SUP-ICU Trial. We also accepted historic PPI/H2RA therapy as a valid indication for SUP.

Standards were that 100% of patients receiving SUP required a valid indication and 100% of patients with an identified risk factor should have SUP prescribed.

Results: 92 patients were prescribed SUP on admission. 62 were prescribed the default bundle, with the remaining 30 receiving alternative PPI prescriptions.

None were prescribed H2RA or alternative classes of SUP.

Of 92 patients prescribed SUP, 67 met the inclusion criteria with top indications being previous PPI therapy, vasopressor use and invasive ventilation; 25 had no risk factors requiring SUP.

Of of the 8 patients not prescribed SUP; 3 had no risk factors but 5 had one or more risk factors requiring SUP.

Discussion: The results of this audit showed that the 92% of patients were prescribed and received SUP. This is similar to larger studies looking SUP prescribing in ICU (86% Erstad et al 1999).

However, a significant proportion of patients had potentially inappropriate prescriptions, similar to other audits (Frandahet al 2012). 25 patients were prescribed SUP without identifiable risk factors, and 5 patients with risk factors had no SUP. It was not determined whether this led to patient harm.

The most common reason for SUP prescription pre-morbid prescription of a PPI or H2RA (34 patients). It was beyond this audit to review these historic prescriptions, but only 8 of these patients had identifiable risk factors for stress ulceration on admission

Although not reviewed it was noted that the all prescriptions without valid indications, were given the default bundle. This raises the possibility that these bundles are being used inappropriately and medications prescribed without thought by admitting clinician. Subsequent data review will look at other inappropriate bundle prescriptions i.e. sedative infusions for patients not intubated and ventilated.

Improvements can be made with a prescribing policy for SUP, continued education of staff and regular review of ongoing need for prescriptions with re-audit to review progress.

Analysis of validity and usability of the Situation Awareness Global Assessment Tool (SAGAT) during Simulation training for Intensive Care teams

Laura Vincent^1,2, Helen Higham^1,2, Paul Greig¹, Rosie Warren¹, Jonathan Chantler^1,2, Stuart McKechnie^1,2, Trevor Venes², Duncan Young¹ and Charles Vincent³

¹Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK

²Oxford University Hospitals NHS Foundation Trust, Oxford, UK

³Department of Experimental Psychology, University of Oxford, Oxford, UK

Abstract

 • Situation awareness (SA) is comprised of three levels: 1) perception, 2) comprehension and 3) projection. SA errors contribute substantially to critical incidents in ICU.¹ Training interventions targeted at improving SA have been shown to enhance performance and outcomes in non-medical environments.²

• Reliable measurement of SA is vital in understanding how training improves performance. Direct “freeze-frame” questionnaires (e.g. the Situation Awareness Global Assessment Tool – SAGAT) have been postulated to provide the most accurate assessment of SA.³

• SAGAT was designed for military settings. It has been adapted for healthcare^{4 5}but has not previously been used in ICU training.

• Institutional ethical review classified the study as educational audit.

• Two scenarios were adapted from our regular ICU simulation training. Scenario A was less complex than scenario B.

• 22 SAGAT questions were generated iteratively by subject matter experts for each scenario and divided across three pause points

• Questions were targeted at each level of SA (e.g. What was the blood pressure? What was the heart rhythm? What are your next steps?)

• SAGAT questions were allocated randomly to one scenario in each session (because training time was limited)

• Participant answers were analysed by calculating percent correct at SA levels and pause points for individuals and teams

• Validity and usability were analysed using qualitative and quantitative methods

• Design of the SAGAT questions took six iterations over one month

• 19 multidisciplinary teams (19 doctors, 38 nurses) participated in the training. All participants completed both scenarios. SAGAT was applied 8 times to scenario A and 11 to scenario B

• SAGAT accuracy scores were significantly higher for scenario A: mean difference 7.8% (95% CI, 1.4 to 14.2) p = 0.01. However, this difference equated to only 2 questions per participant overall

• Secondary findings included significantly different SAGAT scores at different levels of SA (Figure1) and pause points (Figure 2) between participants. However this may reflect the design of the questions or underlying construct being measured as opposed to changing SA

• Administration of the SAGAT tool during training was straightforward. Pause times averaged 8.1 (±1.5) minutes for scenario A and 6.3 (±1.4) minutes for scenario B

• Course feedback overall was excellent

• 86% (49/57) of candidates found SAGAT pauses not at all or mildly disruptive to learning, 14% (8/57) found them very disruptive to learning

• SAGAT showed limited evidence of validity and usability for the measurement of SA in simulation training for multidisciplinary teams in ICU.

• Further research should consider:

  ○ The feasibility of developing SAGAT for further similar AICU scenarios

  ○ Methods of improving the discriminatory capability of SAGAT

Figure 1. Mean SAGAT scores at SA levels 1–3 in scenarios A and B. RM ANOVA revealed significantly higher scores at SA level 3 versus 1 (*p ≤ 0.01) and level 3 versus 2 (§p = <0.01)

Figure 2. Mean SAGAT scores at pause points in scenarios A and B. RM ANOVA revealed significantly higher scores at pause 3 versus 1 (*p ≤ 0.01) pause 3 versus 2 (§p < 0.01), and pause 2 versus 1 (§p < 0.001)

Pharmacist medicines optimisation and error mitigation at critical care discharge – a human solution to an electronic risk

Fraser Hanks¹, Carol Jones²

¹St Thomas' Hospital. Guy's and St Thomas' NHS Foundation Trust, London, UK

²Department of Critical Care, Guy's and St Thomas' NHS Foundation Trust, London, UK

Abstract

Background: Patients are vulnerable to medication error at critical care discharge as it is a period of significant discontinuity^1–2. Medication changes are frequent and include:

• Suspending regular medication – up to 75% is held on admission to ICU.³

• Changes to medication, e.g. dose, frequency, route.

• Initiation of new medication during critical illness.

A comprehensive Critical Care Pharmacist (CCP) review at discharge can prevent medication errors, such as omission (e.g. regular medication), and unintended continuation (e.g. antipsychotics for ICU delirium⁴). Guy's and St Thomas' Foundation Trust (GSTT) currently use different electronic prescribing systems within critical care and on the wards. These systems lack interoperability and medication is transcribed between systems at step-down.

Audit criteria	Prospective CCP review 2018	Retrospective ward pharmacist review - 2016	Retrospective ward pharmacist review - 2014
Number of charts audited	42	71	120
Number of charts with an error	69% (29/42)	N/a	27% (32/120)
Total number of interventions	78	24	39
Mean no. of interventions per chart	1.9	0.3	0.3
Risk stratification	2018	2016	2014
Simple – unlikely to cause harm (e.g. Antibiotic duration)	28% (22/78)	62.5% (15/24)	43.6% (17/39)
Serious – potential to cause reversible harm (e.g. ACEi restarted in AKI)	68% (53/78)	37.5% (9/24)	48.7% (19/39)
Critical – potential to cause irreversible harm (e.g. treatment dose dalteparin not transcribed)	4% (3/78)	0% (0/24)	7.7% (3/39)

Therefore, the following processes are required at step-down:

• Transcription accuracy check between systems.

• Reconciliation of pre-admission medication.

• Review of medication initiated on ICU.

At GSTT ward pharmacists perform the majority of transcription checks as they occur outside of CCP core hours, often the next working day. However, CCPs are better placed to perform discharge medication reviews:

• They have knowledge of the patient and can individualise care.

• They have knowledge of critical care medication issues and rationale for changes.

• They are competent with both prescribing systems.

• The majority of CCPs are prescribers and rectify error(s) immediately.

• Discharge prescriptions frequently contain error. Prospective CCP checking can intercept errors before they have the opportunity to reach the patient.

Recognition of these issues, in addition to an increase in frequency and severity of medication incidents, has resulted in step-down prescribing being added to the Trust corporate risk register. To mitigate this risk we proposed enhanced CCP resource to review all step-down prescriptions, with the aim of improving quality and safety of step-down prescribing.

Objectives: Compare the error rate identified by prospective CCP medication review and compare to previous audit cycles of retrospective ward pharmacist review. Additionally, to stratify the potential risk of these errors.

Intervention: Prospective CCP review of discharge prescriptions during the intervention week on the adult critical care at GSTT.

Results:

Conclusions: Errors are frequent at step-down and the majority of errors have the potential to cause patient harm if uncorrected. This risk can be mitigated by prospective CCP screening of discharge prescriptions. Critical care units should review their discharge processes and consider if additional CCP resource could improve the quality and safety of step-down prescribing.

Patient experience of thirst in high dependency care

Jenny Clark, Joel Meyer, Christina Iezzi, Marco Castillo and Sally Archer

St Thomas' Hospital. Guy's and St Thomas' NHS Foundation Trust, London, UK

Abstract

Background: Thirst is a common, distressing and overlooked symptom in critical care (Rose et al 2013). It is associated with anxiety, desperation and a feeling of powerlessness (Kjeldsen et al 2014). Patients attending a critical care survivors forum at our Trust recalled thirst as a striking symptom of their admission. The aim of this study was to determine the frequency and impact of thirst in a critical care population and to explore measures to improve symptoms.

Methods: Weekly assessment of patient-reported thirst was undertaken on all patients who met the inclusion criteria on a mixed 11 bed HDU for 4 weeks (April 2018). Inclusion criteria were: English speaking, >18 years old, admitted to level 2/3 care (ICS 2009) for >24 hours with RASS +1/-1 (Sessler et al 2003).

Length of critical care stay, airway status and oral intake status were recorded. Patients who reported thirst in response to a screening question then rated their thirst intensity and distress on a 10 point scale, with 0 indicating no distress/intensity and 10 being maximum symptoms (Puntillo et al 2014). Patients were then asked to suggest solutions to ease their current thirst.

Results: 24 patients met the inclusion criteria and participated in the study. Median length of critical care stay was 10 days (range 2–69). Nine assessments were completed with patients on unrestricted intake, two were under fluid restriction, six were nil by mouth and seven patients were on restricted/modified intake due to dysphagia. Seven patients were tracheostomised at time of assessment (3 cuff inflated, 4 cuff deflated).

Half (12/24) of participants reported current thirst and there was a median intensity score of 8/10 (range 4–10). Median thirst distress was 7.5/10 (range 3–10). Most (83%) of patients who were NBM reported thirst (n = 5), with 44% (n = 4) of those on restricted oral intake and 33% (n = 3) on unrestricted oral intake reporting thirst.

There were consistent suggestions for thirst solutions with 4 key themes for symptom management emerging; craving flavours, wanting a cold sensation in the mouth, environmental factors (ie. avoiding mixing patients who are eating/drinking with NBM patients on the ward) and a want for fastest possible resumption of unrestricted oral fluids.

Conclusion: This service evaluation shows a high prevalence for thirst in critical care with a high symptom burden. There is a trend towards more frequent thirst in those NBM patients assessed in the sample. However, even those patients who have no oral intake restrictions report feeling thirsty in one third of this sample. Qualitative feedback shows some consistent themes which could act as a basis for proactive, responsive management of thirst in critical care. Some findings in this service evaluation are limited by small sample sizes, especially when filtering outcomes into specific cohorts. This preliminary work highlights the need to embed thirst assessment and management protocols into clinical working in critical care with the aim of improving patient experience.

This study received ethics committee approval or equivalent. Clinical Governance Approval through the Trust Quality Improvement and Audit Committee (Project number 8316).

The experience of a Regional ICU of implementation of a central line LocSSIPThe experience of a Regional ICU of implementation of a central line LocSSIP

David Penney, Jonathan Pugh and Lowri Roberts

Cardiff University, Cardiff, UK

Abstract

Background: Central venous cannulae are some of the most commonly used intravenous devices on ITU. They have many applications, including fluid resuscitation, administration of centrally-acting drugs, and haemodialysis. However, placing a central line is an invasive procedure in what is a very vulnerable group of patients, and carries with it significant risks.¹ These include, but are not limited to arterial puncture, gas embolisms and pneumothoraces. On several recent occasions, these complications have resulted in death. A large proportion of these arise from missed steps in the procedure – something which can improve.

With this in mind, a LocSSIP was created for our DGH in order to improve safety standards and step-wise documentation of CVC insertion. This project collected data from before and after the LocSSIP introduction, to gauge its efficacy. It also aimed to review the protocol, with a view to rolling out similar LocSSIPs elsewhere.

Results: LocSSIP use resulted in a broad improvement of CVC documentation. Based on our criteria, the CVCs were sufficiently documented in 55% of procedures post-LocSSIP, compared with 25% before the protocol was introduced.

Apart from a few exceptions, stepwise documentation was significantly better once the LocSSIP had been introduced. For example, VBG (venous blood gas) use increased by almost 100% once the LocSSIP had been used – this is a mandatory step (as dictated by AAGBI guidelines 2016)², without which a central line might be wrongly sited in an artery. Similarly, ultrasound guidance for siting the line was improved – cutting down the risk of a missed arterial line. Post-LocSSIP, the removal of the central line was documented in many more cases (28% c.f. 11%). By knowing how long a line has been in situ, we can assess the risk of an infected line. Also, line removal can result in a potentially fatal air embolism.

Conclusion: The introduction of the LocSSIP has resulted in an higher standard of documentation for CVC insertion and removal. Procedures are more commonly meeting the criteria for what is needed in a CVC insertion.

One area which could be improved is the documentation of CVC removal. Numerous examples show air embolisms resulting following CVC removal; clearly a robust system should exist to minimise chances of this recurring.³ There exists elsewhere in the hospital a protocol for CVC removal – which could easily be introduced into a ‘CVC pack’. Alternatively, modifying the checklist to include a ‘removal’ section would encourage its documentation. In terms of data collection, having all the relevant information on one page would be useful.

CVCs are used in ICU, in ED and also in theatre. We suggest there should be a robust, uniform protocol for their insertion and removal – both with adequate documentation. The LocSSIP has shown this is possible in one area; it is our recommendation that this be rolled out to other departments across the hospital.

Peri-haematomal oedema volume following intracerebral haemorrhage is negatively associated with plasma osmolality

Susan Stevenson¹, Virginia Newcombe¹, David Menon¹, Ben Glocker² and Steven McDonagh²

¹University Department of Anaesthesia, Cambridge, UK

²Imperial College London, London, UK

Abstract

Introduction: Peri-haematomal oedema (PHE) is a cause of secondary neuronal injury following primary spontaneous intracerebral haemorrhage (ICH). Absolute PHE volume is independently associated with poor outcome.¹ Minimising oedema volume therefore represents a potential therapeutic target in a disease with currently no specific treatment.² Mechanisms underlying oedema formation include extravasation of osmotically active molecules at the haematoma and perihaematomal parenchyma interface. Sodium is the main determinant of serum osmolality with hyponatraemia being associated with poorer outcome.³ We investigated the relationship between osmolality and PHE volume.

Methods: Patients with primary spontaneous supratentorial ICH, diagnosed on computerized tomography (CT) and admitted to Neurointensive Critical Care Unit (NCCU), Cambridge, between January 2015 and August 2017, were included retrospectively. Using a novel software program to map haematoma and oedema on CT, the volumes of each were determined. Volumes were measured on every CT performed for each patient during their admission. We excluded patients with a single scan only and excluded scans following surgical evacuation. Clinical data was extracted from electronic notes and included demographics, laboratory values and trends. Haematological variables included haemaglobin, platelets, prothrombin time and activated partial prothrombin time. Biochemical variables included sodium, potassium, urea, creatinine, glucose and serum osmolality when measured. Due to the scarcity of measured osmolalities, osmolalities at the time of scan were calculated from the nearest values using the formula:

$$\text{Osmolality}=2[\text{Sodium}+\text{Potassium}]+\text{Urea}+\text{Glucose}$$

Correlations between volumes and laboratory data were calculated and statistically analysed using the program SPSS.

Results: Forty-one patients were identified, 28 of which had multiple scans and 17 had eventual surgical evacuation. The median age was 59 years with the majority (78%) of patients being male. The most common co-morbidity was hypertension (49%) with 29% of all patients receiving medications associations with electrolyte disturbances (diuretics and ACEIs) prior to admission. The median Glasgow Coma Scale (GCS) on presentation was 10 (IQR:7–11).

Using the absolute volume, PHE was observed to reach a peak in the first 1–2 days following ictus and decline over the course of weeks. Pooling the absolute PHE volumes and serum osmolalities of all scans against time, revealed an inverse relationship between the two (Figure 1).

Figure 1.

Serum osmolality and oedema volume against time.

Figure 2.

Relationship between peak oedema volume and serum osmolality.

Fig 1.

Fig 2.

In patients with multiple scans, using peak oedema volume as a marker of oedema severity, only sodium (p = 0.02) and osmolality (p = < 0.001) values taken at time of scan showed a statistically significant negative correlation with oedema volume (Figure 2). No other relationship was observed using any other laboratory values.

Discussion: Within our patient cohort, a negative association between peak oedema volume and serum osmolality and sodium was observed. Enhanced oedema volume may serve as mechanism behind the deleterious effects of hyponatraemia in ICH with optimization of serum osmolality representing a potential therapeutic target.

Use of Stress Ulcer Prophylaxis in Critically Unwell Patients

Krushna Patel¹, Andrew Baigey¹ and Sinan Bahlool²

¹Kings College Hospital, London, UK

²East Sussex Healthcare NHS Trust, Sussex, UK

Abstract

Currently, stress ulcer prophylaxis is prescribed routinely for all patients admitted to the William Harvey Hospital ICU. The evidence suggests that this approach may not reduce the incidence of upper GI bleeds, and may contribute to an increase risk of mortality and hospital acquired pneumonia due to loss of the protective bacteriostatic effect of gastric acid^1,2. Simultaneous treatment of stress ulcer prophylaxis and broad-spectrum antibiotics also contributes to increased risk of Clostridium difficile infection.² There are several factors which put patients at higher risk of GI bleeding. For those patients, the evidence suggests the benefits of stress ulcer prophylaxis outweigh the negatives.³ The two main choices of stress ulcer prophylaxis in the UK are proton pump inhibitors (PPIs) and histamine-2 receptor blockers. There is currently little evidence to support the use of one over another for routine stress ulcer prophylaxis. We conducted an audit to compare the practice in our intensive care unit against the published best practice guidelines for the provision of Intensive Care Services (GPICS), edition 1, 2015.⁴ We suggest for patients that are at high risk of GI bleeding; due to coagulopathy, prolonged mechanical ventilation, recent GI ulceration, traumatic injuries or satisfying two minor criteria (sepsis, long ICU stay, occult GI bleeding for >6 days, or on glucocorticoid therapy) should be prescribed stress ulcer prophylaxis via an enteral route where possible. In patients without any of the above risk factors and absorbing enteral feed, stress ulcer prophylaxis is not routinely recommended. Data was collected from prescription charts over December 2017. Patients were assessed as to whether they met the criteria and if stress ulcer prophylaxis was prescribed appropriately. A total of 43 adult patients, both surgical and medical, were included. Stress ulcer prophylaxis was prescribed for 88% of the cases but on review only half met the criteria for ulcer prophylaxis. Only 53% of these patients received prophylaxis via an appropriate route. Therefore, overall prophylaxis was correctly prescribed in only 35% of the cases. These findings confirm that many patients admitted to ICU are prescribed ulcer prophylaxis unnecessarily or inappropriately. Bases on the results new guidelines have been suggested across the Trust for prescribing stress ulcer prophylaxis. With assessments on a case-by-case bases, we hope to change our practice towards taking a more cost effective and judicious approach to using stress ulcer prophylaxis.

Proposed external validation of the ABC + Lactate score for 30-day mortality in patients with decompensated cirrhosis

Alex Warren¹, Jessica Kamal², Keval Naik³ and Victoria Snowdon³

¹Division of Anaesthesia, Department of Medicine, University of Cambridge, Cambridge, UK

²University of Cambridge, Cambridge, UK

³Department of Hepatology, Cambridge University Hospitals NHS Trust, Cambridge, UK

Abstract

Background: Patients with decompensated cirrhosis often develop organ failure and require critical care. Traditional scoring systems such as Child-Pugh have been found to perform poorly at predicting acute mortality in such patients. Described in 2017, the ABC+Lactate score (Albumin, Bilirubin, Clotting + Lactate) was shown to accurately predict mortality at 1 year in patients with cirrhosis admitted to ICU ^[1], but it has not yet been externally validated. Our objective was to perform external validation of this score on a longitudinal observational cohort study of patients with acute-on-chronic liver failure.

Method: We identified patients with a diagnosis of cirrhosis (proven by imaging or biopsy) admitted to Addenbrooke's Hospital between 1st January 2015 and 1st August 2017 who met the European Association for the Study of the Liver criteria for acute-on-chronic liver failure (ACLF). Clinical and biochemical data from the electronic patient records system were used to calculate several commonly used scoring systems at presentation. Follow-up information at 30 days and 1 year was identified. The ABC+Lactate score and grade were calculated as per the methods described in the derivation cohort. Chi-squared goodness-of-fit test were used to perform testing of the score on this external cohort. Kaplan-Meier analysis and log-rank testing was used to compare mortality between patients in each ABC+Lactate grade.

Results: During the study period, 95 patients were identified who met criteria for ACLF and had sufficient data available for inclusion. Most patients (n = 74, 77.9%) had an alcohol-related aetiology of liver disease. The majority of patients presented with variceal bleeding (n = 28, 29.2%), ascites and/or encephalopathy (n = 44, 45.8%) or sepsis (n = 8, 8.3%).

59 patients (61.5%) were admitted to critical care. The overall 30-day mortality was 24.2% (n = 23), and 1-year mortality was 54.7% (n = 52).

An increased ABC+Lactate score at presentation was significantly associated with 30-day mortality (median score 10.7 in survivors vs. 12.3 in non-survivors, p = 0.031. However, goodness-of-fit chi-squared test based on the pre-defined ABC+Lactate grades was not statistically significant (chi-squared = 2.304, p = 0.183).

Patients with ABC + Lactate Grade 1 (ABC + Lactate score <8) had 30-day mortality of 6.3% and 1-year mortality of 31.3%. Those with Grade 2 (ABC + Lactate score 8–10.9) had 30-day mortality of 27.3% and 1-year mortality of 64.7%, while those with Grade 3 (ABC + L score ≥11) had 30-day mortality of 28.3% and 1-year mortality of 71.3%.

Kaplan-Meier analysis with log-rank test demonstrated no significant difference between these grades (p = 0.174)

Conclusion: A higher ABC + Lactate score is associated with 30-day mortality in patients with cirrhosis and organ failure. However, goodness-of-fit tests indicate that the suggested cut-offs are unlikely to provide a useful clinical prognostication tool in this patient group. More research is required to identify factors associated with survival in this patient group.

Proning in UK critical care departments: a questionnaire survey on current practices

Owen Chambers¹, David Whitmore², Peter Bamford³ and Peter Shirley⁴

¹ST6 Anaesthesia, Aintree University Hospital NHS Foundation Trust, Liverpool, UK

²ST9 ICM/Anaesthesia, Aintree University Hospital NHS Foundation Trust, Liverpool, UK

³ST9 ICM/Anaesthesia, Royal Liverpool and Broadgreen University Hospital Trust, Liverpool, UK

⁴Consultant in Anaesthesia and Critical Care Medicine, Royal London Hospital, London, UK

Abstract

Introduction: Randomised controlled trials have demonstrated that oxygenation is significantly improved in patients with acute respiratory distress syndrome (ARDS) when ventilated in the prone position [1], with more recent studies suggesting a mortality benefit [2]. A 2014 meta-analysis suggests that the use of prone ventilation when employed for greater than 16 hours per day in patients with moderate to severe ARDS improved mortality [3]. These studies culminated in a Cochrane systematic review published in 2015 recommending that prone ventilation for 16 or more hours per day should be actively considered in patients with severe hypoxaemia within 48 hours of mechanical ventilation [4]. Given these recent developments, we set out to gain an insight via a questionnaire into the current practices of proning patients in critical care departments in the UK.

Methods: A questionnaire survey with 14 items was sent to Intensive Care Society (ICS) members. Aspects covered included timings, durations, practicalities, complications of proning and availability of local protocols relating to safe proning practices.

Results: Replies were received from 133 members, almost entirely doctors (97%) working in adult (97%) critical care units with mixed medical/surgical patients (72%). Respondents seemed to have experience with proning patients with 82% routinely performing this for refractory hypoxaemia. There was variation in reported duration of proning per episode; 48% proned for 12–16 hours per day and 47% proned for 16–20 hours per day. A small number (3%) proned patients for >20 hours per day. Complications witnessed by respondents included pressure sores (69%), intravascular catheter displacement (29%), endotracheal tube displacement (20%) and cardiovascular instability (10%) among others. The most commonly affected areas by pressures sores were the face (69%), ears (11%) and mouth (9%). Only 31% of respondents utilised a checklist or protocol for proning patients and only 10% were aware of guidance regarding cardiopulmonary resuscitation in a proned patient. Most had no formal training or guidance on the practicalities of proning patients on their units (69%). A national, ICS/FICM endorsed guidance document would be welcomed by 86.4% of respondents.

Conclusion – The intervention of prone positioning now has a strong evidence base and although labour intensive and would almost certainly be a beneficial and cost-effective intervention [4]. This survey suggests that the prone position is commonly employed in patients with refractory hypoxaemia, however there is little in the way of formalised guidance in its use. As suggested by the findings of this survey, there is a need for a national guideline regarding the safe proning of patients with refractory hypoxaemia with the aim of reducing complications for these patients and improving care.

Soluble junctional adhesion molecules as predictors of morbidity in patients undergoing major abdominal surgery

B Lalabekyan^1,2, Matthew Golding³, Lorena Perez³, Sussan Nourshargh³ and Michael O’Dwyer^1,2

¹Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

²Royal London Hospital, Barts Health NHS Trust, London, UK

³William Harvey Research Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

Abstract

Acute kidney injury (AKI) in patients undergoing major abdominal surgery is associated with increased morbidity and mortality. Junctional adhesion molecules (JAMs) play an important role in vascular permeability and leukocyte transmigration. In experimental models JAMs were demonstrated to contribute to neutrophil infiltration in the liver and leucocyte recruitment in the lungs. It has been also demonstrated that JAMc blockade promotes renal recovery in chemotherapy induced AKI. A correlation between the level of circulating JAMs and incidence of distant organ failure in trauma patients has also been demonstrated. In this study, we investigated the relationship between perioperative levels of JAM-C and JAM-A with the incidence of postoperative AKI and organ dysfunction in patients undergoing major abdominal surgery.

Thirty-three patients undergoing elective major abdominal surgery were recruited. Blood samples were collected preoperatively and at 24 and 48 hours postoperatively. The levels of circulating JAM-C and JAM-A were measured using ELISA technique. The postoperative Sequential Organ Failure Assessment (SOFA) score was recorded daily on days 1–4.

Thirty-three patients (18 males and 15 females) with median age of 66 (46–82) were recruited. Ten patients (30.3%) underwent colorectal surgery, 4 (12.1%) had surgery involving the upper gastrointestinal tract and 19 (57.6%) had hepatobiliary or pancreatic surgery. Twenty-seven patients (81.8 %) had surgery for a malignancy.

A repeated measures ANOVA with a Greenhouse-Geisser correction determined that mean JAM-C and JAM-A concentrations differed significantly between time points [(F(1.289, 36.082) = 6.488, P = 0.01) and (F(1.671, 46.780) = 17.179, P < 0.0005)] (Fig1).

The levels of JAM-C at 48 hours were associated with higher SOFA score at days 2, 3 and 4 (R²= 0.363, P < 0.05; R²= 0.362, P < 0.05 and R²= 0.420, P < 0.05 respectively). Levels of JAM-A preoperatively and at 24 hours postoperatively were associated with higher SOFA score at postoperative days 3 and day 4 (R²= 0.389, P < 0.05 and R²= 0.726, P < 0.05 respectively) and levels of JAM-A at 48 hours were associated with the elevated SOFA score at day 4 (R²= 0.278, P < 0.05).

Elevated levels of JAM-C at 48 hours postoperatively were associated with postoperative AKI (P = 0.003) and a similar association was observed between the levels of JAM-A at 24 hours postoperatively and postoperative AKI (P = 0.02). (Fig 2)

This study demonstrated interesting links between elevated levels of JAM-C and JAM-A in the perioperative period and the incidence of AKI and organ dysfunction in the postoperative period. Further studies could help to investigate these associations further.

Extracorporeal Membrane Oxygenation: Ethical considerations and End of Life Care

Alexander Smith, Christoph Vandenbriele and Suveer Singh

Royal Brompton and Harefield NHS Foundation Trust, London, UK

Abstract

Introduction: Veno-Venous Extracorporeal Membrane Oxygenation (VV-ECMO) provides a bridge to recovery in the treatment of severe acute respiratory failure (SARF). Although most patients survive their first hit, there is a risk of extending suffering in those with irreversible organ failure. Since mechanical issues with the circuit are nowadays often no longer an issue and since ECMO is able to support failing organs for a long period of time, physicians often struggle with a new dilemma: should we continue this high level of care? We describe three representative cases and discuss the wide issues surrounding the initiation and withdrawl of ECMO.

Case Presentations: Case 1 – A 63 year old gentleman with SARF had ECMO withdrawn following a 73 day run due to irreversible respiratory failure. He was alert and orientated but completely dependent on his mechanical support. Therefore, both patient and physicians agreed for consenting him to organ donation before active withdrawal.

Case 2 – A 58 year old gentleman with SARF was successfully decannulated from VV-ECMO after 69 days of mechanical support, before developing ner onset type two respiratory failure and passing away.

Case 3 – A 53 year old gentleman with SARF was diagnosed with advanced lung adenocarcinoma after initiation of ECMO. A communication and understanding breakdown with the next of kin led to delays in initiating best supportive care. The patient deteriorated and passed away without family consensus on care.

Conclusions: There is little guidance in the literature about palliative withdrawl of ECMO. We advocate initiation of an infusion of anxiolytics and sedatives, with slow weaning of carbon dioxide removal to suppress air hunger.

Many patients are able to maintain a borderline organ function, but lost their reserves during the disease process; Therefore, even after successful decannulation, morbidity and mortality remain high. Care has to be taken not to extend suffering with aggressive treatment where there is minimal chance of full recovery.

At time of referral underlying exclusion criteria to ECMO may not have been identified. Optimism when initiating ECMO often leads families to want ongoing aggressive treatment despite irreversible organ failure. It is vital to take into consideration families beliefs and allow time to consider options, but the best interests of the patient remain paramount. Therefore, the initiation of good randomized trials about the in- and exclusion criteria for ECMO-start up should be considered.

The use of neostigmine for treatment of constipation in critical care; a case series

Phoebe D'Orton-Gibson and Eoghan O'Callaghan

Aintree University Hospital, Liverpool, UK

Abstract

Gastrointestinal dysmotility and constipation are common problems in intensive care and are associated with an increase in morbidity and mortality.¹ Previously the mainstay of treatment for refractory cases with pseudoobstruction was endoscopic or surgical decompression.² However more recently neostigmine, a cholinesterase inhibitor has been shown to be effective in this context, within critical care, with few complications. In one study of 21 patients with acute colonic pseudo-obstruction ten of the eleven patients on neostigmine achieved bowel opening compared to none in the placebo group.³ In a further study eleven out of thirteen patients with critical illness-related colonic ileus achieved bowel opening in the first 24 hours of treatment compared to none in the placebo group.⁴

We therefore hypothesized that administration of neostigmine to patients on critical care with constipation would result in bowel opening. A new bowel management guideline was introduced at University Hospital Aintree. This outlined a number of steps including optimisation of a patient's medical state, routine prescription of laxatives and imaging. Neostigmine infusion at 0.4 mg/h is considered after 96 hours without bowels open and titrated up to 0.8 mg/h if there is no response to the first 8 h of infusion.

This case series occurred 01/05/18–31/08/18. All were inpatients on the critical care unit at Aintree University Hospital. These patients had no bowel opening for at least 96 hours and had previously been treated with laxatives. They were excluded if they had any evidence of mechanical obstruction, major bowel dilatation, contraindications to neostigmine or recent history of abdominal surgery. As per protocol an infusion of neostigmine was commenced at 0.4 mg/h for 8 hours and increased to 0.8 mg/h for 16 hours if there was no initial response for up to a total of 24 hours.

5 patients were identified as suitable for neostigmine infusion. All 5 of the identified patients were male. The patients' ages ranged from 19–65 years. All patients opened their bowels while on the neostigmine infusion. Average time to bowel opening was 11.6 hours, range 2–25 hours. No complications reported in other case series (including ischaemic bowel, perforation or bradycardia) were observed.

This case series has shown that neostigmine is effective in achieving bowel opening for patients with refractory constipation in critical care. Whilst no patients in this case series experienced complications, the incidence of serious complications in this context remains to be measured. Neostigmine may nevertheless improve morbidity and mortality in critically ill patients by reducing constipation-related complications.

Critical Care Clinical Psychology – assessing the need and reviewing a new service

Gemma Mercer and Charlotte Cooper

University Hospitals of the North Midlands, Stoke on Trent, UK

Abstract

Introduction: Critical illness, injury and a stay within critical care present a number of challenges to both patients and their loved ones. The most urgent threat to deal with is the one to the patient’s life. Once this has been managed, more challenges can present themselves, particularly in relation to their emotional wellbeing and the psychological impact of surviving a critical illness (Aitken et al., 2013). Recent research has therefore advocated for Clinical Psychologists to be embedded into Critical Care Units across the UK in an attempt to reduce the psychological impact of a stay within Critical Care (Jackson et al., 2014; Wade et al., 2014; Wade et al., 2015). The NICE guidance on ‘Rehabilitation after critical illness in adults’ also recognises the ‘non-physical’ potential outcomes of a critical care stay as being an important target for treatment.

A Clinical Psychology service has recently been introduced into the Critical Care Unit at the Royal Stoke University Hospital, as part of the new multi-disciplinary rehabilitation team. The service consists of a Clinical Psychologist and an Assistant Psychologist working 2.5 days per week in the unit. We have reviewed the referrals sent to the service over a six month period, to examine the psychological needs being identified within this patient group.

Methods: All referrals received by the Clinical Psychology team were recorded over a six month period (March to August 2018). Information was obtained regarding referral source, reasons for referral and action taken.

Results: On average, ten referrals a month were received (range 7–15 per month). Reasons for referral included pre-morbid mental health issues, patients displaying symptoms of low mood and/or anxiety as a result of their stay and behaviours that challenge staff. Referrals for family support were made when it was identified that the anxiety or behaviour of the family was significantly impacting on the patient’s progress on the unit.

Fifty percent of patients were seen by the Clinical Psychology service. Of those not seen, approximately one third of referrals were found to be inappropriate, and the majority of the remainder had either died or were discharged prior to review. Approximately two thirds of referrals were made by the Rehabilitation Co-ordinators or during the weekly Rehabilitation Ward Round, which is likely to be due to these staff members having increased access to the Psychology team.

Conclusion: This review has evidenced that there is a demand for psychological input into critical care, although there is a need to develop the referral process to ensure the service is utilised most efficiently. Pre-morbid mental health problems are a significant reason for referral, however further work is required to identify those who may develop mental health problems as a consequence of their stay. Awareness of the service appears to influence the rates of referrals and suggests further education for the staff group may help to highlight the availability of clinical psychology input.

Comparison of one-year outcomes of younger and older patients undergoing emergency laparotomy admitted to a Critical Care Unit

Michael Jarvis¹, Nicholas Brazel², Som Sarkar³ and Nicholas Watson³

¹Stoke School of Anaesthesia, Stoke-on-Trent, UK

²East Midlands School of Anaesthesia, Nottingham, UK

³Sherwood Forest Hospitals NHS Foundation Trust, Sutton in Ashfield, UK

Abstract

Over the last decade, following the introduction of standards from the National Emergency Laparotomy Audit (NELA), significant improvements in outcomes have been noted, with mortality reduced to 10.6% and nearly 30,000 emergency laparotomies being undertaken across England and Wales each year (1,2). Mortality is higher with older patients, especially with increasing multimorbidity and frailty. Approximately half of patients undergoing emergency laparotomy are 70 years or older, which NELA defines as old age.

We compared 1-year outcomes for patients aged <70 years (<70 s) and aged 70 years and over (≥70s) admitted to the Critical Care Unit at Sherwood Forest Hospitals NHS Foundation Trust following Emergency Laparotomy; the NELA database only has outcomes to 90 days post operatively. NELA standards include patients with a risk of death ≥10% to be admitted to critical care post-operatively; locally we use ≥5% predicted mortality as the standard for critical care admission.

Results: Data were extracted from ICNARC, the NELA database and hospital data systems. Over the 3-years from January 2014 to December 2016, 465 emergency laparotomies were performed, 215 patients admitted to the Critical Care Unit. Of these, 106 were <70s and 109 ≥70s.

Our data showed good outcomes of 30-day mortality of <10% for both groups, with one-year mortality of 14.15% in the <70s and 24.77% in the ≥70s. Interestingly P-POSSUM score was the best predictor of 1-year mortality in the ≥70s. For critical care and 30 day mortality in the younger group, we demonstrate a link between mortality and APACHE II score (p = 0.027 for critical care mortality and p = 0.03 for 30 day mortality) and highest creatinine (p = 0.009 for critical care mortality and p = 0.025 for 30 day mortality), but find less of a correlation in the ≥70s (APACHE II p = 0.16 for critical care mortality and p = 0.047 for 30-day mortality, highest creatinine critical care mortality p = 0.16 and 30-day mortality p = 0.7, respectively). Due to the average short duration of cardiovascular or advanced respiratory support, we failed to show a link between duration of support and mortality. For both groups, there was a large increase in mortality between 30 days and one year. In the <70s, mortality doubled between 30 days and a year (7.55% to 14.15%), and almost trebled in the ≥70s (9.17% to 24.77%). There was no statistical link between APACHE II score or highest creatinine and 1-year mortality in either the under 70s or ≥70s (APACHE II <70s p = 0.76, ≥70s p = 0.68; Highest creatinine <70s p = 0.98, ≥70s p = 0.39).

NELA numbers in brackets

Conclusion: Our data suggests good outcomes in high risk patients admitted to Critical Care and to 30 days postoperatively, but importantly follow up at one year shows a significant increase in mortality after hospital discharge, which may influence decision making when consenting for emergency laparotomy in the elderly.

Admissions to Critical Care	<70 years	≥70 years
Total numbers	106 (13,838)	109 (11,054)
Mortality on  Critical Care	4.72%	5.50%
Hospital mortality	5.66%	9.17%
30-day mortality	7.55% (6.5%)	9.17% (15.7%)
90-day mortality   (NELA only)	(9.1%)	(21.2%)
One-year mortality	14.15%	24.77%

Setting ceilings of care – a threat or a role for the functioning of critical care outreach teams (CCOTs)?

Simi Paul, Vikki Raval and Jamie Gross

Northwick Park Hospital, London, UK

Abstract

Introduction: The main role of the critical care outreach team (CCOT) is to identify and institute timely treatment and appropriate escalation of care to patients who are deteriorating in hospital with a view to improve patient outcomes and reduce unplanned admission to ICU where possible.¹ However, there is an increasing trend of patients presenting to critical care services who may be frail, have significant chronic disease and/or in the last phase of life² and may not be suitable for full escalation of organ support. Despite national efforts to increase the uptake of advance care plans in this cohort³, their uptake is often low and inconsistent and decisions on ceilings of treatment often default to emergency situations when patients are acutely unwell.⁴ The purpose of this study is to investigate what impact this has on the CCOT.

Objective: To investigate the extent to which the CCOT facilitate ceiling of treatment decisions in a busy district general hospital with tertiary services.

Methods: 1-year prospective observational study at Northwick Park hospital between 1/5/2017 and 30/4/2018 that included all patients referred to the CCOT. Data was collected from a routine data source and included the total number of patients reviewed by the CCOT, the proportion who were reviewed by CCOT during a medical emergency team (MET) call, the duration of each CCOT review and the proportion who had ceilings of treatment set at a CCOT review. Data was also collected on the number of MET calls and ICU step-downs that were undesirably missed by the CCOT over the same time period.

Results: During the study period, there were 5694 reviews by the CCOT involving 3273 patients. 59% of reviews (n = 3385) were as part of a MET call. Out of the total number of patients reviewed by CCOT, 11% (356 patients) had a new ceiling of treatment set and 5% (175 patients) had a new DNACPR order set at the time of CCOT review. The mean time for a CCOT review was 53 minutes. There were 121 MET calls that were missed by the CCOT (3.6% of the total) and 27 out of 560 ICU step-downs (4.8%) that were not seen within 24 hours of ICU discharge during the study period.

Discussion: This study shows that a significant proportion of time is spent by the CCOT on facilitating the setting of new ceilings of care for those who might not be appropriate for full escalation in organ supportive treatment. Further studies are required to investigate whether or not these findings are replicated in other centres and if so, whether or not this is an appropriate use of CCOT resources or whether this negatively impacts on competing responsibilities of the CCOT. These include timely intervention and escalation of care in patients who are appropriate, timely review of patients discharged from ICU and training and education of ward staff to recognise and manage the early stages of critical illness.

Can We Feed?

Valerie Davey, Curtis Roberts, Craig Brandwood, Natalie Welsh and Andy O'Malley

Manchester University Foundation Trust, Manchester, UK

Abstract

Background: The Assessment, implementation and management of nutrition in critical care should be carried out by a dietitian with multidisciplinary (MDT) collaboration as a standard of care from GPICS (1).

Introduction: Review of feeding practices at the adult intensive care unit at the Manchester Royal Infirmary identified very limited assurance level. Following presentation at the critical care quality and safety meeting it was agreed for a nutrition working group to be formed. Initially the group had representation from a consultant with a nutritional interest, dietetics and nursing. This has subsequently expanded to include input from speech and language therapy, nutrition specialist nurse and pharmacy.

Aims: The purpose of the group is to

• To engage in a broad range of activities and promote a culture of best nutritional practices in critical care.

• To promote evidence based nutritional practices.

• To improve the nutritional care and clinical outcomes of our critically ill patients.

• To promote the different skills within the MDT.

• Assess and evaluate service provision.

Objectives:

• Update critical care feeding guidelines.

• Produce and deliver teaching to the Critical Care and MDT.

• Evidence based directed assessments.

• Weekly nutrition round

 ○ At a set time.

 ○ Members of the MDT are encouraged to participate including students.

 ○ Identification of nutritional risk to allow prioritization.

 ○ Assessment and discussions centre around an acronym ‘CAN WE FEED’ (2).

Discussion: The Critical Care Nutrition Group and subsequent Ward Round have promoted the different roles within the MDT and the unique skills each speciality has to offer. It has provided a platform to complete service evaluation and identify gaps in service provision as well as providing a forum to ask questions and consider opportunities for changing practice. Patient care has been improved by providing timely coordinated MDT treatment plans and promoting nutrition provision which is well evidenced to improve patient outcome (3).

Conclusion: The group continues to evolve and has the desire and scope to involve other professions. Going forward the group aims to facilitate extended practice roles for the different MDT members e.g. supplementary prescribing, bedside naso-jejunal tube placement and ultrasound muscle measurement. The group and ward round will partake in ongoing service evaluation and hope to highlight research opportunities that will continue to raise the profile of the team and promote the role of nutrition in Critical Care.

Are simpler interventions being overlooked in the treatment of out of hospital cardiac arrest (OHCA)?

Nicholas Toossi and David Slessor

Portsmouth Hospitals NHS Trust, Portsmouth, UK

Abstract

Background: To improve survival from an OHCA early bystander CPR must be initiated.¹ Despite this, only 40% of victims receive bystander CPR in the UK.² There are number of initiatives aimed at improving bystander CPR rates including the provision of targeted CPR training to family members of patients at elevated risk of cardiac arrest.³ We set out to find if victims prior to an OHCA had recently been admitted to hospital or attended an out-patient hospital appointment. Such interactions could represent opportunities to promote bystander CPR for patient’s families.

Methods: All patients admitted to the Queen Alexandra Hospital (Portsmouth hospitals NHS trust) intensive care unit following an OHCA between January 2017 and June 2018 were included. A retrospective observational study was performed analysing which of these patients did not receive bystander CPR. We investigated if patients from this group had recently been admitted to hospital or attended an out-patient hospital appointment for a significant co-morbid condition which could place them at a higher risk of cardiac arrest.

Results: Of the 57 patients who had an OHCA 25% (14/57) received no bystander CPR. 43% (6/14) of these arrests were witnessed by a friend or family member. 36% (5/14) of this group had a hospital admission or specialist outpatient appointment recently prior to their arrest. No documented advice with regards to bystander CPR was provided to the patient or a member of their family during these clinical interactions.

Conclusion: With regards to our study, a considerable proportion of patients who did not receive bystander CPR had recently been involved with medical services prior to their cardiac arrest. This represents a missed opportunity to provide targeted CPR training to family members. Proven modalities at improving CPR rates could be promoted during these clinical interactions including ultra-brief CPR training, DVD training, mobile phone apps etc.³

The prognosis after an OHCA has improved in the past few decades because of advances in interventions used outside and in hospital.⁴ In recent times, huge investment has befallen hospital interventions such as targeted temperature management and immediate coronary angiography with or without percutaneous coronary intervention. Nonetheless, the time it takes to initiate CPR has the greatest impact on survival.³ Therein lies the question, are local critical departments doing enough to support this simpler, albeit less sexy, form of intervention to manage this medical emergency?

Mucoactive agent use in UK Critical Care Units: A Survey of Unit-level Practice

Mark Borthwick¹, John Warburton², Rohan Anand³, Judy Bradley³, Bronwen Connolly⁴, Bronagh Blackwood³, Brenda O’Neill⁵, Murali Shyamsundar³, Chee Yap⁶ and Danny McAuley³

¹Oxford University Hospitals NHS Foundation Trust, Oxford, UK

²Department of Intensive Care Medicine, University Hospitals Bristol NHS Foundation Trust, Bristol, UK

³Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

⁴Guy's and St Thomas' NHS Foundation Trust, London, UK

⁵Centre for Health and Rehabilitation Technologies, INHR, Ulster University, Newtownabbey, UK

⁶School of Medicine, Dentistry and Biomedical Sciences, Queen's University Belfast, Belfast, UK

Abstract

Acute respiratory failure (ARF) is a frequent cause of admission to critical care. Associated mortality from ARF and acute respiratory distress syndrome (ARDS) has been reported at up to 40%.¹ Patients typically require invasive mechanical ventilation as part of management, which may be supplemented by use of mucoactive agents to optimise airway secretion clearance. However there are limited data to characterise practice in UK critical care units (CCUs), and empirically this is variable. We therefore conducted a survey to explore selection, indication for use, implementation, and supporting guidelines for mucoactive agents in UK CCUs.

Survey questions were developed, prepared in Survey Monkey (Survey Monkey Inc, San Mateo, California, USA), piloted by a reference group of pharmacists, and iteratively refined to reduce ambiguity. Respondents were asked to rank unit practice regarding expected indications and benefits of mucoactive agents, indicate therapies in use, and provide copies of guidelines used.

Due to familiarity with medicines use across this environment, pharmacists were targeted as respondents of choice to report practice.² Invitations to participate, including the survey link, were distributed electronically to UK CCU pharmacists via the UK Clinical Pharmacy Association on 04/04/18. Forward distribution to personal contacts was encouraged. The denominator for UK CCUs was defined as the number of units participating in the Intensive Care National Audit and Research Centre Case Mix Programme (England, Wales, Northern Ireland, n = 276) and Scottish Intensive Care Society Audit Group (n = 63). Reminders, including response rate by network, were circulated weekly until the survey closed 25/04/18. Final response data were downloaded to Excel (Microsoft Office, WA, USA) for analysis. Ranked data were converted into Borda³ counts. (ie, where there are seven categories, the responder’s first choice scores 7 points, second choice: 6 points, third: 5 points etc. Scores for each category from all respondents were summed to establish overall ranking).

Pharmacists representing 128 CCUs responded (37.5%; including 2 additional non-registered units). Mucoactives were reportedly used in 106/128 CCUs (82.8%). The most highly ranked indication for mucoactive use was “Thick Secretions” (Figure 1). The most highly ranked expected clinical benefits were firstly “Improved gas exchange”, and secondly “Reduced ventilation time”.

Figure 1.

Indications for mucoactive agents (Top) and expected clinical benefits (Bottom) [LOS; length of stay].

Figure 2.

Mucoactive agents used.

There was high use of topical isotonic sodium chloride solution and systemic carbocisteine (Figure 2). A wide range of topical hypertonic sodium chloride solutions and two concentrations of topical N-acetylcysteine were reported. Only five critical care units provided guidelines directing the use of mucoactive agents (3.9%).

These survey data demonstrate use of mucoactive agents in UK critical care units is frequent and practice is variable. The provision of guidelines to inform practice is uncommon. Whilst improvement in gas exchange and reductions in ventilation time were reported as anticipated therapeutic benefits, some respondents reported no expected benefit from the use of mucoactive agents. These data may reflect the limited high quality evidence available for these treatments, and the need for clinical trials to inform best practice.

Developing a Midodrine (Noradrenline Weaning) Protocol

Andrew Nash^1,2 and Nigel Chee¹

¹Royal Bournemouth Hospital, Bournemouth, UK

²Salisbury District Hospital, Salisbury, UK

Abstract

An inability to liberate patients from critical care is not infrequently due to ongoing vasopressive requirements, usually via a CVC in order to maintain a MAP >60 mmHg or targetted urine output. The management of this clinical conundrum varies, often with inappropriate fluid therapy/Echo guided therapy, or simply time unnecessarily on the ICU with no active intervention. This delay in declaring a patient “wardable” leads to increased hospital acquired infections, non-clinical transfers and cancelled elective major operations.

Methods: Based on a detailed literature review looking at safety data and use of oral/intramuscular vasopressors, a local protocol at Royal Bournemouth Hospital (RBH) was produced for liberating patients from Noradrenline infusions (the primary vasopressor in RBH ICU). Initially phenylephrine was considered, but the use of this outside of obstetrics and without intravenous use was limited. Subsequently, midodrine was identified as the ideal oral vasopressive agent.

The protocol is designed to be led by senior clinicians. Midodrine enterally is the primary intervention where possible.

46 patients liberated from ICU June 16-April 2018

43 had Midodrine enterally

2 patients phenylephrine subcutaneously

1 patient intramuscularly.

Patients were followed up for 3 days, with a review of their observations chart, U&Es and drug chart, as well as review any return to ICU.

Results: ALL 46 patients discharged at 12 hours after initiation of Midodrine to the ward.

26 patients 57% required 1 day of oral vasopressor (i.e up to 3 doses)

14 (30%) days in total.

6 (13%) required >48 hours

No patients returned to ICU.

No acute kidney injury (potential surrogate for low MAP) directly attributable to midodrine

One patient was weaned over 2 weeks.

Discussion: The protocol initially was not followed entirely as per guideline, reflecting local pressures and variation in personal practice. It has recently significantly improved as drug availability has been redressed. The cohort size has doubled in the past 9 months and is providing solid data, which hopefully the MIDAS trial will also reflect. This protocol is in common use at RBH and could become widespread and common place within intensive cares across the region.

Future plans include in the first instance, to collate more data, including limitations of discharge and time of delay from Noradrenaline requirement. The results of the international MIDAS trial will add strength to this work and hopefully concur with the findings, allowing it to transition easily into commonplace intensive care daily practice. Areas for use might include PACUs and ward based management of epidurals for example.

The Effect of Intermittent versus Continuous Enteral Nutrition on Muscle Wasting in Critical Illness: Safety and Feasibility

Angela McNelly^1,2, Danielle Bear³, Bronwen Connolly^3,4,5, Gill Arbane³, Laura Allum³, Philip Hopkins⁶, Matt Wise⁷, David Brealey⁸, Kieron Rooney⁹, Jason Cupitt¹⁰, Bryan Carr¹¹, Nicholas Hart^3,4, Hugh Montgomery^1,2 and Zudin Puthucheary^1,12

¹UCL Dept. Medicine, London, UK

²University College London Hospitals NHS Foundation Trust/UCL, National Institute for Health Research Biomedical Research Centre, London, UK

³St Thomas' Hospital, Lane Fox Clinical Respiratory Physiology Research Centre, London, UK

⁴Guy's and St. Thomas' NHS Foundation and King's College London, National Institute for Health Research Biomedical Research Centre, London, UK

⁵King's College London, Centre for Human and Applied Physiological Sciences, London, UK

⁶King's College Hospital, Intensive Care Medicine, Major Trauma & Anaesthesia, London, UK

⁷University Hospital of Wales, Dept. Critical Care, Cardiff, UK

⁸University College London Hospitals, Dept. Critical Care, London, UK

⁹Bristol Royal Infirmary, Dept. Critical Care, Bristol, UK

¹⁰Blackpool Victoria Hospital, Dept. Critical Care, Blackpool, UK

¹¹University Hospitals of North Midlands, Dept. Critical Care, Stoke-on-Trent, UK

¹²Royal Free London NHS Foundation Trust, Dept. Anaesthesia & Critical Care, London, UK

Abstract

Introduction: Loss of skeletal muscle during early critical illness can be profound¹, with major functional sequelae.² In contrast to continuous enteral feeding, which may inhibit muscle anabolism, intermittent feeding may promote it. The safety and feasibility of intermittent feeding in a randomized controlled intervention study is reported here.

Objective: To determine whether intermittent enteral feeding was safe and feasible during the first ten days of an intensive care unit (ICU) stay.

Methods: We studied adults with multi-organ failure, who were mechanically ventilated for ≥48 hours and remained on one of 8 UK ICUs for ≥7 days. Participants (n = 121) were randomized within 24 hours of admission to receive four-hourly intermittent bolus feed or (control) continuous enteral feed for the 10 day study period

For participants who received feed for ≥48 hours (n = 99), we recorded data up to day 10: the number of and reason for missed feed days, average daily energy and protein intakes, and percent target energy and protein intakes received. Data are reported descriptively and analysed using two sample t-tests or Mann-Whitney U-tests as appropriate.

Table 1.

Energy and Protein Requirements Met, and Mean Daily Energy and Protein Intake. IQR: Interquartile range; CI: Confidence interval.

	Energy Requirements Met, %			Daily Energy Intake, kCal
	Median	IQR	p value	Median	IQR	p value
Intermittent Feed, n = 52	83.2	17.6	0.12	1539	427	0.02
Continuous Feed, n = 47	77.2	29.2		1363	384
	Protein Requirements Met, %			Daily Protein Intake, g
	Mean	95%CI	p value	Mean	95%CI	p value
Intermittent Feed, n = 52	77.9	83.1–72.8	0.12	70.66	75.06–66.26	0.58
Continuous Feed, n = 47	71.6	77.9–65.3		68.65	74.55–62.75

Results: Of 99 participants (mean ± standard deviation age 57.4 ± 15.8 years; 68% male), 52 received intermittent feeding and 47 continuous, on a mean of 77% and 74% of the 10 maximum possible study feed days, respectively. ICU discharge prior to Day 10 was the main reason for missed feed days.

Key Findings: No intermittent feed-related serious adverse events were reported. Two participants transferred from intermittent to continuous feed, and one from intermittent to total parenteral nutrition, due to gastric residual volumes ≥300 mls or vomiting. Those intermittently fed received more calories: median ± interquartile range (IQR):1539 ± 427 vs 1363 ± 384 kCal respectively; p = 0.02. The proportion meeting their daily protein or energy requirements was similar in both groups (Table 1).

Conclusions: Delivery of intermittent enteral feed to ICU patients appears feasible and safe, and associated with greater median energy intake than is continuous enteral feeding.

Grant Acknowledgment: AM: Moulton Foundation; National Institute for Health Research (NIHR) Biomedical Research Centre (BRC) UCL Hospitals NHS Foundation Trust and UCL; BC, NH: NIHR BRC Guy's and St Thomas' NHS Foundation Trust and King's College London. HM: NIHR BRC UCL Hospitals NHS Foundation Trust and UCL. ZP: ASPEN Rhoads Research Foundation. The research team acknowledges support from the NIHR Clinical Research Network. The views expressed are the authors’ and are not necessarily those of the NHS, NIHR or Department of Health.

Ciaglia Blue Dolphin Technique for Percutaneous Dilatational Tracheostomy – A Single Centre Experience Dr D. Sehrawat, Dr P. Mukherjee, Dr. D. Acharya Watford General Hospital, Watford, Hertfordshire, UK

Deepak Sehrawat, Prasun Mukherjee and Devraj Acharaya

Watford General GHospital, Watford, UK

Abstract

Background: Percutaneous dilatational tracheostomy is the most common and well established surgical procedure in ICU due to its ease of insertion and safety. Ciaglia Blue Dolphin(CBD) is a relatively new technique which uses balloon inflation to create tracheal stoma instead of conventional dilatation with intended benefits of quick, reliable, safe dilatation and subsequent cannula placement in one single step. Other added benefits are less chances of tracheal ring and tracheal wall injury, less chance of airway leak and preservation of tissue elasticity.

Aim: To evaluate the Ciaglia Blue Dolphin (CBD) technique in terms of learning curve, procedural difficulty, number and nature of complications.

Methods: Single centre prospective observational study in Watford General Hospital ICU starting from 2010 till August 2018. CBD technique in our ICU is performed by single operator (Dr Acharaya) and standard technique described by manufacturer is used. Tracheostomy is performed under Fibre optic bronchoscope guidance and end tidal carbon dioxide monitor is also used. All procedure related difficulties, immediate and intermediate complications are noted.

Results: Of the total 266 percutaneous dilatational tracheostomies performed during this period, 92 were done using Blue Dolphin technique. In one patient, we used CBD technique after the initial surgical exposure by ENT surgeons in the operating theatre. There was a steep learning curve. In 2014 we performed more CBD tracheostomies (21) compared to Blue Rhino technique (12). During the earlier stages of Dolphin technique in our practice, in two patients we had to convert to Blue Rhino technique after failing to insert tracheostomy tube despite second dilatation. Two more patients needed second dilatation. There was one incidence of balloon rupture which was reported to the manufacturer. There was one incidence of guidewire kinking and curling up in the mouth following balloon dilatation. Three patients had minimal venous bleed. One patient had arterial bleed from anterior tracheal wall which stopped after placing the tracheostomy tube. None needed any surgical assistance or input.

Discussion: From our experience with Ciaglia Blue Dolphin dilatational technique, we can confirm that it is a safe and reliable technique. There is a steep learning curve and as suggested by the manufacturer, it is a single stage tracheostomy and technically easy to perform. Overall, there were no major complications or procedural difficulties. Larger randomised controlled trials are needed to compare this technique with other techniques of percutaneous tracheostomy.

Is it good to talk? A qualitative analysis of a shock call system in the management of cardiogenic shock

Deanna Gibbs¹, Cidalia Eusebio², Julie Sanders^2,3 and Alistair Proudfoot^2,3

¹Barts Health Trust, London, UK

²St Bartholomew's Hospital, Barts NHS Trust, London, UK

³Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

Abstract

Introduction: The acuity, complexity and mortality of cardiogenic shock (CS) mandates mobilization of resources, specifically mechanical circulatory support (MCS), early and effectively. We introduced a shock team and shock call platform (Everbridge Mass Notification) to facilitate real-time, multi-disciplinary (MDT) teleconferencing to facilitate support immediate and multi-disciplinary decision-making in CS. This study sought to explore perceptions of the impact of the shock call system on decision making in CS.

Methods: We employed a descriptive qualitative approach. One focus group interview was conducted with 12 multidisciplinary clinicians experienced in the shock call system. The focus group was audio-recorded, transcribed, and thematically analysed independently by two researchers to capture and describe the clinicians’ experience and perceptions of the platform with a focus on its utility in supporting decision making in CS.

Results: Nine intensivists, 1 heart failure cardiologist, 1 cardiothoracic surgeon and 1 interventional cardiologist participated in the focus group. Fifty-eight shock calls had been activated over a period of 13 months prior to the focus group. Analysis yielded six key themes: a) supported decision making; b) team communication; c) user/technology interface; d) managing the call – purpose and process; e) governance and learning; and f) the evolution of the platform and its wider applications. The first two themes will be presented in detail to highlight the influence of human factors on team engagement and behaviour. The theme of supported decision making described the mechanisms used by call participants in reaching a clinical decision. Respondents articulated the benefits of collegiate support, the credibility brought by group expertise, and the benefits of shared experience and learning. However, differing views were expressed on the necessity of reaching a consensus decision. The interplay between participants was further elaborated in the theme of team communication. Factors such as context, situational awareness, the need for structured communication via chairing of case discussions, and the individuals/specialties involved in each call all contributed to the effectiveness of team communication, and ultimately the perceived benefit of the call for planning patient care.

Conclusion: Human factors play a significant role in clinician perceptions of, and engagement with, a shock call system as a tool for real-time MDT discussions in CS. Technological solutions to support decision making in complex care scenarios must consider the influence of individual and team behaviour on the mechanisms by which decision-making technology is implemented. Our findings are being used to enhance shock call performance, enrich team communication and ultimately improve the utility of the system to optimise patient-care.

A positive NAFLD fibrosis score may predict increased perioperative blood transfusion requirements in patients undergoing cardiac surgery

Behrad Baharlo, Sarah Birch, Michael Webb, Janaarththanan Sakathevan, Mariam Ammoun, Andrew Chukwuemeka, Jonathan Cousins and Parind Patel

Imperial College Healthcare NHS Trust, London, UK

Abstract

Background: NAFLD (non-alcoholic fatty liver disease) shares many risk factors with cardiovascular disease, including the metabolic syndrome. Up to 80% of patients with metabolic syndrome have NAFLD on imaging. It has been postulated that NAFLD may pose a cardiovascular risk above and beyond that conferred by the traditional risk factors. The transfusion of blood and blood products in cardiac surgery is a known risk for post-operative morbidity and mortality.

We studied the impact of the NAFLD fibrosis score on transfusion requirements in a cohort of patients presenting for cardiac surgery.

Method: We prospectively studied 157 patients undergoing elective cardiac surgery at Hammersmith Hospital. Demographic, clinical (medical history, Euroscore II and peri-operative blood results) and transfusion data for that hospital episode were collected. Metabolic syndrome status using the NCEP ATP III criteria (2005 revision) was established. A pre-operative NAFLD fibrosis score (present, intermediate or absent) was calculated for each patient. Univariate analysis of the effect of the NAFLD fibrosis score on transfusion requirements (fresh frozen plasma, cryoprecipitate, platelets and packed red cells) was carried out using ANOVA for three cohorts (total population, metabolic syndrome and non-metabolic syndrome).

As higher risk patients receive more transfusions, where NAFLD groups differed in Euroscore II risk, further analysis against a low risk group (Euroscore II defined mortality <2%) was carried out to minimise the effect of surgical and patient risk to our results.

Results: Of 157 patients, NAFLD fibrosis score was ‘present’ in 50, ‘intermediate’ in 77 and ‘absent’ in 30.

Patients with a ‘present’ NAFLD fibrosis score had increased total products (n = 157, p = 0.02) and red cell (n = 157, p = 0.004) transfusions.

In those without metabolic syndrome, similarly an increase in total products (n = 84, p = 0.012) and a trend towards increased red cell (n = 84, p = 0.07) transfusions was seen with a ‘present’ NAFLD fibrosis score.

The NAFLD score groups for these two cohorts differed in Euroscore II risk (total population p = 0.04, non-metabolic syndrome cohort p = 0.001).

When only low risk individuals were analysed, a trend (p = 0.079) towards increased red cell transfusions was seen in the total population. In the cohort without metabolic syndrome there was a significant increase in red cell transfusion (p = 0.014).

In the metabolic syndrome cohort (n = 73) the NAFLD score groups were well matched for pre-operative Euroscore II (mean = 1.39%, median = 1.09%). An increase in red cell transfusion (p = 0.03) was seen in those with NAFLD score ‘present’.

Discussion: Our results suggest that in unselected patients undergoing low risk cardiac surgery (Euroscore II <2%), there may be an increase in peri-operative blood transfusions with a ‘present’ pre-operative NAFLD fibrosis score.

In patients with metabolic syndrome, there is a significant increase in blood transfusions (p = 0.03) in those with a ‘present’ pre-operative NAFLD fibrosis score.

Our findings warrant further investigation. The NAFLD fibrosis score may be useful in identifying transfusion requirements and peri-operative risk over and above that traditionally recognised, especially in the presence of metabolic syndrome.

Augmented renal clearance in critically ill patients: a single centre cohort study of creatinine clearance in intensive care

Brian Johnston¹, David Perry², Martyn Habgood², Miland Josho³ and Anton Krige²

¹Royal Liverpool University Hospital, Liverpool, UK

²East Lancashire Hospital Trust, Blackburn, UK

³University of Central Lancashire, Preston, UK

Abstract

Introduction: Augmented renal clearance (ARC) is poorly appreciated but conventionally characterized by a creatinine clearance greater than 130 ml/min/1.73 m. Modification of drug doses in patients with reduced kidney function is well established. However, in patients with ARC, dose escalation may in fact be appropriate to avoid sub-therapeutic plasma concentration. Improved guidance on recognition and the optimisation of drug dosing in patients with ARC is urgently needed.

The Royal Blackburn Teaching Hospital (RBTH) critical care unit introduced measurement of daily urinary creatinine clearance in January 2014. We retrospectively analysed all admissions from 2014 for the development of ARC with the aim of characterising those patients who are at increased risk of ARC.

Methods: We conducted a single centre retrospective cohort study of admissions to RBTH ITU between January 2014 – December 2016. Daily 6-hourly urinary creatinine clearance was recorded. ARC was defined as creatinine clearance >130 ml/min/1.73 m. Patients with renal impairment (Serum creatinine levels >110 mmol/l) or patients that underwent renal replacement therapy were excluded from analysis. Demographic data, past medical history and critical care admission data were extracted from the CMP data entered into the Wardwatcher database and the hospital PAS system. Blood results including, glucose and serum creatinine and requirement for vasopressors during admission were recorded. Prevalence of ARC was determined, and repeated measures logistic regression using STATA 15 was used to identify risk factors for developing ARC.

Results: Complete data was available for 1328 patients. Mean age for our cohort was 58.7 years (SD 6.8). ARC was observed in 624 (47%) admissions. The explanatory variables that were examined were gender, APACHE II score, admission diagnosis (sepsis, post-operative without sepsis, pancreatitis, trauma), presence of diabetes mellitus, pre-admission estimated glomerular filtration rate and serum glucose levels. ARC was more common in males 56.9% compared to females 35.5% (p < 0.001). ARC was significantly associated with younger age (p < 0.001). There was no threshold phenomenon with age before which ARC was more likely. ARC was significantly associated with lower APACHE II scores (p < 0.001). There were no statistically significant association between ARC and admission diagnosis, presence of diabetes mellitus, level of cardiovascular support required or serum glucose levels during admission.

Discussion: ARC is common finding in patients admitted to ITU. Our study is consistent with previous reports of ARC that suggests ARC affects between 35%-65% of admissions.¹ In multivariate analysis ARC appears to affect younger male patients with lower APACHE II scores. Similar findings have been reported for ARC and APACHE II however, no association between ARC and other severity of illness scores have been reported suggesting the association with APACHE II may reflect the influence of age on the score. ARC in ITU patients potentially leads to increased renal clearance of drugs and has been related to subtherapeutic levels of beta-lactams, vancomycin, gentamicin, anticoagulants and anti-epileptic drugs.^1,2,3,4

The physiological mechanism responsible for ARC and its relationship with clinical outcomes remain ill defined.¹ Given its high frequency in ITU patients, research urgently need to define the impact of ARC on drug clearance and dosage in these patients.

Prevalence of nutritional-related symptoms in discharged ventilated adult ICU patients – The pilot Symptoms and Nutrition After Critical Care (SNACC) survey Louise Albrich BSc, RD, Yeovil District Hospital; Mary Hickson, PhD, RD, University of Plymouth

Louise Albrich¹ and Mary Hickson²

¹Yeovil District Hospital, Yeovil, UK

²University of Plymouth, Plymouth, UK

Abstract

Background and Aim: Critical illness causes physiological, physical and psychological impairments which determine patients’ recovery ⁽¹⁾. Impairments may affect nutritional intake (e.g. dysphagia) or may be associated with suboptimal nutritional status (e.g. weakness) ^(2,3). The prevalence of such symptoms after intensive care unit (ICU) stay is unclear. This pilot survey aimed to collect preliminary data on prevalence and severity of nutrition-related symptoms reported by previously ventilated ICU patients, whilst exploring associations with length of stay (LOS) and time since discharge.

Methods: A multicentre cross-sectional survey using convenience sampling, invited discharged ICU patients, who had at least 2 days ventilation, to complete a survey (figure 1). Prevalence and severity of 30 symptoms were self-reported along with demographic and clinical data.

Results: 74 participants completed surveys (n = 74/172, 43% response rate) predominantly by phone (47%). 73% (n = 51/74) participants were 51–80 years old and median body mass index was 29 kg/m.² 60% (n = 44/74) had seven days or more of ventilation, median ICU-LOS was 14 days (IQR = 8–25), and hospital-LOS 33 days (IQR = 19–56). Participants reported a median of 13 post-ICU symptoms (IQR = 10–17) at a median 86 days (IQR = 43–178) post-ICU. The most prevalent gut symptom was being ‘not hungry’ (62.2%; n = 46/73) and being ‘low in mood’ the most prevalent psychological symptom (57%; n = 42/74). Both predominantly of ‘moderate’ severity. Physically, 87% (n = 64/74) participants reported having ‘lost muscle’, with most being ‘severe’ (45%, n = 33/74) (figure 2).

Longer hospital-LOS had a weak positive correlation with the number of symptoms (r = 0.23, p = 0.049). Patients reporting weakness were more likely to have longer stays on ICU (>4 days) (n = 52/55, 95% versus n = 14/19, 74%; p = 0.023, Fisher’s exact). Similarly, those reporting ‘not being hungry’ were more likely to have longer ICU-LOS (n = 44/46, 96% versus n = 22/28, 79%; p = 0.047, Fisher’s exact). Patients reporting ‘early satiety’ were more likely to have longer hospital-LOS (n = 39/43, 91%, versus n = 10/16, 63%, p = 0.018, Fisher’s exact). Those reporting nail changes were more likely over 3 months since ICU-discharge (n = 17/26, 65% versus n = 12/35, 34%; c²= 4.61; p = 0.032, Phi = 0.31), whereas those reporting a ‘gain in weight but not muscle’ being more likely to be after 6 months post ICU-discharge (n = 14/42, 33.3% versus n = 1/21, 5%; p = 0.013, Fisher’s exact). Patients reporting ‘muscle loss’ were more likely to have ward-LOS above 14 days (n = 32/64, 50% versus n = 1/5, 20%; p = 0.007, Fisher’s exact).

Discussion: In-patient recruitment provided most participants and telephone completion was preferred. Reports of reduced hunger, desire to eat, and early satiety concur with previous studies, but poor appetite was higher than previously reported (29%) ⁽⁴⁾. Swallowing problem prevalence is consistent with previous data (23%) in the 6 months post-ICU ⁽³⁾. LOS was associated with most symptoms except appetite which differs from previous data ⁽²⁾. Hair loss corresponds with other data (27%) ⁽⁵⁾ and nail changes manifest after three months, indicating the importance of defining the timeframe of symptom assessment.

Conclusion: Many nutrition related symptoms may be prevalent in ventilated patients post ICU, and LOS is an important associated factor. This data can help services prepare patients better for their recovery through timely symptom monitoring, advice and treatment.

Organ donation in patients following out-of-hospital cardiac arrest

William Gaunt, Jamie Hooker and Alistair Meikle

Critical Care Department, Crosshouse Hospital., Kilmarnock, UK

Abstract

Organ donation (OD) can radically improve the quality of life of those in need of an organ transplant. Donation of organs after circulatory death (DCD) represents one way in which the pool of potential organ donors can be increased.

Survivors of out-of-hospital cardiac arrest (OoHCA) who remain neurologically impaired following resuscitation are a group of patients who may be considered for DCD. Prediction of which of these patients, on withdrawal of life sustaining therapy (WLST), are likely to be suitable for DCD can be challenging. For their organs to be utilised for transplantation, death must occur within around three hours following WLST.

Often this group of patients are on minimal cardiovascular and respiratory support prior to WLST, and there is a perception that time to death can be prolonged well beyond the time frame for donation to be feasible.

Many of the published studies advocate referral of any individual who may be a potential donor. This approach is not only costly in terms of resources, but can be challenging for the families of the individual concerned, especially where DCD does not occur due to prolonged warm ischaemic time.

Following a study by Tordoff and Bodenham¹, we were interested to see how our rates of DCD in patients who presented to our ICU after having suffered an OoHCA compared to those published in their study.

The case notes of all patients admitted to our ICU following OoHCA were retrospectively analysed between May 2015-May 2018. We utilised a data collection tool which was based on that of Tordoff and Bodenham.¹ In addition we collected data on ventilator and cardiovascular support parameters immediately prior to WLST.

Of the 32 patients analysed in our study 19 were male and 13 were female. The median age was 60 years. Initial rhythms were asystole (11), VF (10), PEA (7) and unknown (4). Median time of collapse to ROSC was 32 minutes (4–100 mins). 25/32 (78%) patients died within hospital. 21/25 (84%) had WLST. Median time to asystole following WLST was 130 mins (5–8640 mins). 10/21 (48%) patients were referred for DCD, and 6/10 (60%) of families were approached for authorisation. Of 6 families approached 2 agreed (33%). 2 patients proceeded to DCD. Analysis of specific cardiovascular and respiratory support parameters prior to WLST failed to determine robust predictive factors to indicate death was likely within a three hour time frame.

Whilst our study population was limited in size, precluding any formal statistical analysis, our DCD rates and successful identification of suitable DCD donors, appear to be proportionately higher in our ICU (2/32 vs. 1/100). Whilst, from a basic subgroup analysis, it is tempting to draw some conclusions regarding potential identifiable characteristics of our patient population which may be reasonable predictive indicators for successful DCD, further work is necessary, with a more formal analytical approach.

As was previously indicated by Tordoff and Bodenham¹, our data has further highlighted the difficulty in predicting those individuals who have suffered an OOCHA and who may be suitable candidates for DCD.

IL-18 is upregulated in the pulmonary microenvironment in human experimental lung injury and the acute respiratory distress syndrome

Sam Lockhart, Yvonne Dombrowski, Rachel Bell, Michael McKelvey, Andrew Boyle, John Conlon, Danny McAuley and Cecilia O'Kane

Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

Abstract

Background: Interleukin-18 (IL-18) is upregulated in the systemic circulation of patients with the Acute Respiratory Distress Syndrome (ARDS) and predicts mortality. The predominant source of IL-18 in ARDS is currently unclear and it is unknown if IL-18 is upregulated in the pulmonary microenvironment of ARDS. The actions of IL-18 are context dependent, therefore, to harness the potential of targeting IL-18 in ARDS a detailed understanding of its production and regulation in human lung injury is essential.

We hypothesised that IL-18 is upregulated in the pulmonary microenvironment in ARDS, and that this occurs due to local IL-18 production rather than delivery from the systemic circulation.

Methods: We analysed stored plasma and BAL in a phase II clinical trial in ARDS [1]. To elicit the specific source of IL-18 in the context of lung injury we utilised a human model of ARDS in which healthy subjects inhaled lipopolysaccharide (LPS) [2]. BAL and plasma from healthy volunteers that did not inhale LPS were used as controls. IL-18 was measured in bronchoalveolar lavage (BAL) and plasma by ELISA. RNA was isolated from BAL cell pellets, and gene expression measured by RT-PCR.

Results: Plasma IL-18 was significantly elevated at baseline in patients with ARDS (Figure 1A). Importantly, BAL fluid IL-18 was also upregulated 3.5 fold at baseline compared to control (ARDS: 53.7 ± 11.35 pg/ml, Control: 15.4 ± 1.6 pg/ml, P < 0.01), and BAL fluid IL-18 remained >3-fold elevated 3 days after baseline measurement (Figure 1B). To examine if IL-18 was produced in the lung or delivered from systemic circulation we measured IL-18 in BAL fluid from participants 6 hours after LPS inhalation, a timepoint at which we have found IL-18 is not upregulated in the systemic circulation (LPS: 273 ± 26.4 pg/ml, Control: 247.9 ± 14.45 pg/ml, P = 0.59). BAL IL-18 was higher in the LPS group (P = 0.05, Figure 2). Consistent with this, IL18 mRNA was 3.5-fold higher in cells isolated from LPS BAL fluid versus Control while the negative regulator of IL-18, IL-18BP was unchanged. In, addition, the IL-18 target genes IFNG and TBX21 were upregulated 5.5-fold and 3.6-fold suggesting enhanced IL-18 bioactivity.

Conclusion: Our data demonstrate, for the first time, that IL-18 is upregulated in the pulmonary microenvironment and that this persists to at least day 3 of ARDS. Furthermore, in an inhaled LPS model of ARDS, alveolar IL-18 was higher compared to control, without an increase in its inhibitor, and associated with an increase in transcription of its downstream mediators. Further investigation of the role of lung-derived IL-18 in ARDS is needed.

Out-of-hours admission to intensive care after planned surgery increases patient mortality: a retrospective cohort study in a district general hospital

Jennifer Ridenton, Guy Hooper and Michael Spivey

Royal Cornwall Hospital, Truro, UK

Abstract

Introduction: Hospital admission during weekends has been associated with increased mortality and adverse outcomes.^1–6 Policy makers in England have used this evidence to justify a controversial ‘7-day service’ across the National Health Service on assumptions that reduced weekend service provision underlies adverse outcomes.⁷

Previous literature on the ‘weekend effect’ focused on emergency admissions or ward-based care. Data was limited to admission day and didn’t consider admission time. If adverse outcomes correlate with service provision, then patients admitted out-of-hours (OOH) are comparably likely to be affected.

With shortage of healthcare resources, major elective surgery is increasingly performed out-of-hours and on weekends, including cases requiring admission to intensive care units (ICU). Higher nursing ratios and monitoring in ICU should theoretically mitigate risks of these surgeries. Recent evidence suggests that planned surgical admissions to ICU out-of-hours (weekdays 1800-0759, Friday 1800 – Monday 0759) are associated with increased length of stay (LoS) and mortality.⁸

We used data from the Intensive Care National Audit and Research Centre (ICNARC) to examine the ‘out-of-hours effect’ and ‘weekend effect’ on patient outcomes, including LoS and all-cause hospital mortality for planned surgical admissions to ICU.

Methods: This retrospective cohort study obtained and analysed the ICNARC data set. All patients who had elective surgery requiring adult ICU admission in the Royal Cornwall District General Hospital between October 2012 and May 2018 were included. Primary outcomes were hospital LoS and all-cause hospital mortality.

Results: Of the total 897 planned post-operative admissions, 56.4% occurred out-of-hours. Compared with in-hours, OOH ICU admissions were associated with increased LoS and increased crude hospital mortality rate. In patients whose planned admission was delayed until the weekend, we found an increased LoS (Average LoS: 2.7 days) compared with planned in-hour admissions (Average LoS: 2.0 days). Weekday out-of-hours admissions were associated with the highest mortality rate (Actual hospital mortality: 1.9%), versus in-hours (actual hospital mortality: 1.3%). Delayed admissions, irrespective of admission time, were associated with increased mortality. Of the 14 hospital mortalities from planned post-operative ICU admissions, 28.6% were delayed admissions.

Discussion: Despite consistent staffing ratios and higher levels of monitoring in ICU, crude hospital mortality rates were higher for patients admitted out-of-hours and associated with increased LoS compared with ‘in-hours’ admissions. As with previous work, delayed admission to ICU, irrespective of time or day, is associated with adverse outcomes.⁹ Our study addressed limitations of previous literature by examining disparities of patient outcomes due to the ‘out-of-hours effect’ and ‘weekend effect’. Although statistical significance was not obtained due to small study size, the results are consistent with statistically significant data published in Australia and New Zealand.⁸ This study was based on data from a single DGH, therefore the generalisability of our results is limited to similar hospitals.

Conclusion: In addition to the ‘weekend effect’, patients admitted to ICU out-of-hours after planned surgery have an increased mortality rate and LoS. Delayed admissions, regardless of time, are associated with adverse outcomes. This study’s findings mirror those of the ANZICS group and should have implications on planned major surgeries requiring ICU.

Booked admissions to critical care; are they accurate?

Ian Sheldrake¹ and Spike Briggs²

¹Queen Alexandra Hospital, Portsmouth, UK

²Poole Hospital, Poole, UK

Abstract

Safe peri-operative care is essential to effective surgical treatment. Patients are frequently admitted to critical care units post-operatively to prevent and treat significant complications. The proportion of elective surgical procedures cancelled due to lack of availability of a post-operative critical care bed is Quality Dashboard Indicator (ACC13) issued by NHS England.¹ We aimed to determine the number of patients cancelled due to lack of bed availability, and the proportion of patients admitted to the unit having had a bed booked pre-operatively. The existing system used hand-written diary entries with only basic patient demographics. No details of co-morbidities or why the patient needed a critical care bed was recorded. Often, there were no details of the responsible consultant or planned procedure.

Methods: 32 patients were identified in the diary from November 2014, December 2014 and January 2015. Information was gathered from the surgical bookings office and ICIP (IntelliVue Clinical Information Portfolio, Phillips) to ascertain which patients were admitted to critical care. A Critical Care bed booking proforma was developed and distributed to all surgical secretaries. An email was sent to all consultants explaining that surgical teams were now required to deliver the completed form to Critical Care to ‘book’ a bed. The process was re-audited in 2017. 64 patients had elective critical care beds booked across January, February, March and April. Information on individual patients was collected via the same method.

Results: The initial audit showed 57% (n = 16) of booked admissions were admitted to the unit. 4 patients were excluded due to lack of data. 2 patients (7%) were cancelled on the day of surgery specifically because of lack of HDU bed availability. 6 patients (38%) were admitted for Level 3 care. The mean length of stay on the unit was 3 days. The case mix of patients admitted was 75% (n = 12) Maxillofacial/ENT surgery, 13% (n = 2) General Surgery, 6% (n = 1) Gynaecology and 6% (n = 1) Orthopaedics. On re-audit, 51% of patients (n = 30) were admitted to the unit post-operatively. 5 patients were excluded due to lack of data. 2 patients (3%) were cancelled specifically due to lack of HDU bed availability with 13 patients (43%) admitted for Level 3 care. The case mix of booked patients was 23% (n = 15) Maxillofacial/ENT surgery, 31% (n = 20) General Surgery, 11% (n = 7) Gynaecology and 34% (n = 22) Orthopaedics. In contrast, the case mix of admitted patient was 37% (n = 11) Maxillofacial/ENT surgery, 27% (n = 8) General Surgery, 7% (n = 2) Gynaecology and 30% (n = 9) Orthopaedics.

Discussion: A smaller proportion of patients were cancelled due to lack of bed availability. It is difficult to ascertain whether this was attributable to the new booking proforma or other factors. The new system ensures more data on patient co-morbidities is available to ICU clinicians to aid decision making. There is a significantly different case mix between the initial audit and re-audit. This would suggest some specialities and being more proactive than reactive in booking critical care beds pre-operatively as the process has been clarified.

Renal dysfunction occurring in the general ICU patients of a district general hospital is associated with nearly half of antibiotic doses being incorrect

Liam Roberts and David Stanley

Dudley Group of Hospitals NHS Foundation Trust, Dudley, UK

Abstract

Introduction: Sepsis is a common cause of admission to intensive care units (ICU) both in the UK and worldwide with significant associated morbidity and mortality. ICU admissions are also frequently and significantly complicated by sepsis. Antimicrobial therapy is one of the few evidence-based treatments offered in ICU and is a mainstay of the management of sepsis. Abnormal renal function is often concurrent with sepsis, either as a pre-existing co-morbidity or resulting acute end-organ dysfunction. Renal function in ICU patients can be dynamic as patients improve or deteriorate. Effective antimicrobial therapy is dependent on appropriate dosing relative to renal function to ensure therapeutic plasma concentrations. Prescribing and administration of these drugs were reviewed in a district general hospital ICU using an electronic patient record (EPR), which included both laboratory result reporting and prescribing functions.

Methods: Antimicrobials that were included within the Trust's antimicrobial prescribing guidelines and required renal dose adjustment were identified. Those which required daily interval blood tests prior to administration (gentamicin and vancomycin) were discounted. The antimicrobials remaining were: meropenem, levofloxacin and ciprofloxacin. The drug chart records for these drugs covering a continuous 13 month period were extracted from the EPR. These were then analysed to determine whether each administered dose was correct given the most recent renal function result, dose interval and dose mass. Non-compliant dosing was then identified as either under- or overdosing. Each patient's notes were then reviewed to ensure there were no special circumstances (i.e. neurological infection) or microbiology input to account for the dose administered.

Results: 41 doses of Ciprofloxacin were administered. 34 (82.9%) of these were compliant (6/14.6% underdosed, 1/2.4% overdosed). 16 doses of levofloxacin were administered. 13 (81.3%) of these were compliant (1/6.3% underdosed, 3/18.8% overdosed). 1658 doses of meropenem were administered. 1286 (77.6%) of these were compliant (93/5.6% underdosed, 271/16.3% overdosed). When only those meropenem doses given with renal dysfunction present were analysed, 276 (48.2%) of a total of 573 doses were compliant.

Conclusions:

• An insufficient number of ciprofloxacin and levofloxacin doses were administered to allow any substantial conclusions to be drawn.

• Meropenem dosing deviated significantly from the Trust guidelines both in terms of under- and overdosing.

• In the patient cohort with renal dysfunction, this was more significant and is possibly due to the default prescription setting (for normal renal function).

• Overdosing with fluoroquinolones and carbapenems especially is generally not as concerning as underdosing given the relatively non-toxic nature of these classes of drug and the greater risk posed to this group of patients by undertreatment of the infective process underlying sepsis.

• Such a significant shortfall in prescribing requires addressing and thus the following measures have been introduced: pharmacists will ensure antimicrobial prescriptions are checked against renal function on each daily drug chart review, education has been provided to medical staff, a renal dosing advisory has been added to the meropenem prescription – this will display all prescription options on selection forcing the prescriber to discriminate between them.

• Re-audit will be performed to ensure there has been an improvement in compliance.

Functional and physical recovery of Intensive Care Unit survivors at follow-up

James Bruce, Mark Rezk, Greg Warren and Tom Clark

Torbay and South Devon NHS Trust, Torquay, UK

Abstract

Introduction: Intensive Care Unit survivors may experience long-term impairment of their functional and physical capabilities.¹ A proactive rehabilitation programme may help to improve recovery from critical illness. It is our aim to better understand what level of functional and physical impairment ICU survivors experience and optimise their recovery.

Methods: Functional Independence Measure (FIM, 19 domains, scored 1 to 7) and Chelsea Critical Care Physical Assessment (CCCPAS, 10 domains, scored 0 to 5) data is routinely collected for all ICU patients screened at 48 hours as being ‘high-risk’ of needing rehabilitation. Pre-admission FIM score is estimated for each patient and CCCPAS is collected on day 4 and both at ICU discharge, hospital discharge and follow-up clinic. All patients admitted to ICU are assigned an estimated Frailty Score (FS, scored 1 to 9) within 48 hours. This data is stored in a local Trust database for quality improvement purposes alongside basic demographics (sex, age), co-morbid data (SOFA score, past medical history, primary ICU diagnosis, baseline bloods) and length of mechanical ventilation. Descriptive statistics calculated using Excel.

Figure 1.

FIM out of 133. CCCPAS out of 55.

Results: Data analysed from first 100 patients included in extended follow-up programme. 82 patients seen in follow-up clinic at a mean time after hospital discharge of 46 days (95% CI +- 9 days). Of 18 patients not followed-up, 11 had died whilst data was incomplete for 7. Median FS of cohort was 3. Median SOFA 6 (IQ range 3–7). 13 patients were ‘frail’ (FS > 4) and 87 patients were ‘non-frail’ (FS < 5), of which 31 were ‘vulnerable’ (FS 4). All ‘frail’ patients survived to follow-up. Figure 1 shows median, interquartile range and maximum/minimum values for FIM and CCCPAS scores at (1) pre-admission/day 4 ICU (2) ICU discharge (3) hospital discharge and (4) follow-up.

Conclusion: Critical illness appears to have a detrimental impact on functional and physical measures. Recovery after survival in patients assessed as being ‘at high-risk’ of needing rehabilitation is significant in those followed-up in hospital and after discharge by our rehabilitation team.

Reaching bedside staff with targeted education using social media: what does it cost?

Fung Kei Ng^1,2, Sarah Wallace¹, Barry Coe¹, Anna Owen¹, James Lynch¹ and Brendan A McGrath^1,3

¹Manchester University NHS Foundation Trust, Manchester, UK

²Health Education North West (North Western), North West England, UK

³Division of Infection, Immunity & Respiratory Medicine, Faculty of Biology, Medicine & Health, The University of Manchester, Manchester, UK

Abstract

The National Tracheostomy Safety Project (NTSP) is an example of a group publishing guidance and resources to improve the quality and safety of care for critically ill patients ^1,2. However, publications in traditional medical literature often bypass bedside staff and represent a missed opportunity to influence staff behaviours which may translate into better patient care. Many staff frequently use social media, accessed on personal devices, offering an opportunity to directly target relevant staff with tailored educational resources. We aimed to develop a strategy to target such staff as part of the Evidence into Practice award from the Health Foundation, an independent UK charity for better heath and healthcare.

We commissioned a multimedia marketing and production company (FeedFirst Ltd, feedfirst.co.uk, Manchester, UK) to conduct a social listening and profiling exercise, learning from competitors/influencers, and interviews and surveys with target staff groups on e.g. what information they want and how they want it. We produced 26 short (<2 mins) films regarding key elements tracheostomy care and branded these to match our website (www.tracheostomy.org.uk). We profiled our target audience on common social media platforms using Facebook Insights (Facebook, CA) with Facebook analytics providing the most robust results detailing delivery of our content: reach (number of users), impressions (delivery to any user’s page), views (>3 seconds for video) and clicks (user interactions with the media).

Insights identified a potential audience 1.4 million on Facebook, 260k on LinkedIn 272 k on Twitter. It was clear from a global perspective that spikes in social media activity were related to healthcare events or media reports regarding tracheostomy care. We therefore planned a campaign to highlight the NTSP’s involvement in the Global Tracheostomy Collaborative’s International Tracheostomy Symposium in Dallas, Texas, in February 2018. We spent £4,140 on paid social media advertising, generating 629,270 impressions. Our Facebook campaign reached 433,852 users with Facebook reporting a 13.05% uplift in brand identity and calculating 32,200 users would remember our content. We recorded 242,451 views and 6,165 clicks at a cost of £0.02 per view and £0.67 per click.

Our campaign and strategy reached target bedside staff with relevant resources. Our content demonstrated significant social mobilisation, with around two thirds of all views arising from organic peer-to-peer sharing (not paid). We believe that this intelligence and innovative approach to dissemination of educational resources represents value for money.

Reducing harm from temporary haemodialysis (HD) catheters using an additional X-ray wire check

Gavin Crawford and Dewi Williams

Dumfries & Galloway Royal Infirmary, Dumfries, UK

Abstract

The use of imaging to aid the safe placement of an intravenous catheter is now well-established practice. Ultrasound allows the user to visualize and access the vein in real-time, reducing the likelihood of accidental arterial puncture and other associated complications, as well as reassurance of wire position prior to vessel dilation. It is well recognised that this reduces procedural complications (1). X-ray performed after insertion, confirms the expected course of the catheter and position of the catheter tip. Most tunnelled HD catheters are inserted under fluoroscopy however this is not the case for most temporary HD lines.

Free passage of a guide-wire following ultrasound-guided venopuncture tells little of the course of the wire within the body. Using conventional Seldinger technique, malposition of the catheter (e.g. via the subclavian vein following the left Internal jugular approach) would only then be confirmed at the post-insertion X-ray check. HD catheters are stiff and are unable to pass freely round sharp corners, therefore having the potential to cause greater trauma than other more flexible catheters.

Following a recent critical incident (haemothorax) in our institution involving the insertion of a dialysis catheter, we have sought to implement steps to reduce patient harm through the routine use of an additional X-ray prior to dilation of the tract.

We position an X-ray plate beneath the patient prior to commencing the procedure. This gives the user the ability to confirm the correct course of the guide-wire prior to dilation of the tract and insertion of the catheter. We piloted this methodology in 15 patients to confirm if this was feasible in everyday clinical practice. Cooperation of our radiographers was vital as this additional step increases their time at the bedside. We found no problems in this regard and they supported the pilot enthusiastically.

We recognise that this technique requires an additional chest X-ray, however we consider the potential benefits to the patient. This technique allows for timely recognition of wire malposition and re-siting over a cannula without compromising sterility. Where this is unsuccessful, those patients who require fluoroscopy can be identified. (This was necessary in one out of fifteen patients). In one particular case, there was significant malposition of a guide-wire where all other conventional ‘checks’ did not suggest any issue, and most notably where dilation of the tract may well have resulted in patient harm. We therefore suggest this methodology has potential to reduce risk associated with HD catheter insertion in critical care.

A 2-year retrospective review of a Scottish Emergency Department’s airway management guidance: does bougie use improve first pass success rate?

Chloe Yeap¹, Jessica McQuillian¹, Christopher Edmunds^1,2, Tony Sim^1,2, Charles Moore^1,2 and Reilly Tom^1,2

¹Aberdeen Royal Infirmary, Aberdeen, UK

²Intensive Care Unit, Aberdeen Royal Infirmary, Aberdeen, UK

Abstract

Emergency airway management is a shifting paradigm, traditionally the remit of Anesthetists. In recent years, emergency physicians are increasingly taking the lead when managing airways within their department. The Emergency Department setting has been associated with a higher incidence of complication and failure by NAP4. ¹ It has also been demonstrated in the literature that increased laryngoscopy attempts are linked to increasing risk of complications. ² This means that in an Emergency setting the importance of initial success with any airway intervention becomes intrinsically linked to patient outcome and mortality, particularly as the majority of these patients are admitted to the Intensive Care Unit.³

Recent literature has suggested the addition of an airway adjunct, increases the rate of successful airway interventions at first attempt. ⁴ We retrospectively compare the rate of success and complications with or without a bougie for emergency airway interventions.

We retrospectively reviewed all patients needing Emergency Department intubation between July 2015 and July 2017 to identify the first pass intubation success, adjunct use and complication rate. Completed using the local airway registry, data was statically analysed in Microsoft Excel.

197 cases of rapid sequence induction were identified between 1 July 2015 and 31 July 2017. Of these, 51 (26%) were excluded due to incomplete data. Of the remaining 146 cases, there were 106 (72.6%) cases where a bougie was used for intubation and 40 (27.4%) cases where no bougie was used.

The two groups demographics showed no significant difference for sex, age, indication for intervention and predictor of difficulty. Overall first pass intubation success was 80% with 19% at the second attempt and no cases requiring three attempts.

There was no difference in the first pass success rate with the use of a bougie compared without a bougie (bougie: 79.2%, no bougie: 82.5%, p = 0.84). There was no significant difference in first pass success rate between specialties (EM: 81.1%, Anaesthetics: 86.7%, ITU: 66.7%, p = 0.89). 88.9% of EM doctors who used a bougie for intubation succeeded on the first attempt compared with 100% without a bougie. Similarly, 77.3% of Anaesthetic doctors who used a bougie succeeded on the first attempt compared with 91.7% who did not use a bougie.

Common complications were desaturation and hypotension, 19 cases (13%) of each were identified. Most cases of desaturation occurred in the bougie group (bougie: 15.1%, no bougie: 7.5%, p = 0.15). The incidence of hypotension was similar for both groups. There was a higher incidence of oesophageal intubation in the no bougie group (bougie: 3.8%, no bougie: 7.5%, p = 0.47), a sub analysis of EM doctors showed a higher incidence of oesophageal intubation when a bougie was not used.

In conclusion although the benefit of using a bougie for every intubation unclear, the data suggests that EM doctors may benefit from use of bougie, but risk increased desaturation rate as a result. The choice of whether to use an adjunct or not is specialty situation dependant. More prospective work is needed to evaluate the impact of adjunct use in emergency airway interventions.

A Quality improvement fellowship: Developing an in-situ simulation programme for collaborative neonatal resuscitation in a District General Hospital

Francesca Riccio, Caroline McCrea, Matthew Williams, Lisa Toft, Emma Williams and Laura Polkinghorne

Queen Alexandra Hospital, Portsmouth, UK

Abstract

Paediatric acuity in a Queen Alexandra Hospital is high, particularly in relation to the stabilisation and resuscitation of infants. The relative infrequency in the management and resuscitation of neonates leads to unfamiliarity amongst the multidisciplinary teams involved. This in-situ simulation project focused on human factors, patient safety and improving the interface between staff.

A pre-simulation questionnaire was sent to all the critical care, paediatric and anaesthetic teams to assess confidence in stabilising a neonate of <4kgs. Based on the results of the questionnaire, scenarios were developed on actual learning events, including a neonate with bronchiolitis, a septic neonate with an unanticipated difficult airway, and non-accidental injury.

The results of the questionaire and simulation scenarios were as follows:

Questionnaire: 47 responded (Paediatrics 16, Critical care 7, and Anaesthetics 23). Average confidence was low across all specialties, but lowest in anaesthetics.

Simulation scenarios: Four different scenarios were run throughout the year. The scenarios and de-briefs were effective in identifying key learning outcomes, including; using an Ayres-T-piece, setting up a resuscitaire, working in silos, and availability of difficulty airway equipment. Subsequently, the following were implemented: paediatric mini skills sessions, an anaesthetic checklist, change to paediatric front of neck access packs and a paediatric study day. In view of positive feedback, the in-situ simulations and mini-skills sessions will continue for the foreseeable future. Health Education England has awarded £10,000 to develop an in-situ app transferable to other specialties.

Lessons learnt included the realisation that introducing in-situ simulation into a department takes a minimum of six months, primarily due to staff uncertainties about what in-situ training involves. It is important to highlight that simulation is testing the system not the individual. Effective debrief post in-situ simulation can have a powerful impact on individuals and enhance take home messages.

Simulation within the clinical environment improves reliability and safety in high-risk areas by providing great realism and a safe setting to explore latent threats and system errors [1]. Multidisciplinary training and debriefing facilitates sharing of information across specialty ‘silos’ [2]. Effective and concise debriefing allows critical components to be uncovered in a non-judgemental and confidential way [3]. In-situ simulation delivered within the working environment, as opposed to centre-based simulation, is therefore the most appropriate for this kind of programme.

A novel approach to control the correction of sodium imbalance in patients requiring haemodiafiltration

Benjamin Whatley, Andrew Weir and Richard Walker

Torbay and South Devon NHS Trust, Torquay, UK

Abstract

Patients presenting with severe hyper- or hyponatraemia require slow, controlled correction of their sodium imbalance to avoid a brain injury resulting from cerebral oedema, or central pontine myelinolysis. Independent to the treatment of their sodium imbalance some pateints also require haemodialysis, however, used with standard protocals this will rapidly correct a patient’s plasma sodium concentration to that of the dialysate fluid, commonly 140mEq/L. Hence in these patients a specialised approach is required.

There are several case reports in the literature whereby the problem has been successfully overcome by incrementally changing the sodium concentration of the dialysate fluid, through the addition of concentration sodium solutions, or sterile water. However, this technique lacks the option for real time adjustment, as modification can only be made in time with dialysate bag changes, and could expose a patient to additional risks from errors in dialysate adjustment, changes to other electrolyte concentrations, and loss of sterility.

Another method which we have used and which has not been described in detail in the literature is to provide the patient with haemodialysis using standard dialysate, and concurrently provide an infusion of hyper or hyponatraemic solutions to the patient. The rate of sodium (when correcting hypernatraemia) or free water (correcting hyponatraemia) removed from the patient by the renal replacement therapy can be calculated, and replaced centrally using infusions of 2.7% NaCl or 5% Dextrose respectively. The rate of these infusions is then adjusted to correct the sodium imbalance at a safe rate. By using a specifically designed calculator to set the rate of the replacement infusions, changes to haemodialysis settings or the patient’s sodium concentration can be quickly accounted for with a reduced risk of error.

We describe an alternative method for dialysis with controlled sodium correction. Both our and previous approaches have advantages and unique risks to them, and we make no recommendation as to which is more appropriate for use in any given setting. If either is used, education of nursing and junior medical staff is vital to their safe use.

Patients are on an improving trajectory of illness at time of admission to ICU, and patients with a higher likelihood of sepsis show a greater improvement

Jennifer Hunter, Edward Palmer and Steve Harris

University College London Hospital, London, UK

Abstract

Background: Trajectory of acute illness, measured as trajectory of severity score, has been suggested as an independent marker of severity of illness. As underlying pathology is likely to affect trajectory of illness, we analysed severity score trajectories in patients deemed likely and unlikely to have a diagnosis of sepsis prior to intensive care unit (ICU) admission. The new Sepsis-3 definition is based on a change in SOFA score, so understanding the trajectory of sepsis at the time of ICU admission is important for diagnosing sepsis in this context.

Methods: Data collected as part of a large multicenter observational study of ward patients accepted to ICU ((SPOT) light study) were analysed, including physiological measurements taken at the time of ward assessment and in the first 24 hours of ICU admission. Trajectory was defined as the change per hour in severity score between these two time points, calculated based on the ICNARC APS scoring system for patients with a full set of physiological measurements available at both assessments. Physicians rated the likelihood of sepsis at the time of ward assessment from 1 to 4.

Results: Data for analysis were collected for 5,429 patients (494 study months). The subset with complete data for ward and ICU assessments (778 patients) were analysed. Of these patients a diagnosis of sepsis was considered likely 598 (76.9%), and unlikely in 180 (21.2%).

Median trajectory calculated based on ICNARC APS scoring was −1.00 (IQR −4.00–0.00) in the group where sepsis was judged likely, and −0.33 (IQR −3.00–0.21) in the group where sepsis was judged unlikely. The difference is significant (Kruskal-Wallis test, H = 7.87, 1 d.f., p = 0.005). A negative trajectory represents improvement, and a positive trajectory deterioration. There was no significant difference in the severity score at ward assessment between the two groups (Kruskal-Wallis test, H = 0.76, 1 d.f., p = 0.38).

Conclusions: Access to a full set of physiological measurements for patients prior to ICU admission is rare, and here shows that patients are more commonly on an improving trajectory on admission to ICU, regardless of sepsis likelihood. There is a significant difference in the trajectories for the two groups, with patients with a likely diagnosis of sepsis showing a greater median improvement, of 1 severity score point per hour. Lead time bias may contribute to the difference observed, and recent awareness campaigns may have improved the speed of detection and treatment of sepsis compared with other conditions.

Table 1.

Characteristics of patients with two full sets of physiological measurements based on sepsis likelihood.

	778 Patients
	Sepsis likely (598)	Sepsis unlikely (180)
Age	69.0 (58.0 – 77.0)	69.0 (57.0 – 79.0)
Ward APS ICNARC severity score	24.0 (20.0 – 28.8)	23.0 (19.0 – 28.0)
Trajectory of severity score (change per hour)	-1.00 (-4.00 – 0.00)	-0.33 (-3.00 – 0.21)
Time from ward assessment to ICU admission (hours)	3.0 (1.0 – 9.0)	3.0 (1.0 – 8.0)
28 day mortality	30.3%	30.6%

The Implementation of Pet Therapy in Critical Care: A Service Improvement Project

Rose North

Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK

Abstract

Advances in critical care medicine have led to more people surviving than ever before, however many patients report their experiences as traumatic including pain, anxiety, loneliness and depression (1). Critical care stays can be linked to worsened physical and mental health outcomes in the recovery period and can lead to post-intensive care syndrome (PICS) for patients. Medical and nursing staff are now focusing more on the patient as an individual and there is growing interest in non-pharmacological interventions to create a more person-centred and humanized environment within critical care.

Intensive care can make patients feel especially lonely and isolated from their families and normal lives (4). Many people take great pleasure from spending time with animals when at home; pets bring comfort and can have a calming effect on their owners. Research suggests that pets (especially dogs) can be a good distraction from stressful situations and improve overall mental health. Although the studies are in non-critical care areas, the results show benefits in reduction of pain and anxiety which could be useful in my clinical area. I believe that patients who are less anxious, depressed or stressed are more likely to be motivated with their rehabilitation during critical illness and this could therefore reduce their length of stay in hospital. Pet therapy is a low cost and effective way of reducing anxiety, improving mood and decreasing pain.

I chose pet therapy as a topic as I believe in the benefits of pet therapy on patients and I wanted to see if this could be applied to critical care. I have seen patients struggle with low mood and sleeplessness and hope that pet therapy could improve mood by providing patients respite from their illness.

Some hospitals have incorporated pet therapy into practice to reduce suffering and promote recovery, however critical care poses numerous issues due to the severity of illnesses and dependency of patients and a focus on clinical need rather than holistic need. In my area of work, pet therapy is new and is currently only championed by myself therefore there are currently no set guidelines to follow. The uptake of pet therapy is minimal and limited to the availability of myself and the animal handler.

The aim of the first half of this service improvement project is to establish a baseline for the pet therapy service our units currently offer. I have collected data which shows the number of suitable patients compared to the total number which I established using eligibility criteria written for this project. Ideally, I would like 60% of eligible critical care patients who want to receive pet therapy to have it by December 2018. This aim is realistic as I am the only nurse championing pet therapy and the service is offered on a voluntary basis and therefore not feasibleto offer 100% of the time.

Outcomes of older Colorectal patients admitted to a Scottish Intensive Care Unit

Philip Leadbeater^1,2, Tom Evans^1,2 and Murray Blackstock²

¹South East Scotland School of Anaesthesia, Edinburgh, UK

²Western General Hospital, Edinburgh, UK

Abstract

Data collected in the England and Wales National Emergency Laparotomy Audit (NELA) identifies a stepwise increase in mortality (30 day and 90 day) with each decade of life.

In the absence of a formalized local process of identifying older patients undergoing emergency laparotomy and to better understand the impact on our Intensive Care Unit (ICU) we undertook a retrospective audit of all Colorectal surgery patients aged over 70 admitted to our ICU during 2015 and 2016. Patients were identified by specialty coding within the locally collated admissions database with further information obtained from the electronic patient record. Data collected included patient demographics, reason for admission, length of stay, level of care days and mortality status at 30 days, 6 months and 12 months from date of ICU admission.

There were 100 Colorectal surgery admissions during this two year period, representing 90 individual patients, 46 Female (51%) and 44 Male (49%) with a mean age of 78 years (70–94 years).

Table 1.

Mortality at specified time points for all Colorectal surgery patients and the subset of admissions after emergency laparotomy, admitted during 2015 and 2016.

	Mortality in all Colorectal admissions (%, n = 90)	Mortality in direct emergency laparotomy admissions (%, n = 55)
30 days after ICU admission	21%	25%
6 months after ICU admission	31%	35%
12 months after ICU admission	38%	42%

16 admissions were directly after elective surgery. The remaining 84 admissions were comprised of 55 direct admissions after emergency laparotomy; 15 (10 patients) for respiratory failure and 6 (5 patients) for cardiac or renal impairment after recent major Colorectal surgery. 8 admissions were for patients with a primary surgical issue with non-operative management or after minor-moderate Colorectal surgery. Mortality data are summarized in table 1.

These 100 admissions accounted for a total of 568 critical care days, including 297 days of level 3 and 193 days of level 3 care. Patients admitted directly after emergency laparotomy had a median ICU length of stay of 4 days (inter quartile range 2–7 days; total ICU days including same patient re-admissions) accounting for 177 level 3 and 117 level 2 days. Renal replacement therapy was used infrequently (12/90 patients, 8/55 emergency laparotomy admissions).

We hope to use these preliminary data to drive local cross-specialty engagement in developing pathways for early identification of older patients undergoing emergency colorectal surgery.

We have only looked at those patients admitted to the ICU and clearly further work is needed to understand how the local management of older patients undergoing emergency Colorectal surgery measures against the well defined standards proposed in both NCEPOD and NELA reports.

A multidisciplinary approach to post intensive care tracheostomy weaning and the impact of a dedicated team on decannulation rates and outcome in a regional UK major trauma centre

Rachael Moses, Jane Pulsford, Sarah Bunting, Louise Stevens, Helen Al-nufoury, Andrew Fishburn, Claire Slinger, Aashish Vyas and Craig Spencer

Lancashire Teaching Hospital NHS Trust, Preston, UK

Abstract

Background: Percutaneous tracheostomies are commonly performed in intensive care to expedite weaning from mechanical ventilation especially following major trauma, acquired brain injury or severe respiratory failure. Often patients are discharged from the intensive care unit (ICU) to a ward environment with no specialist follow up

	Pre-MDT	Post MDT
Total ICU trachy patients	49	51
Discharged to Neuro/Respiratory with trachy in situ	21	12
Discharged to ward decannulated	7	13
Discharged to ENT ward	15	18
Died (mostly on ICU)	6	1

Aim: To evaluate the effectiveness of a multi-disciplinary tracheostomy team (MDT) at reducing the total length of hospital stay and improving decannulation rates in tracheostomy patients once discharged from ICU

Method: The team consisted of a Consultant, Physiotherapist, Speech and Language Therapist, a Head & Neck Specialist Nurse and a Critical Care Outreach Nurse. The team met weekly on the neurosurgery and respiratory wards and may prescribe treatments or therapies, offer advice to ward staff or carry out interventions. Audit data was gathered for 6 months preceding the establishment of the team and during the 6 month pilot period.

Result:

Based on around 62 patients being discharged to neurosurgical and respiratory wards per year, the permanent introduction of a Trust Tracheostomy MDT has the potential to reduce the time patients spend with a temporary tracheostomy in situ by 50% and in patient bed use by 911 days per year.

Conclusion

A Tracheostomy MDT is an essential service to ensure timely decannulation as well as reducing hospital length of stay and improving overall outcome.

ARDS and Thrombocytopenia: A Catastrophic combination?

Hannah Blagnys^1,2, Sarah Beavis³, Richard Dobson³ and Anand Padmakumar³

¹ST6 Intensive Care Medicine, Chesterfield Royal Hospital, Chesterfield, UK

²East Midlands School of Anaesthesia, Nottingham, UK

³Consultant in Anaesthesia and Critical Care Medicine, Chesterfield Royal Hospital, Chesterfield, UK

Abstract

Mr W, a 23 year old gentleman, presented to hospital with pallor and breathlessness which had been progressing over the past weeks. He had a diagnosis of autistic spectrum disorder but was otherwise previously well. His family however had noticed a decline in his mental health, weight gain and worsening pallor over the preceding months.

His admission bloods showed a haemoglobin of 61 g/L, white cell count of 12.1x10⁹and platelets of 14x10.⁹ His urea and creatinine were also raised.

An initial diagnosis of immune thrombocytopenic purpura (ITP) was made and he received two units of red blood cells along with oral prednisolone. Over the next three days he deteriorated becoming delirious and hypoxic ultimately requiring tracheal intubation and ventilation. He was difficult to ventilate after intubation requiring high airway pressures and FiO2 despite diuresis. His X-ray showed diffuse pulmonary infiltrates suggestive of ARDS.

Mr W’s ongoing management included intravenous methylprednisolone, intravenous immunoglobulins and lung protective ventilation. A full body CT was requested to rule out malignancy. His CT head, confirmed later by MRI, showed multi-territory mixed venous and arterial acute infarcts suspicious of embolic disease.

Further haematological tests included ADAMST-13 which was negative ruling out thrombotic thrombocytopenic purpura (TTP). Lupus anticoagulant, anticardiolipin IgG and IgM were all positive giving the likely diagnosis of catastrophic antiphospholipid syndrome (CAPS). Supportive treatment was continued including ventilation, IV heparin, corticosteroids and a course of immunoglobulins. Mr W’s thrombocytopenia and oxygenation slowly improved without plasma exchange.

Antiphospholipid syndrome (APS) is a systemic autoimmune disease characterised by recurrent thromboembolic events in the presence of antiphospholipid antibodies.¹ When this syndrome is accelerated resulting in multi-organ failure it can be defined as catastrophic antiphospholipid syndrome (CAPS).² Preliminary criteria for diagnosis include;

1. Involvement of three or more organs, systems and/or tissues

2. Development of manifestations simultaneously or in less than a week

3. Confirmation by histopathology of small vessel occlusion in at least one organ or tissue

4. Laboratory confirmation of the presence of antiphospholipid antibodies²

Catastrophic APS is rare, occurring in less than 1% of those diagnosed with APS.³ In those patients with CAPS, pulmonary involvement arises in approximately 68% of cases, and ARDS account for nearly one third of those.⁴ Pathogenesis of this is poorly understood, however antiphospholipid antibodies have been found in the broncho-alveolar lavage fluid of patients with ARDS which may support the hypothesis that antiphospholipid antibody may promote the process of ARDS.⁵

A precipitant for catastrophic APS is not always found, however it is most likely to be infection.⁴ Surgical or invasive procedures such as ERCP can also precipitate a crisis and thus clinicians should be aware of this in their day to day practice. Other causes include drug treatment, obstetric complications, SLE flares and withdrawal of anticoagulation.⁴ The most commonly used treatments are anticoagulation, corticosteroids, plasma exchange and intravenous gammaglobulins.²

Catastrophic APS is a rare condition and has a mortality rate of 50% despite treatment.² It may present with ARDS, thrombosis or thrombocytopenia requiring physicians to have a high index of suspicion when treating such clinical pictures.

The impact of a dedicated multidisciplinary social media team at a national critical care conference

Rachael Moses¹, Aoife Abbey², Nitin Arora³, Robert Cronin⁴, Steve Mathieu⁵, Ganesh Suntharalingam⁶, Jamie Strachan⁷, Tanjinder Sanghera⁸ and Jonathan Downham⁹

¹Lancashire Teaching Hospital NHS Trust, Preston, UK

²Queen Elizabeth Hospital, Birmingham, UK

³Heartlands Hospital, Birmingham, UK

⁴Gloucestershire Hospitals NHS Foundation Trust, Gloucestershire, UK

⁵Portsmouth Hospitals NHS Trust, Portsmouth, UK

⁶London North West University Hospitals, London, UK

⁷Oxford University Hospitals NHS Foundation Trust, Oxford, UK

⁸Barts Health Trust, London, UK

⁹South Warwickshire NHS Foundation Trust, Warwick, UK

Abstract

Introduction: Social media is a powerful product of this generation. In the last five years alone the number of ‘tweets’ sent per day has increased by a factor of 100,000. 500 million ‘tweets’ are now sent per day. Using a designated twitter ‘hashtag’ allows a conference to promote interaction before during and after the event

Aim: To assess the impact and reach of the annual Intensive Care Society (ICS) conference by instituting a dedicated social media (SoMe) team

Method: A dedicated SoMe team, modelled with aim of representing the critical care multidisciplinary team, was set-up approximately one year in advance. The approach included early adoption of a single official hashtag which was registered with ‘Symplur’, an online platform that collates and provides analytics for hashtags. We integrated ‘twitter moderators’ into every session, live-streamed delegates’ tweets on screens, created real-time infographics on the talks and also used our hashtag to encourage attendance at networking/social events. We had facilities for creating live stream and recorded interviews with speakers. We also created daily ‘summing up’ videos which were uploaded across facebook, youtube, periscope and twitter.

Results: The impact of a dedicated SoMe team:

	ICS SOA 2017 (over 3 days)	ICS SOA 2016 (over 3 days)	Change from 2016–2017 (%)
Total Impressions	25.19 million	9.68 million	+160%
Total Tweets	15226	7482	+103%
Total Particpants	2376*	1107	+114%
Avg Tweets/participant	6	4	+50%

^*The number of paying delegates to the conference was 1061.

Conclusions: A multi-professional SoMe team significantly increased the impact and reach of a national intensive care conference in the UK. This approach could be adopted for other international conferences to share education, learning and involvement in the future.

Tracheostomies: Are The Unweanable Really Weanable?

Rachael Moses¹, Aashish Vyas¹, Jessica Blakemore¹, Helen Al-nufoury¹, Claire Slinger¹ and Natasha Alexander²

¹Lancashire Teaching Hospital NHS Trust, Preston, UK

²Brighton and Sussex University Hospitals, Brighton, UK

Abstract

Introduction: There has been a steady rise in the use of percutaneous tracheostomies in the intensive care unit (ICU) since they were first introduced over 30 years ago. Many patients are discharged from ICU with tracheostomies, most with little specialist follow up meaning huge financial implications and poor patient outcome. As part of a pilot scheme across the region the Lancashire and South Cumbria long term ventilation service (LSCLTVS) accepted patients with permanent tracheostomies and high cost care placements to see if there was any potential to wean. This single case study details the pathway for one

Methods: A 54-year-old male who had suffered an intracerebral hematoma 3 years previously required craniotomy and tracheostomy due to high secretion load and inability to protect his airway. He had suffered with copious secretions, recurrent infections and poor swallow, meaning weaning had been unsuccessful.

He was referred to the LSCLTVS to assess potential for weaning. Initial nasendoscopy revealed large volumes of pooled saliva based secretions, adduction of the vocal cords, some degree of laryngospasm, redness to the endobronchial tree and no significant swallow.

Combined Physio and Speech and Language Therapy commenced for upper airway desensitisation alongside artificial airflow voicing using 10L oxygen delivered via subglottic tracheostomy port. Rescue breathing was used to re-educate the upper airway and later cuff deflation and finger occlusion for voice elicitation. Further therapy using a passy muir speaking valve (PMV) and rescue breathing allowed voice production and airway clearance.

Results: Once continuous cuff deflation was tolerated along with PMV for up to 12 hours daily we performed a cap trial for 45-minutes to assess readiness for decannulation, which he tolerated exceptionally well with vocalisation, no upper airway dysfunction, minimal suctions and ability to cough and clear into his mouth. This demonstrated he was ready for inpatient assessment and decannulation

Conclusion: Following a 5 day inpatient stay for repeat nasendoscopy and assessment the patient was successfully decannulated after 2 months of therapy, having previously had a tracheostomy for nearly 3 years. 3 months later he was able to eat and drink after swallow rehabilitation. The cost saving for decannulation alone is over £92,000 per annum. This case is an example of the importance of continual re-evaluation of long-term tracheostomy patients and the importance of laryngeal weaning to maximise patient outcome and quality of life.

Quality Improvement Project: Delay in ICU Admission Clerking During Busy Periods and Associated Mortality

Edward Watson and Marcela Vizcaychipi

Chelsea and Westminster Hospital NHS Foundation Trust, London, UK

Abstract

Introduction: Critical care units receive unpredictable emergency admissions throughout the day and a concentration of elective surgical patients in the evening. However, due to ward rounds, handovers and reduced night staff, patients admitted after 6pm may experience a significant delay in being seen, which may be associated with adverse events.

Aim: To establish the effect of ICU busyness on patient attention, and the association between early attention and mortality.

Methods: Inclusion criteria: All patients on the ICU admissions database (Accubase, 17/04/1993–13/08/2018, n = 11,001). Exclusion criteria: Incomplete data set (admission time/date, clerking time, discharge date, age, outcome, APACHE II), age <18 (n = 289), still on ICU (n = 6), admission time = '00:00’ (n = 1,770), clerking interval = '0 mins’ (n = 5,201). Statistical analysis: The interval between arrival on ICU and completion of clerking was used as a surrogate marker of junior doctor busyness. The mean interval & arrival times of n = 100 sub-cohorts (clustered chronologically by admission time) were presented graphically for ease of interpretation. Similarly, clerking delay & mortality were plotted against each other, a correlation was tested for using Spearman’s rank. A Mann Whitney U test was used to test clerking delay between busy (8–12am + 6pm-12am) & non-busy times.

Results: N = 4,300. Clerking delay was higher in busy periods (3.5 hours (3.1–3.9) vs 2.6 (2.4–3), MWUT Z = −3.7, p = 0.0003). There was a highly significant correlation between delay & mortality (Spearman's R = 0.74, p < 0.0005), with an associated mortality rise of 1.5%/hr.

Conclusions: Medical staffing levels and daily activities have an impact on the speed at which new admissions are seen. It is reasonable to assume that this may also apply to other timely duties. Longer time to be clerked was associated with a higher mortality, but this is difficult to interpret due to significant confounders, such as sick patients require longer clerking times, morning admissions were more likely to be medical patients who deteriorated overnight, and late surgical admissions were more likely to have had longer and more invasive surgery.

Audit: Adjusting Shift Patterns to Match the Time Profile of ICU Admissions

Edward Watson and Marcela Vizcaychipi

Chelsea and Westminster Hospital NHS Foundation Trust, London, UK

Abstract

Introduction: The majority of elective surgery patients leave recovery at 18:30. With both SHOs working 08:00–18:00, they were only present for the arrival of 35% of all new admissions and 41% of planned surgical admissions. Furthermore, with only skeleton staff to juggle new referrals, the evening ward round and the handover, patients arriving from recovery after 6pm frequently faced long delays before being formally admitted.

Aim: To assess the impact of adjusting the SHO hours to match the admission load.

Methods: One SHO was kept on the 08:00–18:00 shift, the other was changed to a 12:00–22:00 shift for three days a week. Inclusion criteria: All patients admitted between 01/02/17 (SHOs introduced). Exclusion criteria: Incomplete data, those patients admitted on a day without an evening SHO on the rota. Outcome measures: Percentage of admissions within the extended shift span, reduction in clerking delay, reduction in Recovery LOS, reduction in Resus LOS. Statistical analysis: Clerking interval, Resus LOS and Recovery LOS were compared for patients admitted to ICU from 18:00 to 22:00 between the period 05/08/15 (when SHOs were introduced to the unit) – 01/08/17 (when the SHO shift change was implemented) and the period 02/08/17–06/02/18.

Results: Admission coverage: By shifting the start time back by four hours to 12:00–22:00, the SHOs were present for 41% of all admissions and 86% of planned surgical admissions.

Clerking interval: Moving one SHO to a later shift time resulted in a small drop in clerking-interval time from 2.3 hours (IQR 1.3–4.1) to 1.7 hours (IQR 0.6–4.4), though this did not reach statistical significance.

Recovery: There was minimal change between the baseline time of 3.0 hours (IQR: 1.7–4.8) spent in recovery and after the shift change: 2.8 hours (IQR: 1.9–4.8)

Resus: LOS was also very similar before (4.2 hours, IQR: 3.8–8.4) and after (4.9 hours, IQR: 3.7–6.9) the shift change implementation.

Conclusions: The significant increase in number of patients admitted whilst an extra SHO was on duty during the evening, and the drop in clerking interval by approximately half an hour suggested that the intervention was a successful adaptation to our admission profile. However, the negligible change in Resus and Recovery LOS implies that other factors external to ICU staffing, such as bed blockade, are the limiting factors in timely ICU admission.

Audit: Prevalence of Out of Hours ICU Discharges and Associated Mortality & Morbidity

Edward Watson and Marcela Vizcaychipi

Chelsea and Westminster Hospital NHS Foundation Trust, London, UK

Abstract

Introduction: The Intensive Care Society’s Core Standards for Intensive Care Units stipulates that discharge from Critical Care should occur between 07:00 and 21:59.¹ This is due to both the negative patient experience and the association with increased mortality.² Since 2004, priority for allocation of ward beds goes to A+ admissions. Although multifactorial, this has led to an increase ICU bed block. The percentage of delayed ICU discharges due to a shortage of ward beds at C&W rose from 1.3% to 17.1% with the implementation of the 4-hour target. In many units it is now commonplace for discharges only to occur when a new admission requires the bed, which may occur overnight.

Aim: To review the incidence of patients discharged out of hours (OOH) and calculate their relative mortality.

Methods: A retrospective cohort study investigating the mortality and hospital LOS of ICU patients who were discharged OOH. Inclusion criteria: C&W ICU whole database (Accubase) 30/04/1993–08/01/2018 (n = 10,652), included: n = 4,698 Exclusion criteria: Still an inpatient (n = 130), died on ICU (n = 1,663), palliated on ICU (n = 35), discharged to a more specialist critical care environment (n = 251), discharged at ‘00:00’ (n = 1,759), incomplete data (n = 2,116). Outcome measures: Percentage discharged OOH; Mortality (at hospital discharge). Statistical analysis: Chi squared test will be used to test for an increase in mortality, if significant, a risk ratio will be calculated.

Results: Incidence: Overall 6.4% (301) of patients were discharged OOH from 1993–2017. Mortality: The mortality patients discharged out of hours was higher than that of those discharged in hours (OOH: 9.0%, IH: 5.2%, X²= 7.74, p = 0.005), and a relative risk of 1.71 (95% CI 1.18–2.49, z = 2.81, p = 0.005). They had a longer hospital LOS: OOH median 8 days (IQR 3–23), IH median 6 days (IQR 1–15). Of the 27 OOH discharges who did not survive to hospital discharge: 12 required urgent review overnight, one of which died on the night of discharge. Documented reasons for review included: arrythmia (AF, VT), pulmonary oedema, hypotension, sepsis, reduced conscious level, GI bleeding, hypoglycaemia and mucus plugging.

Conclusions: OOH discharges are few, but are associated with an increased mortality. It also confers risk to those patients who cannot be immediately admitted: both by delay in admission time, but also by diverting the attention of skeleton staff toward discharge planning.

Quality Improvement Project: ICU Bed Blockage from Non-critically Ill Tracheostomy Patients

Edward Watson and Marcela Vizcaychipi

Chelsea and Westminster Hospital NHS Foundation Trust, London, UK

Abstract

Introduction: Tracheostomies are a common critical care intervention for facilitation of weaning from mechanical ventilation, airway protection and/or secretion management. After weaning from organ support they may be suitable for transfer to a lower acuity bed on a specialty tracheostomy ward for ongoing care. Due to the limited number of tracheostomy beds and the high hospital LOS, suitable patients may remain on ICU for prolonged periods. In the context of maximal ICU capacity, this bed blockage will contribute to delayed admissions and reduced standard of care for high acuity patients.

Aim: To review the number of ICU bed days blocked by patients waiting for step down to tracheostomy beds, as well as an estimate of the costs of inappropriate bed expenses.

Methods: We retrospectively reviewed C&W ICU tracheostomy patients across a year to ascertain the number of potential step-down days contributing to bed blockade. Inclusion criteria: Patients admitted to the C&W ICU with tracheostomies between 20/07/16–19/07/17 (n = 31). Exclusion criteria: Burns ICU patients (n = 5), who are not suitable for transfer out of a burns ward. Outcome measures: Delayed transfer of care (DTOC) days and total cost of higher acuity bed space. Statistical analysis: Delayed transfer of care (DTOC) days were taken to be the number of days between a patient being classified as ready for transfer to a tracheostomy care ward by the duty ICU consultant and the day of transfer. Misallocation of funds was calculated as the difference in GBP between a level 2 ICU bed (estimated according to Together for Health – A delivery plan for the critically ill, 20131) and a ward level bed: £444pn (£857 – £413).

Results:

Admissions: 25 patients with tracheostomies were cared for on ICU across the year. Two were admitted with long-term tracheostomies, one was admitted twice. 17 were medical, 9 were surgical. The total number of tracheostomy bed days was 796.

Timings:

	pre-trache (days)	trache (days)	post-trache (days)
Medical (not including long-term  tracheostomies)	8 (6–10)	21 (15–36)	1 (0.25–11)
Surgical	9 (4–10)	15 (7–26)	2 (0.25–5)

Outcomes: 54% (14) were decannulated on ICU, 27% (7) died on ICU with a tracheostomy in situ, 12% (3) were discharged to the ward and 8% (2) to the community. DTOC and associated cost differential: 5 patients incurred a total of 106 DTOC days, equating to approximately £47,064 of funds spent on higher acuity bed days. The main limitation to step-down was number of trache-trained nurses on the specialist ward.

Conclusions: Increasing the number of trache-trained nurses is a low-cost strategy that would enable rapid step-down of trache patients and reduce ICU bed blockage by >100 days/year.

Creating a multidisciplinary airway teaching program on ITU

Jims J^1,2, Benjamin Cullinger^1,3, Gosia Bal¹, Jess Payne¹, Charlie Marriott^1,3, James Watson⁴, Ankur Sethi¹ and Peter Walker¹

¹Barking, Havering Redbridge NHS Trust, Romford, UK

²Central London School of Anaesthesia, London, UK

³Bart’s and the London School of Anaesthesia, London, UK

⁴Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

Abstract

Background: Airway management forms a vital part of the management of critically ill patients. However intubation of critically ill patients is not without its risks. Due to the various patient factors and environment factors present in the critical care unit, failure of first intubation attempt may be as high as 10–20%(1), with potentially catastrophic consequences. The case of Elaine Bromley highlighted that optimal management of airway difficult requires not only technical competency, but also good human factor management(2). Difficult Airway Society (DAS) has created various guidelines to aid safe management of unanticipated airway difficulty, including the joint 2017 guidelines for intubation of critically ill patients(3).

The aim of our teaching program is to promote the dissemination of knowledge on airway difficulty management amongst the junior doctors and junior nurses, who may not have experience managing an airway emergency, but may be expected to assist in such situations. We also planned to highlight some of the human factor considerations through in situ simulation.

Method: Pre-intervention questionnaires focused on the DAS 2017 critically ill patient guidelines were distributed amongst the doctors (both airway-trained and non-airway-trained) and nurses on our critical care unit. The responses were used to highlight areas for intervention.

Interventions included DAS guideline teachings for the non-airway trained critical care doctors; DAS guideline teachings for critical care nurses; in-situ simulation sessions on the unit with focused debrief on aspects including team communication and situational awareness. There is also ongoing work to have DAS guidelines and our custom ‘kit-dump’ diagram laminated and made readily available on the unit.

Results: Pre-intervention questionnaires highlighted some areas requiring education. This was addressed through the DAS guideline teaching sessions. Staff demonstrated significantly better knowledge of the DAS guidelines (see Fig 1); as well as higher subjective confidence score after the teaching sessions. The simulation sessions and the structured debrief led to in depth discussion about the preparation and management of failed intubation scenarios, as well as discussion about human factors considerations such as raising concerns, close loop communications and assertiveness. Both teaching modality were well received from the staff.

Discussion: It’s known that critically ill patients are significantly more likely to suffer from intubation related complications, and these may be attributable to environment factors as much as patient factors. Indeed, inadequate provision of skilled staff and assistants was one of the aspects highlighted in the NAP 4 audit as well as DAS publications (3, 4). We have demonstrated that through this multimodal teaching program, it is possible to raise the knowledge level amongst the staff and practice the non-technical skills in a controlled environment. The introduction of this project has also stimulated interest and receptivity to the use of checklist and other aide-memoires amongst the staff members. We hope that a combination of teaching, simulation and making aide-memoires readily available on the unit will help to reduce the risk of intubation related complications.

Breast cancer in the emergency department

Sorcha Hodgson¹ and Emma Jackson²

¹Blackpool Victoria Hospital, Dept. Critical Care, Blackpool, UK

²Lancashire Teaching Hospital NHS Trust, Preston, UK

Abstract

Introduction: Figures by the national cancer intelligence network (NCIN) report that 23% of new cancer diagnoses occur as emergency presentations either via the emergency department or through emergency GP referral. For the majority of cancer’s, it is associated with a worse one-year survival. We report an usual case of severe sepsis leading to a new diagnosis of breast cancer.

Case summary: A 71-year-old lady presented to the ED having been found by her neighbour unresponsive with a GCS of 8, following a 48-hour period of feeling generally unwell. On arrival she was found to be profoundly hypoxic tachypnoeic and haemodynamically unstable. Further assessment noted a large malodorous fungating mass on her right breast the most likely cause being a fungating breast tumour. The patients GCS improved following fluid resuscitation but she remained profoundly hypotensive requiring admission to Critical Care for invasive monitoring, cardiovascular support and further resuscitation. Imaging of the chest was unremarkable. Once haemodynamically stable, she was taken to theatre for an emergency mastectomy, which was packed and left open for a second look. Post operatively the patient had a period of haemodynamic instability most likely due to a septic shower following surgery. Resuscitation continued and by day 5 the patient no longer required cardiovascular support and was suitable for primary extubation. She underwent a staging CT which showed she had no spread of disease and histology showed a triple negative invasive ductal carcinoma.

Discussion: Breast cancer is the most common cancer in females (1) the majority of patients are diagnosed following referral to secondary care under the ‘2 week wait’ (TWW) rule. Screening of eligible women remains the best route for detecting early breast cancer. Despite both initiatives a small proportion (1–4%) of women receives their diagnosis of breast cancer as a result of an emergency presentation. These patients subsequently have the poorest survival rates (2).

Locally advanced breast cancer (LABC) describes malignancy which is at its most advanced stage with no metastasis it accounts for 10–30% of all new breast cancer diagnosis globally (3). LABC encompasses a spectrum of clinical presentations including women presenting with fungating masses as the result of infiltration of the primary tumour into the surrounding skin. Fungating masses can lead to a range of distressing and psychologically upsetting symptoms for the patient and pose a complex management challenge for clinicians due to pain, bleeding, malodour due to exudative discharge and infection (4) which in this isolated case resulted in a life-threatening organ dysfunction requiring intensive care admission.

The incidence of sepsis is increasing globally and it is the most frequent reason for admission to Intensive care unit (ICU) and despite more intensive and invasive management of sepsis- associated organ dysfunction; it remains the most common cause of death in ICU (5). Fundamental to the survival of patient with severe sepsis is the early recognition and treatment. Following the patient’s initial resuscitation and then stabilisation in ICU, surgical source control of the fungating mass was essential.

The impact of a 24/7 PCI service on cardiac arrest admissions to critical care at a district general hospital

Richard Pugh¹, Lawrence Pugh²

¹Ysbyty Glan Clwyd, Rhyl, UK

²Cardiff University, Cardiff, UK

Abstract

Background: Current European Resuscitation Guidelines advocate urgent coronary angiography following out of hospital cardiac arrest (OOHA) of likely cardiac cause (1). Betsi Cadwaladr University Health Board (BCUHB) serves a North Wales population of approximately 690,000 on three acute district general hospital sites. A primary percutaneous coronary intervention (PCI) service was introduced at Glan Clwyd Hospital (GCH) in June 2015, initially operating Monday-to-Friday 08:30 to 17:00 before moving to a 24/7 service April 2017. GCH is now considered one of three cardiac arrest centres for Wales. The aim of this study was to investigate changes in workload for the primary PCI site (GCH) arising from new post-cardiac arrest care pathways and associated outcomes. Method: Retrospective review of the Ward Watcher database of GCH critical care unit, defining OOHA of cardiac aetiology (OOHA-C) on the basis of primary and secondary diagnoses and free text entry.

Table. Admissions to ICU following cardiac arrest 2013–18, GCH

Year Total no. admissions following cardiac arrest No. OOHA No. OOHA-C No. OOHA-C undergoing urgent angiography/PCI (%) ICU bed days Overall OOHA ICU survival 2013–14 25 12 9 2 (22%) 37 25% 2014–15 39 25 24 3 (13%) 79 48% 2015–16 27 16 14 5 (36%) 50 50% 2016–17 30 20 17 7 (41%) 70 55% 2017–18 69 49 39 31 (79%) 202 53%

Results: Numbers of patients admitted to GCH critical care unit following OOHA increased (as expected) over the 5-year period, with increase in bed occupancy of approximately 150 days (see Table). The largest increase was observed in patients with OOHA-C, but an increase was also observed in admissions due to OOHA of likely non-cardiac origin (OOHA-NC). A significant increase in proportion of patients with OOHA-C undergoing coronary angiography, and non-significant increase in overall unit survival (as shown) and hospital survival (17% to 39%) were observed. Urgent coronary intervention was associated with a significantly greater likelihood of survival to unit (67% vs. 38%; p = 0.004) and to hospital discharge (60% vs 25%; p = 0.000) among the OOHA-C population. Discussion: Development of 24/7 PCI service at BCUHB has led to a substantial increase in critical care admissions following OOHA for the PCI site. Although we have not been able to examine sub-groups of OOHA-C patients with ST- or non-ST elevation myocardial infarction, urgent coronary angiography was associated with an overall significant increase in survival to unit and to hospital discharge among OOHA-C patients.

Renal Replacement Therapy in the Intensive Care Unit at Aberdeen Royal Infirmary

Yong Hui Soo and Paul Gamble

Intensive Care Unit, Aberdeen Royal Infirmary, Aberdeen, UK

Abstract

Introduction: About two-thirds of patients with acute kidney injury in the intensive care unit will require renal replacement therapy (RRT).¹ RRT in the intensive care unit is associated with a high mortality rate of 50–60%. The use of systemic anticoagulation heparin has been associated with the risk of haemorrhage. The introduction of regional anticoagulation with citrate has demonstrated longer circuit life and less bleeding complications compared to systemic anticoagulation with heparin, and is recommended by international consensus guidelines.²

Aim: The purpose of this study is to look at the modality of RRT and the mortality associated with RRT in the intensive care unit at Aberdeen Royal Infirmary

Methods: This was a retrospective analysis looking at patients who received RRT in the intensive care unit at Aberdeen Royal Infirmary between August 2017 and January 2018. Patients were identified via ward watcher and were followed up for a 90 day period using ‘Trakcare’.

Results: 56 patients were identified during the study period and followed up for 90 days. The male to female distribution was 31:25. The age of the patients ranged from 17 to 82 years old, with the median age being 65 years old. 36 patients had risk factors for renal impairment (19 patients had diabetes, 21 patients had vascular disease while 23 patients had ischaemic heart disease). The majority of patients who underwent RRT in this intensive care unit had a unit admission diagnosis of septic shock (45%), followed by other (29%), post cardiac arrest (20%), pharmacological overdose (3%) and trauma (3%). Of these patients, 49 patients underwent continuous veno-venous haemodialysis while 4 patients underwent acute intermittent haemodialysis. Three patients who underwent RRT were previously established on chronic intermittent haemodialysis. Of the 56 patients who were identified and followed up for 90 days, 25 patients were alive. Three of these patients progressed to chronic kidney disease stage 3 while one patient went on to commence intermittent haemodialysis. There were no adverse events recorded on DATIX in relation to RRT.

Discussion: Our unit mortality in patients who received RRT was 43% while the 90 day mortality was 55%. This figure is lower than that of a previous study in 2015 where mortality was found to be at 65%. There was no difference in severity score (APACHE II) between the survivals and non-survivals. 8 patients were excluded from this study as they were admitted for less than a day. Circuit life was noticably prolonged in this group of patients, however there is currently limited data regarding this.

Conclusion: Renal replacement therapy using regional anticoagulation with citrate is safe and not associated with adverse outcomes

Hands off! Intensive care staff perceptions on improving patient safety during admissions

Malachi Needham ^1,2

¹John V Farman Intensive Care Unit, Addenbrooke’s Hospital, Cambridge, UK

²Addenbrooke's Hospital, Cambridge, UK

Abstract

Background: Admissions to the intensive care unit (ICU) occur from a variety of areas such as; the ward, emergency department (ED), theatre and inter-hospital. This survey aimed to explore the perceptions of admissions, to identify staff members feelings regarding admissions, and what can be done to improve the way admissions are performed.

Methods: 71 clinical staff members from one general adult ICU within one hospital participated via a structured online questionnaire. All members of the ICU were invited to complete the questionnaire from consultant to healthcare assistant (HCA). Data was collected over a one month period.

Results: 20 participants (28.17%) indicated that admission handovers currently do not work well. 38 participants (53.52%) felt they did not receive all of the information about a patient during an admission. 50 participants (71.43%) reported that they feel able to ask questions or raise concerns during the admission handover. 62 participants (87.32%) would like more structure to admission handovers. 24 participants (33.81%) reported they find admissions stressful. 44 participants (62.86%) felt that there are too many people talking admissions. 47 participants (67.14%) indicated that their priority was getting the patient on the monitor and bedside ventilator. 4 participants (5.6%) mentioned that ‘communication’ was an issue and needed to be improved on. 14 participants (19.7%) reported that ‘information’ was important during an admission. 2 participants (2.8%) mentioned ‘hands off’ approach to receiving admissions.

Discussion: The main theme demonstrated was the desire to improve the way admissions are undertaken. Suggestions were made to adopt a ‘hands off approach’ in a ‘structured’ format. This could then lead to a reduction in communication errors. Not receiving all of the necessary information about the patient during an admission was a concern, suggesting that patient safety is reduced. Conduct and composure of individuals links with the hands off suggestion. Staff members present during an admission having all of the necessary information about the patient could improve patient safety. Teamwork and communication were recurring themes. Involvement of everyone, encouraging individuals to speak up and creating an environment of no hierarchy fear became evident. All members of the ICU team answered the questionnaire, providing a good representation of the ICU.

Conclusion: Improving the way admissions are structured needs to happen in order to improve patient safety. Key issues such as not receiving all of the necessary information and communication difficulties could be resolved through the use of a structured, hands off approach. This guidance should be based on a combination of working together as a more effective team within the ICU. A structured approach will ensure the team will be working towards one common goal. This study addressed the aim because it highlights the need for a structured approach to the admissions process. I have designed this admission flowsheet to resolve the mentioned issues. I have created this admission flowsheet. I am going to attempt to implement this in practice and then conduct further research into whether it has made a difference to patient care.

Implementation of a ‘call and response’ verbal checklist for rapid sequence induction and intubation in the critical care unit improves patient care and safety

Ashton Burden-Selvaraj and Nicholas Bunker

Royal London Hospital, London, UK

Abstract

Background: Rapid sequence induction and intubation (RSI) in critically-ill patients is associated with increased adverse airway incidents in the Intensive Care Unit (ICU). The reported complication rate of emergency intubations is variable, ranging from 4.2% to 39%.¹ Factors that positively correlated with effective airway management in a national audit were communication, organisation and strategy during RSI.² There is low compliance to national recommendations to ensure pre-emptive communication of difficult airway situations and routine use of a checklist.³ The aim of this project is to improve patient care and safety in the ICU by refining non-technical skills and teamwork during RSI with the use of a ‘call and response' verbal checklist.

Method: A quality improvement project comprising of a pre-intervention and post-intervention prospective observational audit was conducted over a 6-month period at an inner London level 1 trauma centre ICU. Practice was audited against the Faculty of Intensive Care Medicine and Intensive Care Society national standards. The main intervention was a ‘call and response’ verbal checklist integrated into an interactive poster at each bed space. Completion of each checklist item was recorded as a binary outcome ‘yes’ or ‘no’. Data on personnel involved in RSI and frequency of adverse incidents were also recorded. Comparison of outcomes were analysed using either Chi-squared test or Fisher’s exact methods.

Results: Data was collected on 46 intubations, 23 pre-interventions and 23 post-interventions. Age of patients ranged from 17 to 77 years with mean ± standard deviation of 53.6 ± 16.4 years. The majority of patients were male (58.7%). A third of intubations were performed out-of-hours with over 60% of them for primary respiratory problem; type 1 or type 2 respiratory failure. There was statistically significant improvement in team communication components: team leader obvious (P = 0.009), team members introduced (P = 0.002), roles allocated (P = 0.004), difficult airway plan communicated (P = 0.002), and checklist used (P = 0.001). Whilst there was improved compliance with the majority of the checklist items in the post-intervention audit, there was notable improvement in non-invasive blood pressure cycling (P = 0.047), tube tie availability (P = 0.021), vasopressor attachment (P = 0.004), fluids attachment (P = 0.022) and nasogastric tube aspiration (P = 0.004). There were 16 (17%) adverse events in the pre-intervention audit and 5 (5%) in the post intervention audit, a significant reduction in the proportion of all adverse events combined (P = 0.018). Grade and speciality of doctors did not affect outcomes. The intervention has been accepted as the new local safety standards for invasive procedures (LocSSIP) in the unit.

Conclusion: Implementation of a ‘call and response’ verbal checklist for RSI in a critical care unit can improve patient care and safety by minimising human factor errors relating to team communication and improving compliance to national safety standards. Occurence of adverse airway incidents can be reduced with a cohesive and standardised team approach, incorporating both visual and dictatorial prompts to the procedure.

Staff experiences of stressful events in critical care – are we supporting them correctly?

Sarah Paterson, Giulia Sartori and Munita Grover

Northwick Park Hospital, London, UK

Abstract

Introduction: Intensive care is a rewarding specialty but looking after critical patients can be demanding and the risk of burnout is high. Recent studies highlighted that debriefing programs are associated with improved quality of care and favourable patient outcome, but their effect on burnout remains unclear^[1,2,3]. We surveyed the staff of a district general intensive care unit to determine their stress level after significant clinical events e.g. cardiac arrests and what support they were offered or would like.

Objectives: To determine stress levels among staff after stressful events and what support they had received and would like to receive in the future.

Methods: A questionnaire was distributed to staff on a busy district general intensive care unit (ICU) over one month. The ICU has a mix of medical and surgical patients. The first section was the stress questions of the Depression Anxiety Stress Scales (DASS) which has been individually validated. The second section covered the level of support offered following a stressful event and what support staff would want in the future.

Stress level	Respondents
Normal (0–14)	40
Mild (15–18)	8
Moderate (19–25)	4
Severe (26–33)	2
Extremely severe (34–42)	1

Question	Percentage
How often do you take work related stress home?	35%
How often are you supported by peers?	68%
How often are you supported by seniors?	68%
How often have you been offered a debrief after a stressful event?	26%

Results: We had 56 responses to the questionnaire from a variety of staff including doctors, nurses and allied health care professionals. The main stress results are shown in the table below.

Staff indicated how often the following events occurred by marking a point on a line with 0% at one end and 100% at the other. This was converted into a numerical value and then averaged with results below.

Out of the 56 members of staff surveyed 48 said that they feel they would benefit from a regular opportunity to discuss significant clinical events at work.

Discussion: The fact that the majority of stress scores among staff was normal is reassuring, as is their feeling of being supported by peers and seniors. However, 7 people had above moderate stress (12%) and the actual rates may be much higher. One study looking at burnout in critical care nurses in France highlighted symptoms of severe burnout in 33% of respondents [4]. 86% of our respondents indicated that they would like a regular opportunity to discuss cases so we did trial a regular discussion once a week. This was well attended by the junior doctors and generated discussion and learning points. However, there was less engagement from nursing staff. We feel the barriers to this come from work load and short staffing. Overall, the risk of burnout among intensive care staff remains high and most staff said they would like a method to discuss this, however, works needs to continue to develop this further and in a way that is accessible and suited for all staff members.

The sticking point: Challenges associated with clinical-grade mesenchymal stem cell therapy in an ovine model of ARDS and ECMO

Jonathan Millar^1,2, Nicole Bartnikowski¹, Nchafatso Obonyo^1,3, Maximilian Malfertheiner⁴, Viktor Von Bahr⁵, Mengyao Yang¹, Meredith Redd⁶, Margaret Passmore¹, Katrina Ki¹, Sanne Pedersen¹, Nathan Palpant⁶, Jacky Suen¹, Danny McAuley² and John Fraser¹

¹Critical Care Research Group, University of Queensland, Brisbane, QLD, Australia

²Wellcome- Wolfson Institute for Experimental Medicine, School of Medicine, Dentistry & Biomedical Sciences, Queen’s University Belfast, Belfast, UK

³Division of Medicine, Imperial College London, London, UK

⁴Department of Internal Medicine II, Cardiology and Pneumology, University Medical Center Regensburg, Regensburg, Germany

⁵Department of Physiology and Pharmacology, Section for Anesthesiology and Intensive Care Medicine, Karolinska Institutet, Stockholm, Sweden

⁶Institute for Molecular Bioscience, University of Queensland, Brisbane, QLD, Australia

Abstract

Background: Mesenchymal stem cells (MSCs) have attracted attention as a candidate therapy for many inflammatory diseases, including ARDS (1). Interest in these cells arises because of their pleiotropic immunomodulatory abilities in-vivo. During acute inflammation, MSCs appear to be highly immunosuppressive, influencing both innate and adaptive immune responses (2).

To date, trials of MSCs in ARDS have excluded patients supported with extracorporeal membrane oxygenation (ECMO). The use of MSCs in ECMO patients, while attractive, raises some safety concerns. Recent ex-vivo experimentation has demonstrated the propensity of MSCs to adhere to membrane oxygenators (3).

Given the paucity of evidence to support the safe use of MSC therapy during ECMO, we conducted a prospective, controlled trial of clinical-grade human MSCs (hMSCs) in a clinically relevant ovine model of severe ARDS supported with vv-ECMO (CELTIC study).

Methods: In fourteen animals, experimental ARDS was induced using a combination of intravenous oleic acid and intrabronchial E. coli lipopolysaccharide. Thereafter, animals were established on vv-ECMO and assigned to one of two groups; intra-bronchial hMSC (3x10^8 cells) treatment (n = 7) or intra-bronchial vehicle (n = 7). At 23 hours after injury, ECMO was stopped, a standardised recruitment manoeuvre was performed, and the animals were returned to pre-ECMO lung-protective mechanical ventilatory settings. At 24 hours animals were euthanised. Mechanical ventilation, ECMO, anticoagulation, and haemodynamic management were protocolised in both groups. ECMO parameters were continuously recorded, including the trans-membrane oxygenator pressure gradient. At the completion of the study, the membrane oxygenator was deconstructed, rinsed, and stained for MSC cell surface markers.

Results: All animals completed the experimental protocol. Baseline physiological parameters were evenly matched between control and hMSC groups after injury, with a mean PaO₂/FiO₂ ratio of 58 ± 24 mmHg and 61 ± 19 mmHg respectively. Prior to the termination of ECMO, haematocrit levels and the magnitude of anticoagulation (APTT ratio) were similar between groups. There was a significant increase in trans-oxygenator pressures in the hMSC group, accompanied by an increase in the pump rpm to flow ratio, although this was not significant (Figure 1). Immunofluorescence performed on deconstructed oxygenators revealed hMSCs adherent to oxygenator fibres in the treated group (Fig. 1).

Conclusion: This is the first in-vivo trial of hMSC therapy administered during vv-ECMO. Our results present a challenge to the translation of cell therapy in the context of ECMO. In this study, despite intra-bronchial delivery, hMSCs exhibited the ability to enter the systemic circulation and adhere to membrane oxygenators in sufficient quantities to cause an increase in the trans-oxygenator pressure gradient. These findings must be balanced against the promising immunomodulatory properties of MSCs during ARDS (4), a subject of further analysis within the CELTIC study.

Apneic Oxygenation Was Associated With Decreased Desaturation Rates During Rapid Sequence Intubation in Multiple Australian and New Zealand Emergency Departments

Adrian Perera^1,2, Hatem Alkouri^3,4, Toby Fogg^5,6,7, John Vassiliadis^5,6, John Mackenzie^3,8 and Yashvi Wimalasena^1,5,9

¹Emergency Department, Lismore Base Hospital, Lismore, NSW, Australia

²Emergency Department, The Royal Free NHS Foundation Trust, London, UK

³Emergency Care Institute, Agency for Clinical Innovation, Sydney, NSW, Australia

⁴Faculty of Medicine, The University of New South Wales, Sydney, NSW, Australia

⁵Discipline of Emergency Medicine, Sydney Medical School, University of Sydney, Sydney, NSW, Australia

⁶Emergency Department, Royal North Shore Hospital, Sydney, NSW, Australia

⁷CareFlight, Sydney, NSW, Australia

⁸Senior SS Prince of Wales Hospital, Sydney, NSW, Australia

⁹Greater Sydney Area Helicopter Emergency Medical Service, NSW Ambulance Service, Sydney, NSW, Australia

Abstract

Background: Rapid sequence intubation (RSI) in patients with critical illness or injury carries significant potential complications, with the incidence of desaturation being the most common. The technique of Apnoeic Oxygenation (ApOx) has been demonstrated to reduce the incidence of desaturation although evidence of benefit has been conflicting.

Aim: The aim of this study was to compare the incidence of desaturation between patients who received ApOx, as recorded in the Australia and New Zealand Emergency Department Airway Registry (ANZEDAR) study.

Methods: We investigated the data from ANZEDAR (n = 3710) which was collected over 24 months. This study is a retrospective analysis of prospectively collected ANZEDAR database of patient demographics, indication of intubation, oxygenation devices and techniques and patient complications from 43 EDs. To evaluate the effect of ApOx on patient desaturation, we fitted a binomial logistic regression model.

Results: Applying the selection criteria to the ANZEDAR database 2519 intubations were analyzed, 1669 (66.3%) patients received ApOx while 850 (33.7%) did not. Desaturation in the ApOx cohort was 10.4% compared to standard care 13.7%. ApOx had a protective effect for desaturation (OR = 0.72 95%CI 0.54–0.97). Each intubation attempt was associated with associated with increased risk of desaturation (OR = 2.06 95% CI 1.32–3.22). When a difficult airway was anticipated OR of desaturation was 1.89 (95%CI 1.35–2.48)

Conclusion: This multicentre study, using data collected in real world practice, provides evidence supporting the use of ApOx to reduce the incidence of desaturation in ED RSI. ApOx is a pragmatic intervention utilising readily available equipment. The greatest benefits are seen in the anticipated difficult airway failing first pass intubation –thus we support the use of ApOx in all ED patients requiring RSI.

WHO’s talking? Enhancing multidisciplinary communication on the ICU with lessons learned from the surgical safety checklist

Tanya Hall, Kate Regan, Fiona Baldwin and Gavin Atherton

Brighton and Sussex University Hospitals, Brighton, UK

Abstract

Introduction: Poor communication has been reported to be a contributing factor in 60% of all hospital adverse events (1). The use of the WHO team brief and surgical safety checklist in operating theatres is well established. The aim of the brief is simple; to encourage dialogue within multidisciplinary teams and therefore avoid preventable harm to our patients. A twice-daily multidisciplinary team (MDT) brief is also a regular feature of labour ward handover, involving anaesthetists, obstetricians and midwives. We hypothesised that implementing a similar brief within the Intensive Care environment may improve communication and enhance patient care.

Our Trust has 39 critical care beds across 2 sites, with over 150 nursing staff and a regularly rotating junior medical workforce. There are separate nursing and medical handovers but historically no integrated multidisciplinary communication between nursing and medical staff. The team working together will have often not worked together previously. Our primary objective was to ascertain whether an evening MDT briefing would be viewed to have a positive impact on communication and team working within the ICU.

Methods: We conducted a month long survey of all ICU clinical staff. We surveyed current handover practices, people’s knowledge of the team’s skill mix and staff morale. We asked staff if they felt an evening multidisciplinary handover would be useful and asked for suggestions on what they felt would be helpful to include in such a handover (2).

Results: The results of the survey showed that 66% of staff were unsure of the identities and the skill mix of their colleagues (3). Many did not know how much ICU experience the Registrar had or whether they were airway or transfer trained. The survey also confirmed that there was no formal discussion between medical and nursing staff regarding potential admissions and transfers. Our aim was to implement a positive change to address this. 85% of staff surveyed welcomed the introduction of an MDT brief (4).

Incorporating staff feedback, we designed a multidisciplinary handover checklist to be used at the start of the nightshift. This checklist aimed to structure and encourage a conversation between the senior nursing and medical staff present for the shift. The key components identified as important in the checklist included roles, skill mixes, specific concerns regarding patients and expected admissions (5).

Conclusions: Our survey highlighted a lack of understanding of roles and potential areas of poor communication between staff on the ICU. We anticipate that the introduction of the checklist will enhance team communication, staff morale and the running of the unit with a positive impact on patient safety. We intend to resurvey to see if there has been a perceived improvement in interdisciplinary communication and handover.

Hyperoxia Prevention – A quality improvement project in the intensive care unit (ICU) of a tertiary cardiac centre

Matthew Cadd, Nadir Ratanshi, Ben Stretch and Daniel Melley

St Bartholomew's Hospital, Barts NHS Trust, London, UK

Abstract

Aim: To prevent hyperoxia in a population of patients at risk of harm.

Background: Oxygen is a free radical gas which has the potential to cause tissue injury. Studies of oxygen therapy in a range of patient populations have consistently shown worse outcomes in patients who are hyperoxic.¹ In patients following ST-elevation myocardial infarction oxygen therapy is associated with increased infarct size, more frequent arrhythmias and a higher risk of subsequent infarct.² Further, in patients who have return of circulation following out of hospital cardiac arrest there is an association between arterial hyperoxia, in-hospital mortality and poor neurological outcome.³ Finally, in a recent trial in critically ill patients conservative oxygen administration resulted in lower Intensive Care Unit (ICU) mortality compared with conventional oxygen therapy.⁴ Avoidance of hyperoxia is not yet universally recognised as a clinical target for patients in ICUs.

Method: The multi-disciplinary team in a cardiac intensive care unit instituted a quality improvement project to reduce the incidence of hyperoxia in a population of patients at risk. At baseline, oxygen saturations of all patients were measured each day at three time points for one week. Normoxia was defined as peripheral oxygen saturation of 94–98%, and hyperoxia as 99% or 100%. Educational sessions were held, focusing on the importance of hyperoxia avoidance. These included a presentation to the consultant body and doctors in training, a small group discussion with the matron and Band 7 nurses, presentations during the morning ‘safety huddle’ for the clinical team and lastly target setting during ward rounds. Patient oxygen levels were measured one month later.

Results: At baseline 130 (53.9%) measurements of oxygen saturation in 17 patients were in the hyperoxic range. After initial intervention, 47 (21.8%) of observations in 15 patients were hyperoxic.

Conclusion: Hyperoxia was not widely recognised as being harmful and had a high prevalence at baseline. Substantial improvement was seen following the institution of a QI project for a group of at patients particularly at risk of hyperoxia (following myocardial infarction and out of hospital cardiac arrest). This on-going project will seek to prevent hyperoxia in patients admitted to the cardiac ICU. We will repeat assessment with fortnightly measurement of patient oxygen saturations and feedback team performance to the whole MDT. We will implement further improvement interventions as required to meet the project aim.

Swallow pathophysiology and dysphagia treatment tolerance in cardiac surgical intensive care patients following tracheostomy and/or prolonged intubation

Sallyanne Duncan¹, Danny McAuley¹, Margaret Walshe² and Bronagh Blackwood¹

¹Wellcome- Wolfson Institute for Experimental Medicine, School of Medicine, Dentistry & Biomedical Sciences, Queen’s University Belfast, Belfast, UK

²Department of Clinical Speech and Language Studies, Trinity College Dublin, Dublin, Ireland

Abstract

Background: Prolonged intubation and/or presence of a tracheostomy tube are recognised risk factors for dysphagia. A small number of prospective studies evaluating swallow pathophysiology using fibreoptic endoscopic evaluation of swallowing (FEES) in this population have been reported. To date no studies have described physiologically based treatment approaches and the feasibility and acceptability of recommended treatment intensities as per guidelines for the provision of intensive care services (GPICS 2015) has not been tested (1).

Methods: A service evaluation was completed in a single Cardiac ICU centre to gather prospective data on the presence, severity and recovery of dysphagia according to duration of intubation. An exploratory sample of forty patients received bedside swallowing evaluations by the speech and language therapy (SLT) team following extubation. Those patients presenting with severe dysphagia underwent FEES on average 3 days later to further investigate swallow physiology. These assessments were completed 15 days following surgery. A treatment protocol was commenced combining swallow resistance exercises and thermal tactile stimulation therapy (TTS). Exercises involved a tongue-hold-swallow, known to improve contact between base of tongue and pharyngeal wall by increasing contraction of pharyngeal wall muscles. TTS used a cold probe to contact and stimulate both faucal arches lying anterior to the oral tonsils to promote a more timely swallow response. Each swallow exercise and TTS set was repeated 10 times per session. Four treatments a week, scheduled to last 45 minutes each, over a three week period were provided.

Results: Fifteen of the forty patients, intubated for >5 days, showed highest incidence and severity of dysphagia. Swallow function took longest to recover in this group. Ten patients consented to FEES. There were more males than females, mean age 76, mean intubation duration 12 days, 8 patients had combined surgery and 2 had valve surgery. Following FEES, SLT recommended nil orally for 6 patients, due to inefficient, unsafe swallowing and high risk of silent aspiration (Table 1). The remaining 4 patients commenced restricted oral trials with nasogastric feeding. Seven of the 10 patients agreed to participate in treatment. Sessions were frequently truncated due to patient fatigue and lasted on average 24 minutes. No patient tolerated 45 minutes of direct treatment per session. Treatment outcomes are highlighted in Table 2.

Conclusion: This project demonstrated a high prevalence of impaired oropharyngeal swallowing associated with prolonged intubation and tracheostomy after cardiac surgery. It is unclear from the findings in this group if treatment was effective. Recommended treatment intensity as per national guidelines was poorly tolerated.

Table 1.

FEES findings for 10 patients.

Delayed onset of pharyngeal swallow	80%
Food/Fluid residue in pharynx after swallow	100%
Unprotective cough	70%
Silent aspiration	50%
Abnormal laryngeal findings	60%

Table 2.

Treatment outcomes for 7 patients.

Diet recommendations	No of patients.
Normal diet/fluids	2
Modified diet	3
Nil orally & nasogastric feeding	2

The Use Of Computer Generated Discharge Paperwork to Reduce Errors and Increase Efficiency

Jeevakan Subramaniam¹, Brandon McCaffrey¹, Jack Siebert², Ana Howarth³ and Dan Harding¹

¹Department of Intensive Care Medicine, Queen Elizabeth Hospital, London, UK

²University College London, London, UK

³St George's University of London, London, UK

Abstract

Computer based admin work represents up to 69% of trainee time spent at work[1]. There are often errors in this paperwork which poses a risk to patient safety[2]. Structured proformas do not make a significant difference to the quality of discharge paperwork, yet many hospital departments rely on them when writing discharge summaries[3]. This project aimed to audit discharge summaries and implement a macro-enabled Microsoft Word document to improve the quality of discharge summaries from the ICU at Queen Elizabeth Hospital, Woolwich.

18 essential items were identified as necessary in all discharge summaries including demographic information, diagnosis, presentation and plan. A macro-enabled document was created with these fields, which once filled generates letters to both the GP and ward staff when stepped down. This form underwent two feedback cycles with trainees, including changes to deny autocompletion without filling all essential fields, and highlighting incomplete fields.

One month of discharge summaries prior to intervention were audited against one month of discharge summaries post intervention. Trainees filling out discharges were also asked to self-report time taken to fill out discharge summaries, pre and post intervention. Chi Squares were conducted for entries that did not improve to 100%. Time spent was analysed using the Mann-Whitney Test.

The tool developed for this project greatly increased the overall completeness of the discharge summaries and step-down handovers – All essential fields either improved to 100% or were significantly improved (p < 0.01) barring one. Mean time spent also significantly improved (p > 0.05). The tool saved a mean time of 13 minutes per discharge summary, but did not have a 100% success rate – due to trainees using punctuation marks to circumvent the tool.

In conclusion, while traditional proformas do not affect the completeness of discharge summaries, more sophisticated tools which mandate completion of essential information can be useful in ensuring continuity of care. Such solutions can be more agile than large systems implemented in hospital software and can undergo multiple feedback cycles from trainees quickly. These tools can also save time for trainees, saving the department money and freeing up valuable trainee time to attend to patients, engage in learning[4] and increase job satisfaction[5].

Synchronization analysis between heart rate variability and quantitative EEG parameters in a patient with progressive cerebral edema

Jeong-Hwa Lee, Han-Bin Lee, Jun Young Jang, Sang-Beom Jeon and Hyun-jo Lee

Department of Neurology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, The Republic Of Korea

Abstract

In neurocritical care, there has been great interest in the development of an early warning system to predict secondary complications after severe brain injury. Heart rate variability (HRV) has been used to assess cardiac autonomic activity in critically ill patients. Furthermore, HRV monitoring is recently shown to be an effective method for preclinical detection of sepsis and delayed ischemia after subarachnoid hemorrhage, using existing physiological data. The present study describes the case of a 36-year-old female who admitted to a neurointensive care unit for cerebral venous thrombosis. Concern for seizures led to continuous EEG monitoring (cEEG) but did not demonstrate any ictal activity. HRV values were automatically measured and recorded every five minutes and used in conjunction with cEEG. About 10 hours after admission, faster frequencies on cEEG were attenuated and the alpha-delta ratio (ADR) on quantitative EEG (qEEG) decreased, with relative increases in HRV variables (SDNN, RMSSD, pNN50, and total power). Approximately 20 hours after this episode, her right pupil dilated and CT scans revealed malignant cerebral edema. Subsequent correlation analysis using Pearson’s test was performed. Decreasing ADR (r = −0.31, p < 0.0001) or increasing absolute asymmetry index (r = 0.22, p < 0.001) were significantly correlated with increasing HRV variables in this patient. One possible hypothesis for this finding is the increase in sympathetic activity during the status of acute increase in intracranial pressure which was described in previous basic and clinical reports. However, the sympathetic hyperactivity in patients with a severe brain injury could not fully explain this paradoxical autonomic reaction including parasympathetic tone. To the best of our knowledge, this case report was the first to analysis HRV monitoring synchronized with qEEG data immediately on admission, which may have an implication for future research defining that the clinical and pathophysiological significances of HRV in patients with progressive cerebral edema.

Early identification of psychological symptoms in critical care with a new software system helps target interventions aimed at reducing long term morbidity

Timothy Baker¹, Vilas Navapurkar¹ and Katy Surman²

¹Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK

²University of Cambridge, Cambridge, UK

Abstract

25–65% of patients who survive critical illness suffer from Post Traumatic Stress Disorder¹. An inability to effectively communicate is the most stressful aspect of being an ICU patient^2,3 often because of a delay or failure to treat uncommunicated symptoms. Early interventions have been shown to have the greatest impact on reducing psychological morbidity. ⁱ We initiated a quality improvement project to reduce the symptoms experienced by ICU patients. myICUvoice^Ó was introduced to enable better communication between patients and nurses. Consequently allowing clinicians to better treat the symptom burden of individual patients. Investigations into the efficacy of myICUvoice are on-going. We report an early observation.

All patients were provided continuous access to a tablet with myICUvoice^Ó loaded onto it. myICUvoice^Ó is a ICU specific symptom communication tool that allows all patients to self-report symptoms. Nurses were encouraged to ask the patients to complete a daily survey to understand “what they are feeling?” as opposed to a less specific “how they were feeling?”. Questions focused on physical, emotional, ICU intervention and pain related symptoms. Patients were also able to self-report any of the 84 symptoms listed at any time during their admission. Response data was compared to the ICNARC dataset for these patients.

In 6 months, 278 patients used the tool. Anxiety was the most frequently reported psychological symptom (reported by 31% of patients). Frequency of this symptom was similar across all admitting conditions. Correlating length of stay with anxiety demonstrated a trend towards a reduction in the number of anxiety reports as the length of stay increased. After classifying patients as “heavy” or “light” users a clear distinction was seen: heavy users were far more likely to report being anxious. No symptoms were clearly associated with increased anxiety but 32% of anxiety reports occurred concurrently with xerostomia.

In this early analysis patient numbers are not sufficiently powered to attain any statistical significance. As we continue to use this tool routinely we are gaining new knowledge about what ICU patients feel and how this varies between different conditions. With a better understanding of the incidence and at risk populations, we are increasingly able to better target early interventions with our specialist psychologist. The early results of a related and on-going ethnological study suggests that heavy users are likely to be those with tracheostomies who are CAM ICU negative and RASS 0. We are developing ways to proactively improve the symptoms for this cohort.

Development of an ‘eResus’ bundle for use in the Emergency Department

Robert Morgan, Emma Colley, Jessica Spedding and Stigaard Laird Helen

Royal United Hospital, Bath, UK

Abstract

Background: The resuscitation area of the Emergency Department (ED) can be a stressful environment in which to provide patient care. Good decision making can be hampered by crowding, patient acuity and working within large, sometimes unfamiliar multidisciplinary resuscitation teams. Rapid access to relevant clinical guidelines enhances the delivery of high quality care under such conditions, frequently by means of an electronic application (1–3).

This quality improvement project (QIP) is based at Royal United Hospital (RUH) Bath, a 565-bed district general hospital and major trauma unit. The RUH provides acute medical care for both adults and paediatrics from a population of approximately 500,000 people across Bath, North East Somerset and Western Wiltshire.

ED guidelines are currently found in a single lengthy PDF document. Even for those familiar with the document, locating individual guidelines and their resuscitation elements can slow down delivery of care. Individual resuscitation bays do not have desktop computers so real-time near-patient access to guidelines is unavailable, requiring practitioners to leave the bedside to access essential information.

Objectives:

• Develop a digital eResus package providing rapidly accessible up-to-date local and national guidelines for use in emergency situations.

• Quick links to other essential emergency information, including checklists and complicated resuscitation drugs.

• Install bedside touchscreen computers allowing rapid, user-friendly access to time-critical guidance.

• Enhance communication and the direction of clinical care.

• Develop a clear system to ensure version control.

Methods & Results:

First PDSA cycle:

• Plan – A preliminary survey of doctors and nurse practitioners before introducing eResus showed that times to access specific guidelines from the PDF ranged from 20 seconds up to 4.5 minutes (median 1.5 minutes).

• Do – 10 adult guidelines were adapted into the eResus format and staff were timed accessing them on a dedicated laptop.

• Study – Staff found the content of eResus guidelines very useful and suitable for use in emergencies. Time to access guidelines was reduced (range 22 to 42 seconds, median 27 seconds).

• Act – We completed adapting the remaining resuscitation-specific guidelines for the eResus package and developed a business plan to apply for touchscreen computers in each resus bay.

The second cycle is currently underway.

• Plan – All guidelines have been incorporated into eResus, including sub-groups of adult and paediatric clinical guidelines, adult and paediatric resus drugs, the WHO checklist and rapid sequence induction checklist.

• Do – Time to access various guidelines on a dedicated computer has shown access times between 5 and 15 seconds. Funding for touchscreen computers and their installation has been secured from the Organ Donation Trust, and the order has been placed.

• Study – Once the touchscreen computers are installed, we will undertake a further review of time to access guidelines along with a questionnaire to establish team satisfaction.

Conclusions: This work remains in progress. However, compelling evidence suggests that this project should meet its objectives. Guidelines will be more resuscitation-focussed, with only the pertinent information available. It appears that eResus will improve both patient safety and clinician confidence.

Mechanical circulatory support remains a complex tool in the management of acute myocardial infarction with refractory cardiogenic shock

James Barnacle, Francisca Caetano, Hatem Soliman Aboumarie, Shahana Uddin and Susanna Price

Royal Brompton and Harefield NHS Foundation Trust, UK, London, UK

Abstract

Background: Mechanical circulatory support (MCS) including ventricular-assist devices (VAD) and venoarterial extracorporeal membrane oxygenation (VA-ECMO) have revolutionised the treatment of refractory cardiogenic shock (CS). Despite advances in interventional cardiology, the commonest cause of CS remains acute myocardial infarction (AMI), with mortality remaining approximately 50%.¹ Recent guidelines have downgraded recommendations for the use of intra-aortic balloon pump (IABP) in patients with AMI, and suggest that other MCS may be considered, either as a bridge to recovery, longer-term VAD or transplantation. Here, the use of online calculators to predict outcome/survival are recommended^2,3, however, it is recognised that the acuity of presentation may necessitate implementation of MCS as a bridge to decision.

Case: A 56-year-old man presented to his local hospital with symptoms of crescendo angina followed by chest pain suggestive of myocardial infarction/ischaemia. ECG on admission showed anterior ST-elevation MI. He was immediately transferred for primary percutaneous coronary intervention (PPCI) with stenting to 100% LAD and 75% mid-RCA occlusions. Despite this, he developed CS with pulmonary oedema, and an IABP was inserted. Initial echocardiography demonstrated severe LV dysfunction with EF <30% and moderate RV dysfunction. Despite this, he remained INTERMACS I (worsening cardiac index and mitral regurgitation and rising pulmonary capillary wedge pressure despite escalating inotropes), therefore an Impella 2.5 (Figure 1) was inserted as an interim measure to reduce mitral regurgitation (Figure 2) and allow transfer to a VAD and transplant centre.

He was commenced on a Levitronix left ventricular-assist device (LVAD), however, this was complicated by uncontrolled bleeding, and eventually the decision was made to palliate on grounds of futility – ongoing haemorrhage and multi-organ failure (haemofiltration-dependence, vasogenic brain oedema with seizures, worsening lung function, and no improvement in cardiac function) despite optimal critical care interventions.

Discussion: MCS is increasingly used in CS management, including after AMI, and is recommended in current guidelines until organ recovery, or as a bridge to decision (BTD).³ However, percutaneous MCS does not appear to reduce mortality in this population^4,5, although they are still used as an interim option, or to offload the heart in the context of ECMO. This case demonstrates the challenges relating to MCS including the disappointing outcomes relating to CS complicating AMI requiring dual antiplatelet therapy. Despite his relatively young age, no associated cardiac arrest, prompt diagnosis, early referral to an ECMO centre and use of intermediary and adjunctive support devices, he progressed to worsening multi-organ failure and died. It highlights the signal for caution in the use of MCS in CS complicating AMI in current guidelines, and reinforces the importance of the post-AMI CS population as the next focus for research for interventional cardiology and critical care.

Figure 1.

Transthoracic echocardiogram showing Impella in situ across the aortic valve. RV; right ventricle, LV; left ventricle, AV; aortic valve, LA; left atrium.

Figure 2.

Transthoracic echocardiogram with colour Doppler demonstrating an Impella in situ, as well as mild mitral regurgitation (arrowed, blue). LV; left ventricle, RV; right ventricle, RA; right atrium, LA; left atrium, MR; mitral regurgitation, DA; descending aorta.

National Early Warning Score and time to antibiotics – an audit of emergency department practice at Northampton General Hospital NHS Trust

Anish Khandhia, Jonathan Hardwick and Jonny Wilkinson

Northampton General Hospital NHS Trust, Northampton, UK

Abstract

Background: The identification and timely management of patients with sepsis remains a challenge in secondary care. Multiple studies have demonstrated that early provision of antibiotics improves outcome in sepsis. The Sepsis Trust ‘Sepsis Screening Tool’ (SST) aids the identification of patients with potential sepsis who would benefit from early antibiotic administration. It should be applied to those patients showing signs of physiological disturbance (National Early Warning Score (NEWS) >3) where infection is a suspected cause. Commissioning for Quality and Innovation (CQUIN) targets dictate that 90% of patients with one or more ‘Red Flag’ criteria (signs of end organ dysfunction) should receive antibiotics within one hour of presentation. The aim of this audit was to determine local compliance with these standards.

Methods: Emergency department data across an 11 month period in 2017 were retrospectively analysed to identify all patients with a presenting NEWS >3 with clinical suspicion of sepsis. The number of patients screened, the number with Red Flag criteria and the time from presentation to antibiotic administration wa reviewed. NEWS were divided into the ranges 3–5, 6–8, 9–11 and >11 and the time to antibiotics for each NEWS range was assessed.

Results:

Screening

522 patients presented with NEWS scores >3 and suspected signs of infection. Of these 477 (91.3%) had a completed SST. Red Flag criteria were present in 389 (81.2 %) patients of whom 269 (69.1%) received antibiotics within one hour of presentation.

NEWS and Time to Antibiotics

The number of patients in each NEWS range is illustrated in Figure 1.

Conclusion: The results demonstrate that over 90% of eligible patients were appropriately screened for sepsis. Additionally the proportion of patients receiving antibiotics within one hour falls close to the target of 90%. We also have demonstrated that as NEWS increases, the time to antibiotic administration decreases.

Figure 1.

The median time across all groups to antibiotic administration was 41.7 minutes.

Figure 2.

The relationship between NEWS and time to antibiotic administration.

Actions arising from this audit are to produce clearer written criteria for the application of the SST in our emergency department, a change to our electronic patient management system to highlight these patients to staff, and a written escalation of care policy to ensure Consultant-grade clinical input earlier. These interventions will be implemented and a re-audit undertaken in 2019.

Diarrhoea in critical care is common, rarely infective in origin, and associated with greater length of stay and increased mortality

Mayur Murali, Clare Ly and Hugh Montgomery

Whittington Hospital, London, UK

Abstract

Introduction: Diarrhoea, defined as ≥3 loose or liquid stools/day¹, affects 9.7–41% of Intensive Care Unit (ICU) patients.² It impacts negatively on patient dignity, mobility and self-care; intensifies nursing workload; and increases patient morbidity through fluid and electrolyte imbalance, dermal injury and impaired nutritional state. Such factors may contribute to increased patient length of stay (LOS), cost of care and mortality with which it is associated.³ The pathogenesis of diarrhoea in the critically ill is poorly understood⁴, but factors such as infection, the administration of feed or medication (e.g. laxatives), and changes in the gut microbiome, its structure, function and perfusion, may play a role.⁵

Methods: The Whittington Hospital is a 360-bed teaching hospital in North London. Computerised records were analysed to identify patients >18 years old who were admitted to its 15-bed mixed medical/surgical ICU in the 60 months between 1/2/2013 to 1/2/2018. The Whittington has clearly defined protocols for sending stool samples for microbiological or virological analysis based on the Bristol stool chart scoring system. Patients who had a stool sample sent were judged to have had an episode of diarrhoea. Demographic data was collected, and laxatives received were documented. Measures of outcome including length of stay (LOS) and mortality were recorded. Patient data were extracted from the ICU database and stool data from the hospital laboratory software (ICE Anglia). Data were analysed using Microsoft Excel 2010.

This work was registered as an audit at the Whittington, and was regarded as service evaluation. All data was anonymised. As such, no ethical approval was required.

Results: 3,737 patients were admitted to the ITU (mean age 61 +/− 18.9), 1,912 (51.1%) male, 1,328 (35.5%) from surgical admissions, median APACHE II score 14 (interquartile range 9–19). Diarrhoea was found in 199 patients (prevalence of 5.3%). Patients with diarrhoea had a higher APACHE II score and were less likely to be surgical. Thirteen (7%) had pathological microbiological stool samples. One sample was positive for Campylobacter, and 12 for C. diff. There were no positive virological samples. Thirty-four/199 (17.1%) received laxatives <24 h prior to onset of diarrhoea, while 3/12 (25%) of C. Diff cases received laxatives within 24 h of a positive sample. Median ITU LOS was 2.5 days (interquartile range 1.1–5.5 days) and ITU mortality was 11% (412/3,737). Patients admitted with diarrhoea experienced longer median ITU LOS (median (IQR) 2.3 (1.0–5.0) days vs 10 days (5.0–22.0), p < 0.001) and greater ITU mortality (9.5% (356/3,727) vs 18.1% (36/199), p = 0.005).

Discussion: This study found diarrhoea in ITU to be relatively common, affecting 1 in 20 admissions, and associated with increased ITU LOS and mortality. There was a low yield for stool investigations (7%), and use of laxatives prior to diarrhoea onset was common (17.1%). Our findings suggest regular review of patient’s gastrointestinal status, with prompt recognition, investigation and treatment of diarrhoea, should be performed. Clear protocols should be developed for prescribing and administering laxatives. The low diagnostic yield suggests rationalising of investigations could help reduce financial costs.

Renal replacement therapy in the ICU: A case for ward-based RRT

Sofia Hanger, Ashley Douglas and Viv Sathianathan

London North West University Hospitals, London, UK

Abstract

The incidence of patients requiring renal replacement therapy (RRT) for chronic kidney disease (CKD) is increasing, with over 63000 patients receiving RRT in 2016. These patients often require admission to a critical care setting when admitted to hospital with an intercurrent illness. We reviewed patients who received RRT in our large, district general hospital ICU who were known CKD patients (established on community RRT or not) and patients who were transferred to our regional renal unit for ongoing care. We focused on the requirement and duration of organ support and how long patients were fit for discharge from critical care, except for the need for ongoing RRT. This enabled us to establish how many potential patients could benefit if a ward based RRT service was introduced.

This is a retrospective observational study. Patients with a primary admission diagnosis of chronic kidney disease and all patients transferred to the regional renal unit from ICU for ongoing care with new or pre-existing renal disease over a 12 month period were identified. Patients who were palliated prior to discharge, were transferred to another ICU or had incomplete notes were excluded. Their primary diagnosis, total length of stay, organ support requirement and length of stay requiring RRT only were analysed using a combination of ICNARC database data, discharge summaries and patient notes. Basic cardiovascular monitoring was sub-divided into patients receiving invasive blood pressure monitoring only and patients who required additional inotropic or vasopressor support.

58 patients were identified, of whom 7 were excluded, out of a total of 796 patients over 12 months. 84% of patients had pre-existing chronic kidney disease and 41% of patients were already established on long-term RRT.

44 patients required RRT only, with or without invasive blood pressure monitoring, for one or more days during their ICU stay. Of these, 25 patients did not require any organ support additional to this during their entire ICU admission. The mean duration patients required RRT only was 5.8 days. The total combined length of stay requiring RRT only (with or without invasive blood pressure monitoring) for all patients was 297 days, of which 121 days was in pre-existing dialysis patients. This equates to 5% of our total ICU bed day capacity over this 12 month period.

These results show that the development of a ward-based RRT service has the potential to create an additional 297 ICU bed days per year in a busy district general hospital. The target population for this intervention would be inpatients requiring no or minimal other organ support who are not suitable for outpatient dialysis or are awaiting transfer for initiation of long-term RRT. Such a service has the potential to reduce ICU bed utilisation in a time of unprecedented demand. demand.

National Restructuring of Medical Specialty Training to include Critical Care, But How do we train this group of trainees?

Nadia Paes¹, Thomas Williams² and William Angus¹

¹Wirral University Teaching Hospital NHS Foundation Trust, Liverpool, UK

²Health Education England: North West Postgraduate Medicine and Dentistry, Manchester, UK

Abstract

Background: A radical national restructuring of medical specialty training has been proposed from 2019 that will see Core Medical Training (CMT) replaced with the Internal Medicine (IM) Stage 1 training programme. This will include a compulsory critical care component

As few CMT intensive care unit (ICU) posts critical currently exist, there is little evidence to support how intensivists should best train medical trainees in this environment. This study provides insight into this area, examining the educational experiences of CMTs undertaking ICU placements at a large teaching hospital in a 2 year period.

Methods: A mixed methods approach of questionnaires and interviews was used. Participants were CT2 at the time of the placement, therefore CT2, ST3 or staff grade specialty registrar equivalent during the study. Fifteen of sixteen eligible trainees completed questionnaires. Inductive thematic analysis was used to analyse open question data and produce an interview guide. Semi-structured interviews explored questionnaire themes further with eight trainees. A hybrid inductive and deductive thematic analysis method, utilising a coding framework produced from the questionnaire data, was used to structure the analysis while also coding for emergent themes.

Results: Increased understanding of intensive treatments was a key outcome from the study. CMTs were better able to understand levels of care, escalation and patient suitability and had improved awareness of negative patient outcomes following intensive intervention. Understanding of intensive treatments was supported by situated learning and by the multidisciplinary team (MDT) via a community of practice (CoP). Improved understanding of critical care treatment supported targeted and more appropriate referrals to ICU when trainees returned to their general medical roles. This was particularly noted by trainees that went on to become medical registrars.

The value of learning in a senior led environment was a further study outcome. Learning alongside an ‘expert' allowed rapid development beyond baseline capabilities. Similarities to an anaesthetic approach of one-to-one teaching, grounded in a cognitive apprenticeship approach, were noted. The improved teacher-student relationship, compared to general medical experiences, facilitated teaching better aligned to learners needs and opportunity for feedback. Immediate, targeted and constructive feedback was provided during discussion of referrals. High senior presence was, however, associated with negative features of decreased autonomy and circumvention of juniors’ opinions by referring teams and the MDT.

Feedback and experience of managing high acuity patients, predominantly through reviewing referrals and Medical Emergency Team (MET) calls, resulted in a perceived improvement in performance. Confidence, communication and prioritisation skills in particular were felt to develop as a result. MET calls allowed trainees opportunity to learn a systematic approach to assessing acutely unwell patients, observe a range of medical registrar peers and led calls. Supported learning was valuable, but autonomy with referrals and MET calls allowed testing and reflection.

This study showed ICU experience supported CMTs in their transition to medical registrar roles and provides insight into provision of training for internal medicine trainees in critical care under the propsed restructuring of medical specialty training.

Management of Devastating Brain Injury in a District General Hospital – Reflections on the FICM Consensus Statement

Guy Parsons ^1,2

¹NIHR Academic Clinical Fellow in ICM, Oxford, UK

²Specialist Registrar in Intensive Care Medicine, Milton Keynes University Hospital, UK

Abstract

Background: Early prognostication in perceived Devastating Brain Injury (DBI) is difficult and there is considerable debate on best practice.(1) The dearth of controlled studies limits the evidence basis and is unlikely to be rapidly remedied. A recent consensus statement endorsed by the Faculty of Intensive Care Medicine aims to provide guiding management principles which serve as a useful framework for evaluating current practice in DBI.(2)

Method: We reviewed the notes of all patients presenting to Milton Keynes University Hospital between April 2017 and April 2018 with “any neurological condition … assessed at … admission as an immediate threat to life or incompatible with good functional recovery and where early limitation or withdrawal of therapy [was] being considered” – the statement’s perceived DBI definition.

After exclusions 17 of a potential 56 cases were included. We evaluated these against the statement principles and further criteria developed locally to inform ongoing prospective audit. Timelines were drawn evaluating the timeliness of investigations, the promptness of Intensive Care admission, and the time allocated for prognostication. The reasons for and timing of withdrawal of life sustaining treatments (WLST) were reviewed in order to understand local practice in the transition from active management to palliative care.

Our relationship with our region’s specialist Neurocritical Care unit was also examined in terms of considerations for transfer of care and the reasons for acceptance or rejection of transfer.

Results: The majority of cases reviewed demonstrated a clear awareness of the importance of conducting time-sensitive investigations without delay. The median time from attendance to CT scan of the head was less than 1 hour. The median time from attendance to Intensive Care admission was 2.5 hours.

All patients were provided with a period of active management to maintain physiological stability, however this period ranged from under 3 hours to nearly 12 days. The broad range of duration for active management was delineated into two main categories. The first and larger; those patients where rapidly deteriorating physiological and/or neurological parameters prompted early discussion of WLST. The second; those patients who remained cardiovascularly and neurologically stable who underwent extended observation for prognostication. 93% of perceived DBI patients admitted to our Intensive Care unit died.

65% of patients were discussed with our Neurocritical Care unit for consideration of transfer, of these 73% were declined. Consequently the majority of patients were supported within our Intensive Care unit, 64% with rapidly deteriorating physiology or brainstem death prompting early consideration of WLST. 75% of patients with a WLST decision were referred to organ donation.

Conclusion: Our study records real-world experience of managing DBI in a DGH and raises a number of points of interest when considering resource allocation, duration of support and the relationship of peripheral to specialist Intensive Care units which will be further expanded herein. This work provides a structure for the ongoing analysis of the implementation of the consensus statement’s principles and a useful picture of DGH DBI management. Regional expansion is planned to capture broader trends.

Can simulation within critical care influence organizational learning?

Edit Castren, Suhas Kumar and Stephen Hutchinson

Norfolk & Norwich University Hospital, Norwich, UK

Abstract

Research into organizations has demonstrated three main factors which are essential for the organization to learn, be flexible and reduce errors. This is defined as organizational learning, with three building blocks consisting of a supportive learning environment, introducing concrete learning processes, and having a leadership which reinforces learning. We modified a survey designed by Harvard Business Review to assess these aspects of organizational learning within our critical care unit. This questionnaire-based survey was composed of 47 questions assessing the three building blocks.

A baseline survey in 2015 with 48 respondents had identified a poor organizational learning culture in our critical care. Whilst introducing in-situ simulation, we decided to evaluate its impact on organizational learning. We repeated the survey in 2018 following 3 years of a regular in-situ simulation program. Fourty-four healthcare professionals completed the second survey in 2018, 31 respondents had taken part in simulation and 13 respondents had not taken part in any simulation.

When comparing the results, critical care staff agreed or strongly agreed (simulation vs no simulation) that “it is easy to speak up” (68% vs 46%), “it is easy to discuss problems and disagreements” (61% vs 31%), “staff are open to alternative ways of getting things done” (45% vs 38%), “differences are welcome” (39% vs 38%). In our unit, staff strongly disagreed or disagreed that “mistakes are held against me” (61% vs 54%), “differing opinions aren’t valued” (39% vs 46%), and “the unit is not open to new ideas” (42% vs 38%). These changes seem to mostly support an improvement in the first building block, creating a supportive learning environment. There was no difference between the simulation and non-simulation groups in most questions relating to aspects of developing concrete learning processes and aspects of leadership.

Briefing is a fundamental part of the simulation sessions, with a deliberate effort directed towards creating a safe learning environment. Perhaps, therefore it is not surprising to see a positive influence of simulation on the first building block of organizational learning.

Whilst simulation is an effective learning process, we saw no evidence from the surveys that staff perception of concrete learning processes within our unit has improved. This may reflect a dose-response effect, many staff having experienced only a small number of simulation episodes. But also, may reflect the need for a more diverse range of approaches to address all the learning needs. Integrating simulation with clinical governance and regular involvement of senior members of staff may be a way to drive new learning processes and help reinforce learning.

The limitations of the study are that the results may be multifactorial as during this time there has been significant changes in staff, and implementation of departmental changes (e.g. daily multidisciplinary team meeting).

Difficult airways outside operating theatres: Gaps in knowledge, skills and equipment

Henry Nash, Gabriella Wong, Natasha Lawrence, Jayachandran Radhakrishnan and Vallish Bhardwaj

Mid Essex NHS Trust, Broomfield, UK

Abstract

The appropriate management of a difficult airway requires the simultaneous presence of personnel with adequate knowledge and procedural competency along with access to appropriate equipment. This combination can be difficult to achieve out of theatre environments and represents a significant safety risk for patients, as highlighted by NAP4.¹ We conducted an audit in our hospital to identify the magnitude of this problem.

Methods: We conducted a two part audit in Broomfield Hospital, a district general hospital offering tertiary services in burns and plastics, head and neck and upper GI surgery. The first part was a survey of the medical and nursing staff in our general ICU to quantify the knowledge and skills in managing difficult airways. In particular, we checked for training and experience in managing difficult airways, knowledge of national guidelines and familiarity with locally available equipment. The second part was an inspection of the difficult airway trolleys to verify compliance with the 2017 Difficult Airway Society (DAS) guidelines.²

Results: We obtained responses from 49 medical and nursing staff. Results were poor in both the knowledge and skill domains. 67.3% (n = 33) of the medical staff were unaware of the DAS 2017 guidelines. 58% (n = 28) of the respondents were unaware of the equipment available in the difficult airway trolley. Only 53% of the doctors in critical care were aware that the emergency front of neck access kit in the difficult airway trolley allowed for both surgical and Seldinger techniques. There was also a disconnect between experience and equipment in critical care, with 73.6% of the doctors reporting more experience and confidence with the McGrath videolaryngoscope compared to the more immediately available KingVision scope.

Difficult airway guidelines³ were present in all difficult airway trolleys, but the guidelines in the critical care were outdated (2004 guidelines). None of the difficult airway trolleys included the 2017 DAS guidelines for airway management in the critically ill patient. Also, no locally approved checklists were available for use in this cohort as recommended by the DAS guidelines and the NAP4 report.

Discussion: Difficult airways requiring invasive interventions are rare especially outside theatres. However remote areas present significant problems with training, maintaining skills, staying up to date with guidelines, and areas like ICU can easily be missed. Improving these issues requires a systematic approach.

We have implemented several changes, with other proposed solutions planned. These include an airway lead; standardisation of airway trolleys and updating of guidelines, with a clear schedule for follow up and maintenance; further standardised training and an additional session in the induction programme; and teaching sessions for the ICU nurses. We’ve also introduced a local intubation checklist to assist in adequate preparation and planning prior to intubation.

The aim is to provide consistency in skills and equipment familiarity for people working in the department.

Conclusions: Knowledge of DAS guidelines is inconsistent & a standardised skill set is not always available. However highlighting this is an important first step to improve knowledge and skills, and allows these issues to be addressed.

Rib fractures: Implementing a Chest Injury Pathway in a district general hospital is associated with improvements in analgesia, standardization of care, and decreases in overall and pneumonia mortality

Oliver Quick¹, Neil Roberts², Emma Harrison¹, Laura Shepherd¹, Rhys Owens¹ and Ben Warrick¹

¹Royal Cornwall Hospital, Truro, UK

²Derriford Hospital, Plymouth, UK

Abstract

Background: Rib fractures represent a significant proportion of trauma seen in Emergency Departments. Analgesia and respiratory support represent the cornerstones of management. Previous audit showed low rates of patient-controlled analgesia and regional anaesthesia, variable pathways of care and increased mortality compared to national average. A pathway was introduced to improve care. This audit assessed compliance and effect and compared to pre-pathway data.

Methods: Retrospective audit of all adult patients with rib/sternal fracture from trauma, admitted for active treatment to a district general hospital over 6 months (Aug 2017-Jan 2018). Patients identified through TARN and WebPACS imaging system, cross-referenced then imaging and notes reviewed. Demographics and characteristics of injury were recorded, along with markers of care such as level of trauma call, maximum imaging, critical care and analgesia, and outcomes including length of stay (LOS) and 30-day mortality.

Results: 69 patients identified for inclusion after review of 191 imaging reports and 80 sets of notes (increased from 43 previously). Overall 30-day mortality was 8.69%, reduced from 11.6%. 39 patients documented use of pathway (60%). 64 patients had sufficient data for further analysis. Median age 68 (range 18–100). Median of 4 fractures (range 0–10). 11 had flail chest (17%). Median Charlson Comorbidity Index was 4 (range 0–15). 12 patients (18%) were treated for pneumonia, increased from 16%. Mortality was 0% in those treated for pneumonia, decreased from 29%. Median hospital LOS 7 days (range 1–37), increased from 6. Median ICU LOS was 4 days (range 1–7), increased from 3.

11 patients (17%) received regional anaesthesia, increased from 2 (5%) previously. 5 (8%) received erector spinae catheters. 7 patients (11%) received epidural anaesthesia (1 when ES catheter failed). 28 (44%) patients were prescribed PCAs, increased from 37% previously. 21 patients (33%) patients referred to cardiothoracics (increased from 23%), 2 had rib fixation (3%) (decreased from 7%).

5 patients (8%) received hospital trauma call (decreased from 35%), 5 (8%) ED trauma call (increased from 7%). 55 patients received CT imaging (86%), increased from 76% previously.

Initial destination of care was CDU for 34 patients (53%), Critical Care for 15 (23%), 6 to MAU (9%), 6 to Respiratory (9%), 4 (6%) to other medical wards. Admission to surgical and orthopaedic wards has dropped to zero. 3 (out of 50) patients (6%) deteriorated on ward requiring escalation to ICU for advanced analgesia or HFNC (9% previously). 4 patients received NIV. 3 patients received HFNC. 0 patients required invasive ventilation.

Conclusions: Despite a large increase in number of patients, implementing a Chest Injury Pathway has been associated with increases in regional anaesthesia and patient-controlled analgesia, increased level of appropriate imaging, more patients getting a standardized pathway of care, less requiring secondary Critical Care admission, and decreases in mortality, particularly in pneumonia. The large decrease in patients receiving hospital trauma response requires further investigation but it is encouraging that improvements of care take place despite this. Further work is needed to improve pathway compliance in the face of competing resource demands, particularly selecting patients for rib fixation.

Efficiency in Critical Care Research: Can surplus clinical blood samples be utilised for research?

Jonathan Halcrow¹, Stephen Wright², Anthony Rostron^1,2 and Kate Musgrave^1,2

¹Institute of Cellular Medicine, Newcastle University, Newcastle Upon Tyne, UK

²Newcastle upon tyne NHS Foundation Trust, Newcastle upon Tyne, UK

Abstract

Acute respiratory distress syndrome (ARDS) accounts for the death of 1 in 10 intensive care unit (ICU) patients. Large-scale prospective studies undertaking cytokine analysis are required to validate the existence of subphenotypes within ARDS. Such studies are complex in critical care as critically ill patients often lack the capacity to consent and research blood tests may worsen anaemia commonly seen in ICU patients. Routinely, ICU patients undergo daily blood tests, with a significant surplus of blood left over after testing. Surplus patient blood is routinely stored for 72 hours then discarded.

Aim: This feasibility study aimed to investigate cytokine stability over time in surplus routinely collected blood, when stored under routine conditions.

Method: Unprocessed critically ill patient (n = 4) and healthy volunteer (n = 4) blood samples were stored under routine conditions for 72 hours. Healthy volunteer samples (n = 4) were stored at 21°C, 4–8°C and under routine conditions. Plasma concentrations of Interleukin-6, Interleukin-8 and Interferon-gamma, cytokines important in ARDS, were measured at baseline, 24, 48 and 72 hours using an ELISA.

Results: All cytokines remained stable in unprocessed blood stored under routine conditions for 72 hours. Storage temperature had no significant effect on the stability of Interleukin-6 (p = 0.1636), Interleukin-8 (p = 0.2101) or Interferon-γ (p = 0.1562) over time in unprocessed blood. Interleukin-8 concentration was significantly higher in samples stored at 21°C, compared to samples stored at 4°C (p = 0.0358) however Interleukin-8 levels remained stable overtime when stored at either 21°C or 4°C.

Conclusion: Surplus blood stored under routine conditions presents a stable and efficient source of biomarkers necessary to identify ARDS subphenotypes. Using surplus blood reduces the need for additional, potentially harmful, blood tests in future research and may present an efficient methodology for future studies seeking to validate the existence of subphenotypes within ARDS.

Risk Assessing and Managing Aggression and Violence in Mental Health Patients in a Tertiary Intensive Care Unit

Cheryl Achary¹, Andrew Wood², Lynne Barrass³, James Pennington³ and Marie Healy³

¹Dual Anaesthethesia-ICM Registrar, Royal London Hospital, London, UK

²Dual Anaesthesia-PHM Registrar, London Deanery, London, UK

³Consultant in Anaesthesia and Critical Care Medicine. Royal London Hospital., London, UK

Abstract

Violent and aggressive episodes are common in the acute medical setting, particularly in areas caring for elderly patients¹ and those with cognitive or mental health disorders.² Incidents of physical assault reported to security staff at our institution have risen by 60% between 2016 and 2017.³

In response to a critical incident involving a patient suffering from an acute psychotic episode, our tertiary critical care unit has initiated processes to equip staff in the management of violence and aggression associated with similar patients admitted to the intensive care. These include a series of practical training sessions with security staff on safe restrictive interventions, new guidelines to both risk assess and manage the spectrum of aggression in mental health patients and plans to integrate these new guidelines into the staff induction program.

Our guidelines focus on five core steps, to be exercised in parallel with the clinical critical care needed by the mentally unwell patient:

Step One. When a patient with known or suspected acute mental health illness is admitted, staff are advised to consider their potential to be violent or aggressive. Risk factors cited as predictive of violent behaviour were gathered according to current evidence base.²

Step two involves advising on a collateral history.

Step three encourages early involvement of the psychiatry team to conduct a mental health assessment and provide advice on managing both the acute mental health pathology and potential for aggression.

Step four utilises the Broset^4,5checklist and advises on preventative strategies in patients identified as moderate or high risk. This guideline details non pharmacological measures including de-escalation strategies and indications for restrictive interventions, with a checklist to highlight patient safety measures when using manual restraint manoeuvres. This step also includes a rapid tranquilisation pathway detailing a structured escalation plan when pharmacological measures are urgently needed.

Step five addresses crisis situations where despite measures violent behaviour escalates and it becomes necessary to conduct an emergency rapid sequence induction, possibly in a remote area. The standard operating procedure details two main phases; initial dosing to gain control of the situation with/without i.v. access in preparation for an RSI, then advice on RSI and post RSI elements to address. We highlight the need for cooperation between psychiatric and medical teams to coordinate an individualised care plan and avoid prolonged intubation periods.

Following such an event we highlight the importance of attentive psychosocial support for those involved.

These guidelines and training sessions were organised to address the emerging needs of our patient population, but also to empower our staff in moments of crises. By focusing on risk assessment and early preventative measures we hope to avoid violent episodes in our intensive care. Where they do still occur, we hope these guidelines provide a structured, safe and evidence based method of keeping both our patients and staff safe.

Blood conservation strategies in critical care: Another Bloody QIP

Marta Goncalves Pinto and Stephen Hutchinson

Critical Care Complex & Anaesthetics, Norfolk and Norwich University Hospital, Norwich, UK

Abstract

Anaemia is quite common in critical care patients and has a significant negative impact on their recovery.¹ The causes for anaemia in these patients are multifactorial. However, iatrogenic blood loss caused by repeated sampling, blood wastage and the resulting anaemia can be prevented.² Determining the factors that contribute to waste, implementation of new procedures and education of staff is crucial.

This improvement project attempted to reduce unnecessary blood loss due blood sampling in critical care patients. PDSA cycles included: 1) The creation of flowcharts for each step in the process of daily blood sampling. 2) Change of the serum blood tubes to plasma tubes sampling. 3) Creation of a CCC profile on ICE web. To avoid repeated blood sampling. 4) Prescribed blood sampling to allow medical staff to delete the default requests. 5) Data collection to determine the impact of a closed arterial sampling system.

All interventions contributed to a culture change within the ICU and together with a focus on transfusion restriction has resulted in a significant reduction in unnecessary transfusions. The impact of individual elements will be described. In terms of quantitative reduction, changing the serum blood sampling tubes for a smaller version we were able to reduce 2 ml of blood being taken per bottle, a 40% reduction. The blood sampling audit revealed a mean blood volume drawn per day per bed space of 53.3 ml. With the use of a closed blood sampling system this was reduced to 24.7 mls. All told, an approximate 54% reduction in blood volume was drawn per patient, bed day.

The present improvement project shows that the use of a multi strategy approach can reduce the volume of blood drawn for diagnostic testing. Standardisation of blood testing protocols regarding specific volume and frequency- the use of small volume sample bottles, reduction of any unnecessary sampling, and enhancement of pre-analytical sampling techniques to reduce repeat sampling, all contributed for the blood conservation in critical care. However, the use of a closed arterial blood sampling system was the most effective tool to reduce diagnostic blood loss.

Cardiac Intensive Care Follow Up Clinic- Lessons to be Learnt

Claire Boynton¹, Bogdan Covina² and Smitangshu MUKHERJEE²

¹Harefield Hospital, London, UK

²Imperial College Healthcare NHS Trust, London, UK

Abstract

Aim: To explore the benefits of a cardiothoracic post intensive care follow up clinic.

We focused on three main areas:

• Reduction in PTSD.

• Using data to feed back into the intensive care unit therefore improving patient experience.

• Referring patients to other services, which may have been overlooked.

Introduction: Intensive care follow up clinics are recommended in Nice guidelines on rehabilitation after critical illness. (1) Many intensive care units now have follow up clinics aiming to tackle the impact of critical illness on the patient and their family. The focus of these clinics is not only physical rehabilitation but also non-physical (mental and psychological health) (2). Cardiac surgery patients are historically looked after in a specialist intensive care environment and generally have a shorter stay in intensive care than other intensive care patients.

Our new ITU follow-up clinic has provided 3 mains areas for us to examine.

• The feedback from the patients, which can then be used to feed back into the intensive care for service improvement.

• The direct impact on the patient of their stay on ITU. The clinic provides an opportunity for patients to talk about their experience both physical and non-physical and provides an opportunity to explain and address those issues

• It allows the chance for referral to other services, eg cardiology, when this is appropriate

Method: The data from 26 patients over 6 months who had a stay of 5 days or more on cardiac intensive care. The patients who stayed under 5 days in the pilot programme, did not tend to have any impact on the their physical or non-physical health. These patents tended to be fast tracked through ITU.

A structured interview on the telephone and subsequently in clinic was used to gain information about the patient’s stay on intensive care and their subsequent problems, the data was analysed. Data was collected on how many referrals were made to other specialities.

Results: 19/26 (70 %) of patients had some up to all of their memory of their stay on ITU

23/26 (88%) of patients rated their stay “good or excellent”

10/26 (38%) felt that they had not fully recovered from their ailment. This ranged from cardiovascular, pain, mobility, weight loss or memory issues.

Suggestions for improvement of the unit included better communication and further information.

3 patients required onward referrals:

1 Patient had severe shortness of breath, required admission under cardiology for DCCV.

1 Patient was admitted under cardiothoracics and subsequently required a stent.

1 Patient with severe shortness of breath, was detected on the telephone interview and referred to cardiology.

Conclusions: The use of an intensive care follow up clinic provided a service which was of benefit for patients’ physical and mental health. In addition there is an opportunity for the patients to provide feedback to the service which can be used for quality improvement of the intensive care unit. Further development of the service to meet the needs of the patients is ongoing and will continue to benefit those patients.

Surviving Sepsis Beyond Intensive Care: Survival rates at one decade

Reema Patel¹, Sally Wood¹, Abigail Ford¹, Andrew Hutchinson¹, Vivienne Weston¹, Mark Simmonds¹ and Marc Chikhani²

¹Queen's Medical Centre, Nottingham University Hospitals NHS Trust, Nottingham, UK

²Anaesthesia and Critical Care, Division of Clinical Neuroscience, School of Medicine, University of Nottingham, Nottingham, UK

Abstract

The multidisciplinary Sepsis Action Group at the Queen’s Medical Centre, Nottingham University Hospitals NHS Trust, has worked persistently for over a decade to improve the care of patients with sepsis in Nottingham. Compliance with the suggested international treatment recommendations has been audited after publication of each iteration of the Surviving Sepsis Campaign guidelines.¹ Early engagement and data collection now provides the opportunity for commentary on the long term survival of our cohort.

Institutional ethical waiver was granted by the Caldicott Guardian. Initial data was collected using a bespoke and previously validated sepsis audit tool.² In 2005/6 and 2009/10 adult patients were identified with significant positive blood cultures and included if they fulfilled the pathophysiological criteria for sepsis. A more streamlined approach was taken in 2013/14, identifying patients as those admitted to critical care with the diagnosis of sepsis, prior to microbiological diagnosis. The duration of the audit was over a five month period between November and March for the first two cycles, and over a 27 month period for the third cycle, the period between November 2013 and March 2014 will be used for survival comparison. Hospital data systems were interrogated to determine duration of survival at one, five and ten years. There was no missing data.

Figure 1.

Percentage sepsis survival from time of diagnosis of severe sepsis or septic shock.

There were a total of 46 patients who met the criteria for study in 2005/6, 89 in 2009/10 and 101 in 2013/14. Demographics were similar between each group with median age of 66 years. Number of patients with full adherence to the recommended treatment items at 3 and 6 hours was 7 (15%) and 4 (9%) in 2005/6, 13 (15%) and 9 (11%) in 2009/10 and 77 (77%) and 55 (55%) in 2013/14. The summary mortality results are presented in table 1. The comparative cumulative survival in years following inclusion into the study is shown graphically in figure 1.

Despite initially reported plateau in mortality³ from 28 to 365 days, there is actually continued attrition until the end of the time constraint. The long term mortality in all three groups is in excess of 50%. This is despite the apparent improvement in compliance with timely treatment interventions. These surprising results match emerging observations of other cohorts^4,5 and highlight the paucity of long term critical care outcome data and the importance of developing post critical care rehabilitation programs and robust outcome registries to objectively measure the survival benefit of interventions.

Table 1.

Summary mortality data.

Time period	Total number of patients	Death at 7 days [n (%)]	Death at 28 days [n (%)]	Death at 1 year [n (%)]	Death at 5 years [n (%)]	Death at 10 years [n (%)]
2005/6	46	7 (15)	12 (26)	18 (39)	20 (43)	27 (59)
2009/10	89	7 (8)	11 (12)	28 (31)	39 (44)	47 (53) to date
2013/14	101	17 (17)	27 (27)	36 (36)	51 (51) to date	51 (51) to date

Evaluation of reusable and disposable bronchoscopes by user preference when performing simulated bronco-alveolar lavage on a manikin

Elizabeth Whetton¹, Andre Simons² and Brendan McGrath¹

¹Manchester University NHS Foundation Trust, Manchester, UK

²University of Manchester, Manchester, UK

Abstract

Single-use bronchoscopes have evolved in recent years such that their suction capabilities allow the ability to adequately perform broncho-alveolar lavage (BAL) and bronchial wash (BW). Their use may also be more cost effective compared to reusable bronchoscopes in intensive care units (ICUs).^1,2 As a result, procedures using single-use scopes are becoming more commonplace. The aim of this study was to evaluate the user preference metrics and feedback of different reusable and single-use bronchoscopes when performing simulated BALs on manikins.

Thirty participants were asked to perform a BAL on a pre-intubated manikin using the Ambu aScope3 slim/regular/large and reusable Olympus BF bronchoscopes. Data were collected on user preference metrics for each bronchoscope. Participants were asked to indicate their agreement with the six statements detailed in Table 1 with a score from 1 (strongly disagree) through 4 (neutral) to 7 (strongly agree) and unstructured feedback on each device.

The results are demonstrated in Figure 1 below:

The bronchoscope with the lowest mean score was the Olympus BF-P40. Users commented on the ergonomics of the device specifically reporting difficulty attaching the syringe, instilling and aspirating saline whilst holding the scope. The single-use scopes all scored well on ergonomics with comments referring to their lightweight nature. The lowest median score for a single statement was for the aScope3 Slim relating to the device’s suction capability. Conversely, the Ambu aScope3 Large consistently scored 7 making it the highest scoring scope for that criteria. The image quality of the Olympus BF-260 scope was rated the highest. However, participants felt that image quality did not impact of their ability to perform the procedure successfully. Other than this overall the scores were comparable between bronchoscopes.

Figure 1.

Median user preference scores for statements 1–6 for each bronchoscope.

The results indicated that the participants found the single-use scopes comparable to reusable scopes for performing BALs with the exception of the aScope3 Slim, which has a smaller diameter working channel. By comparison the aScope Large has a larger diameter channel with improved suction capabilities and may be the most suitable aScope for performing BAL. Our study was limited by its manikin design, but these data indicate that bronchoscope ergonomics and suction capabilities are more important factors than image quality for users performing simulated BAL.

Table 1.

Six statements for assessment of each bronchoscope.

1. Easy and intuitive to use suction system for BAL

2. Ease of use to attach syringe and sample container

3. Suction capability was adequate for performing BAL

4. Easy navigation to target lobe

5. Ergonomics of device was satisfactory

6. Image quality was clear and adequate to verify accurate position in target lobe and completion of BAL

Too many bleeps & not enough beds: Audit of referrals to critical care 2018

Olivia Chu¹, Carl Groves¹, Richard Field¹, Kristina Ooi¹, Maya Sussman¹, Methini Himayakanthan¹, Harvinder Mann¹, Faiza Buhari¹ and Nitin Arora²

¹Heartlands Hospital, Birmingham, UK

²Consultant. University Hospitals Birmingham. Heartlands., Birmingham, UK

Abstract

Critical care units are often highly demanded with scarce availability, providing for a wide heterogeneity of patients. With increasing complexity of patient case-mix and advance in medical technology, demand for ICU beds is expected to continue to rise. Freedom of Information Act inquiry found local Critical Care bed occupancy reached 100% at the end of 2017. ICU referrals are made to a single registrar who has to manage multiple referrals with a limited ICU capacity, especially out-of-hours. Our aim was to assess the quality of the ICU referral and workload.

Data was collected from 3rd January to 7th February 2018. This was based on the Google Forms platform allowing the clinician access via smartphone. This was based on referral guidance set out by NICE & Intensive Care Society.

80 responses received from Heartlands Hospital. The majority of patients were over 55 years old, of which 28% were aged 55–73. 62.5% of patients were male. All specialities were represented in the referral data, the Emergency Department making the most referrals (35%). Referrals were most commonly made by Registrars and Consultants (45.7% & 9.9% referrals respectively). The Critical Care Outreach Team & Registrars gave clearer information in referrals. 47% had not been discussed with responsible speciality Consultant; 27% were described as ‘situation not applicable’.

69.2% of referrals to Critical Care had at least 2 acute systems issues with majority having a baseline of co-morbid condition. 79% referred did not have a ceiling of treatment defined prior to or at point of referral. A ceiling of treatment was present when the patient has been discussed with a speciality Consultant prior to referral. 70% of referrals were deemed as ‘appropriate’. Concerns raised by ICU clinicians regarding referrals focused on lack of treatment initiation, poor illness recognition and lack of senior speciality involvement. Capacity limitation within unit was also highlighted. Ultimately, 54% of referred patients were admitted into ICU with 18% remaining under CCOT care.

The ICU receives a heterogeneous patient group with admission APACHE scores above that of similar centres. Critical care workload continues to increase but has still been able to accommodate as per clinical picture rather than restriction of capacity. Ceiling of treatment lacking in many cases with concerns raised regarding patient management on the wards. Further education & empowering of ward staff may be required to improve the situation.

Recommendations:

• All patients should be reviewed by at least parent team Speciality Registrar prior to referral to Critical Care.

• All referrals should be discussed with parent team Consultant prior to referral to Critical Care

• All patients accepted to Critical Care need to be reviewed by parent team Consultant within 24 hours of admission.

Rare neurological complications of Infective Endocarditis and clinical dilemma; a case report

Muhammad Ali Chaudhary, Tamas Esze, Adnan Akram Bhatti and Ferenc Kovari

North Middlesex University Hospital, London, UK

Abstract

A 51 years old gentleman with history of lichen planus was admitted in ITU with loss of consciousness and seizures at home. He was having fever and had audible murmurs, which raised the suspicion of infective endocarditis and septic embolism. His GCS was 7/15, CT head showed subarachnoid hemorrhage in cerebral hemispheres and parenchymal bleed in right parietal lobe. His blood results showed WCC of 25 and CRP of 356. Trans-esophageal echocardiogram (TOE) showed large vegetation in left ventricle, aortic regurgitation and dilated aortic root. Blood cultures isolated Staphylococcus aureus and its origin was thought to be previously active lichen planus which further lead to vegetations. This presentation is quite rare, as infective endocarditis and primary dermatologic disorder like lichen planus rarely present as a cause of ITU admission. He was commenced on antibiotics, improved and was extubated. Repeat CT head showed no new findings. However, we had a fear that without the removal of the septic and embolic source from his heart, he might face a bad outcome. We faced a difficult dilemma – should we encourage the surgery to reduce the risk of another – potentially fatal – stroke, or we should be precautious as there is an active bleeding in the brain which might deteriorate during surgery? Unfortunately his GCS dropped suddenly to 6/15 resulting in re-intubation and repeat CT head showed new subarachnoid hemorrhage of left anterior frontal lobe. Later on developed a renal failure due to excessive rhabdomyolysis, required dialysis, and absconded on full support 13 days after admission. There are many pathological mechanisms proposed for neurological complications such as septic emboli induced haemorrhagic transformation and pyogenic arteritis occurring due to septic emboli causing arterial wall necrosis [1]. Furthermore presence of highly virulent organisms in skin flora should not be overlooked and in one study it was found increased prevalence of infective endocarditis in atopic dermatitis [2]. Infective endocarditis (IE) is catastrophic once complicated; with IE alone having mortality rates reaching 20% to 40% [3]. Neurological complications occur in 20% to 40% of cases [4]. Non-aneurysmal sub arachnoid hemorrhage due to infective endocarditis is a very rare presenting feature having incidence of 1% only [5]. These complications are commonly associated with staphylococcus aureus endocarditis [6]. It has been cited that relatively younger patients with IE with or without large vegetation and raised inflammatory markers are at increased risk of major embolic complications during their inpatient course [7]. Management is to prevent these complications by using early antibiotic therapy [8]. Antimicrobial treatment has been shown to reduce the risk of neurological complications by 33% to 75%. Cases that had hemorrhagic transformation requiring surgical intervention within four weeks of the event showed higher mortality (75%) as compared to the group of patients who had delayed operation (40%) [9]. However some findings suggest earlier surgical intervention has better outcome and risk of post-operative neurological deterioration is low [10]. Therefore timing of surgery is grey area without specific answer and it needs further investigation.

Improving quality in critically unwell inter-hospital transfers – a region wide trainee survey

Tom Syratt¹, Owen Chambers¹ and Arjun Ardeshana²

¹Cheshire and Mersey Critical Care Network, Liverpool, UK

²Regional Fellow, Countess of Chester Hospital, Chester, UK

Abstract

Whilst Mersey is a geographically small region, many of our tertiary centres are on individual sites. This leads to frequent emergency transfer of critically ill patients for specialist care. Despite the size of the region, internal audit suggested the median time taken for transfer to our tertiary neuro-surgical unit from base hospitals to be 320 minutes. Transfers are undertaken by various grades of trainee anaesthetists therefore we aimed to assess their confidence in conducting these transfers. We have then suggested a strategy to reduce the time and improve the quality of time critical transfers, specifically transfers for primary percutaneous coronary intervention (PPCI) for patients requiring emergency neuro-surgical care.

We surveyed all anaesthetic trainees within the Mersey deanery and received 42 responses across all grades (CT1-ST5+), all of whom had received some form of transfer training in the past. Experience amongst core trainees varied widely with only 16% having performed ³5 transfers, rising to 80% by ST3-4 and 100% by the time trainees reached ST5+. When rating their confidence (0 being no confidence, 10 being completely confident) in performing these transfers, there was a significant increase as trainees became more experienced, rising from an average of 5.3/10 amongst core trainees to 8.2/10 in those ST5 and above. Core trainees expressed low confidence (5.2/10) in the initial management of patients requiring ventilation for ST elevation myocardial infarction, subarachnoid haemorrhage or traumatic brain injury. When asked what concerned the trainees the most about performing transfers, 71% stated being involved in a critical incident during the transfer and 50% cited the receiving team being critical as a concern. When asked, 94% of trainees felt a checklist with written guidance would be helpful in transferring these groups of patients.

Through collaboration with Mersey’s tertiary cardiothoracic centre and neurosurgical centre, we have produced checklists with guidance, key points and important clinical parameters on the emergency care and transfer of ventilated PPCI patients and those requiring neurosurgical intervention. We hope that with these checklists, trainees will be more confident and competent in transferring these patients, which will be reflected in improved transfer times and quality of care in these critically unwell patients.

Interstitial lung disease outcomes in the intensive care unit

James Schneider¹, Stuart Williams² and Owen Dempsey³

¹Critical Care Department, University College London Hospital, London, UK

²Intensive Care Unit, Aberdeen Royal Infirmary, Aberdeen, UK

³Respiratory Department, Aberdeen Royal Infirmary, Aberdeen, UK

Abstract

Background: Interstitial lung diseases (ILD) are a diverse group of disorders ultimately resulting in impaired gas exchange, as a consequence of thickening of the pulmonary interstitium. Critical illness mandating mechanical ventilation can be very challenging in these patients due to impaired alveolar compliance, increased airway resistance, pulmonary hypertension, and poor oxygen diffusion capacity. Often patients are diagnosed in the later years of life, which contributes to further reduced physiological reserve.

Aims and Methods: This study aimed to evaluate the incidence and outcomes of patients who presented to the Intensive Care Unit (ICU) with an established diagnosis of ILD. The WardWatcher (SICSAG) Database was used to select all patients admitted to the Intensive Care Unit in Aberdeen Royal Infirmary between July 1999 and April 2018. Descriptive analysis was performed for study population demographics, ILD subtype, APACHE III diagnosis, and clinical outcomes.

Results: The study population comprised 119 patients with proven ILD (males n = 79, 66.4%), with a median age of 70 (IQR 60–76) years old. This included a total of 10 (8.4%) readmissions during the study period. The median duration of ICU stay was 3.6 (IQR 1.3–9.0) days. The mean APACHE II Score was 25.4 (std ± 7.4), which equated to an estimated predicted mortality rate of 48.7% (std ± 25). The variety of ILD subtypes which presented to the ICU can be found in Table 1. Regarding ventilatory support, the majority of patients underwent endotracheal intubation, n = 83 (69.7%) with 3 (3.6%) re-intubations, encompassing a median duration of 4 (IQR 2–7) days. A total of 25 patients (21%) received a tracheostomy during their stay. The ICU mortality rate was 60.5%, with an overall hospital mortality of 76.5%.

Conclusion: Critically unwell patients with underlying ILD are an elderly population, who have very poor clinical outcomes following admission to the ICU. The majority of patients require mechanical ventilation due to acute respiratory pathology, and have a poor prognosis following this, with respect to readmissions, tracheostomy, and mortality. Careful assessment and informed discussion of the risks involved should be offered to patients with ILD in the critical care sector prior to ICU transfer.

Table 1.

Interstitial Lung Disease Subtypes.

Subtype of Interstitial Lung Disease, n = 109	Frequency, n (Proportion %)
Idiopathic Pulmonary Fibrosis	48 (44.0%)
Unclassifiable ILD	16 (14.7%)
Environmental Exposures	12 (11.0%)
Rheumatoid	12 (11.0%)
Drug-induced	8 (7.3%)
Other	4 (3.7%)
Sarcoidosis	4 (3.7%)
Other Autoimmune	2 (1.8%)
Vasculitis	2 (1.8%)
Infective	1 (0.9%)

An audit on bowel management in a tertiary center critical care unit

Geoffrey Hong Kiat Yong and Huw McConnell

Newcastle upon tyne NHS Foundation Trust, Newcastle upon Tyne, UK

Abstract

Background: Our impression is that there has been a haphazard approach towards bowel care within the unit, leading to a large prevalence of gut dysfunction, which has a negative impact on patient outcomes.

Objectives: To measure the rates of loose stools and constipation within the unit, and to assess compliance with the local bowel management protocol.

Methods: The audit was conducted across 6 weeks, from 23rd April to 4th June 2018 in the critical care unit of the Royal Victoria Infirmary, Newcastle upon Tyne. A total of 376 patient days were observed.

To measure the rate of loose stools and constipation, we have created stickers containing a simple questionnaire on the patient's current bowel status, which were pasted on each daily monitoring chart. A standardised proforma was then used to collect data from the completed stickers each day.

We also captured a snapshot of the compliance with the local protocol over the course of 1 week. The sample size was 26 patients. Our protocol recommends that all patients should be commenced on 15 mg Senna nocte on admission unless contraindicated, and any patients who have not opened their bowels for 2 or more days will require a per rectal(PR) exam to guide further management.

Results: 20.74% of patient days were spent with loose stools, while 31.65% of patient days were spent constipated. In comparison to the previous audit conducted in 2015, constipation rates remained about the same(29% vs 31.65%), while there was a significant drop in loose stool rates(43% vs 20.74%).

The audit showed poor compliance with the local guidelines. 13 out of 26(50%) patients admitted within that week were not commenced on any form of aperients on admission. 14 out of 26 patients did not open their bowels for 2 or more days within that week, of which only 6(43%) had a PR exam conducted.

Conclusions: This audit confirms our initial impression of poor bowel care within the unit. The drop in loose stool rates is unlikely to be an effect of the local protocol due to poor compliance rates; and may either be due to chance or ascribed to the larger sample size in this audit. Our aim is to reduce the rates of loose stools and constipation further. There has been evidence showing that a systematic bowel management protocol can reduce these rates(1,2).

In view of these, we are currently working on a refined bowel management protocol, which is easier to follow and accessible to all staff. However, it is difficult to tell if the protocol is effective without achieving a good compliance rate. Hence, we have presented our findings to all senior clinical staff in the department, and conducted bedside teachings around the unit to the nursing staff, in order to raise awareness on this issue and improve compliance rates. We aim to re-audit at a later date to measure the effect of the new protocol on the prevalence of loose stools and constipation.

A rare case of extensive acute oesophageal necrosis or black oesophagus

Adnan Akram Bhatti, Muhammad Ali Chaudhary and Muhammad Esakji

North Middlesex University Hospital, London, UK

Abstract

Introduction: Acute oesophageal necrosis or black oesophagus is a rare condition with a very low prevalence of less than 0.5 percent [1] and only a handful of cases have been reported in literature. It has an unknown aetiology but it is hypothesized that it occurs due to hypovolemia leading to vascular ischemia of the lower oesophagus resulting in increased vulnerability to refluxed pepsin and superficial injury which ultimately causes oesophageal necrosis. The most common presentation is upper gastrointestinal bleeding, melena and anaemia [3]. We present a case of a woman with oesophageal necrosis secondary to hypovolemia who developed sepsis and multi-organ failure.

Case presentation: A 47-years-old female with a past medical history of alcohol dependence and Wernicke-Korsakoff syndrome presented with hematemesis, malaena and per-vaginal bleeding. She was anaemic and developed hypotension with rising lactate. She was managed as a case of upper gastrointestinal bleeding and sepsis, with fluid resuscitation, antibiotics and subsequently underwent oesophago-gastro-duodenoscopy (OGD) which showed inflammation and superficial ulceration of upper oesophagus and blackening of distal oesophagus consistent with oesophageal necrosis but no active bleeding. Her condition deteriorated acutely, and she developed multi-organ failure requiring mechanical ventilation, inotropic support and renal replacement therapy. Despite provision of maximal organ support, the patient continued to deteriorate and passed away.

Discussion: Acute oesophageal necrosis or black oesophagus has multiple aetiologies and risk factors. It usually occurs in patients with shock, thromboembolism, gastric outlet obstruction, malnutrition, alcohol intoxication, peptic ulcer disease and hiatus hernia. It has very characteristic features on OGD which shows a black distal oesophagus with ulceration with or without proximal extension [3]. Histopathological features include mucosal ulceration, necrotic tissue and hemosiderin deposits. Many pathophysiological mechanisms have been proposed. The most acceptable ones attribute ischemia as the cause of injury which occurs due to hypovolemic states. Other explanations include thrombosis of celiac trunk or dissection of the thoracic aorta inducing ischemia. Moreover, decreased perfusion pressure and prolonged exposure to gastric juices due to gastric outlet obstruction or volvulus are other proposed mechanisms of injury [1]. Management involves antibiotics, bowel rest and high dose proton pump inhibitors. Due to the severity of the disease, threshold for surgical intervention is low and indicated only if there is perforation, mediastinitis or full thickness necrosis [4]. The prognosis of acute oesophageal necrosis is poor and partially depends on comorbidities. Despite aggressive management, it has a high mortality rate of around 35% [2].

Usage patterns of total emergency patients and critical emergency patients: National study of emergency patient flow

Sung Min Lee

Chonnam National University Hospital, Gwangju, The Republic Of Korea

Abstract

Background: This study uses the National Emergency Department Information System (NEDIS) data to analyze the flow of emergency and critical emergency patients and to identify the patterns of emergency medical service usage in Korea.

Methods: The relevance index (RI) and commitment index (CI) were calculated from the 2016 NEDIS data. In this study, the number of clusters was determined using NbClust, and cluster analysis was used to analyze the usage patterns of emergency and critical emergency patients.

Results: The RI and CI were calculated using 8,389,766 cases of 214 districts. The results of the RI and CI suggested that there were three types of clusters among the emergency patients. In Cluster 1, 54 districts (25.2%) had low RI and high CI, and it was of outflow type. Cluster 2 was categorized as the influx-type in 58 districts (27.1%) irrespective of RI and low CI. Cluster 3 was categorized as the self-sufficient type found in 102 districts (47.7%), with high RI and high CI. The cluster analysis of the critical emergency patients was divided into two types. Cluster 1 was categorized as outflow type with high CI found in 129 districts (60.3%), while Cluster 2 was categorized as inflow type with low CI found in 85 districts (39.7%).

Conclusions: This study elucidated the regional status of usage patterns of emergency and critical emergency patients in Korea. This study can serve as a basis for the establishment and selection of emergency medical service areas and vulnerable emergency medical service areas.

Outcomes of high-flow nasal oxygen therapy in adults with acute hypoxaemic non-hypercapnic respiratory failure and a ward-based ceiling of care

Abigail Lau, Zoe Taylor, Joanne Coleman and Peter Featherstone

John V Farman Intensive Care Unit, Addenbrooke’s Hospital, Cambridge, UK

Abstract

High-flow nasal oxygen (HFNO) therapy is commonly used in acute hypoxaemic non-hypercapnic respiratory failure.¹ Treatment can be commenced in ward areas, and is generally well tolerated by patients, allowing it to be continued for several days. However, not all patients improve with HFNO, and a previous service evaluation undertaken in our hospital (a 1100-bed tertiary centre with all specialties except cardiothoracics) demonstrated a high 28-day mortality (71.4%) in a subgroup of patients who had treatment limits in place at the initiation of therapy.² We therefore sought to review the use and outcomes of HFNO in adults with acute hypoxemic non-hypercapnic respiratory failure and a ward-based ceiling of care.

This service evaluation was registered and approved by our institution. Using retrospective electronic case note review, we identified all patients over 18-years of age, with a ward-based ceiling of care, who were commenced on HFNO by the Addenbrooke’s Rapid Response Team between September 2017 and June 2018.

A total of 50 patients (mean age 77 years; age range 45–97 years) were identified. All had significant comorbidities (e.g. advanced malignancies and severe cardiac, neurological and respiratory pathologies), which had guided decision-making regarding treatment limits. Pneumonia and cardiogenic pulmonary oedema were the most common precipitants of acute hypoxaemic respiratory failure. HFNO was initiated between the hours of 18.00 and 07.59 in 66% of cases, and patients received treatment for a median of 2.5 days (range 1–18 days). In-hospital, and 28-day mortality rates were 52% and 62% respectively. Mortality was higher in patients aged ≥71years compared with those aged 45–70 years, and was further increased in those with 3 or more chronic comorbidities. Ten (20%) patients died while receiving HFNO. Of the 25 patients who survived to hospital discharge, 14 were discharged home, or home with a care package, and 11 were transferred to a hospice, nursing home or other secondary care centre.

Our results confirm that patients in this cohort have an extremely high mortality. HFNO is often initiated out of daytime hours, and continued for a protracted period. While HFNO provides symptomatic relief of dysponea in terminally ill patients¹, it is also likely that therapy prolongs the dying process. Patients may benefit from closer working relationships between critical care outreach, and palliative care teams.

Mitral regurgitation saved with percutaneous mechanical circulatory support: the role of the heart teams

Andrew Harper, Francisca Caetano, Brijesh Patel, Anthony De Souza, Rob Smith, Richard Trimlett and Susanna Price

Royal Brompton and Harefield NHS Foundation Trust, London, UK

Abstract

Introduction: Severe mitral regurgitation (MR) associated with cardiogenic shock (CS) and pulmonary oedema is associated with high mortality. Historically, intra-aortic balloon counterpulsation (IABP) has been used to support the circulation prior to cardiac surgery, but in very severe MR the level of support may be insufficient to allow unloading of the left ventricle and provide adequate cardiac output. This case describes the use of percutaneous mechanical circulatory support with Impella to facilitate successful mitral valve surgery in a patient with CS.

Clinical case: a 66-year-old man presented with a past medical history of hyperlipidaemia and gout presented to his local hospital with severe dysponea which rapidly deteriorated requiring intubation and ventilation. He had clinical and biochemical features of CS, requiring boluses of aramine and adrenaline to maintain cardiac otuput. Echocardiography showed evidence of severe MR due to ruptured chordae and a flail P2.

He underwent coronary angiography (unobstructed coronary arteries) and insertion of intra-aortic balloon pump (IABP) inserted locally before being transferred for institution of more advanced mechanical circulatory support (MCS). Following Impella® 2.5 insertion he was admitted to the intensive care unit for ongoing supportive care with a view to improving organ perfusion and resolving pulmonary oedema prior to surgical/transcatheter intervention. During this time, he was supported additionally with noradrenaline, adrenaline, dobutamine and haemofiltration.

He was discussed by the shock team as well as the mitral multi-disciplinary team (MDT), and as he improved significantly over the first 48 hours (complete resolution of pulmonary oedema, minimal inotrope requirement, normalisation of liver function) he underwent successful surgical repair 3 days after initial presentation, requiring no MCS post-operatively. He was successfully extubated 12 days following admission to intensive care, and transferred to the high dependency unit (HDU) whilst awaiting repatriation to his local hospital. He returned to outpatient clinic approximately 2 months after his surgery. He remained asymptomatic following discharge and no longer requires renal replacement therapy.

Discussion: Current guidelines relating to CS recommend initial support with standard inotropic and ventilatory interventions, and early implementation of MCS, whilst intervening to reverse the underlying cause (1). Historically IABP alone has been used in patients with severe MR, but mortality remains high, usually related to the degree of unresolved pulmonary oedema and multi-organ dysfunction around the time of surgery. A number of devices that provide a higher level of ventricular offloading as well as circulatory support exist (including Impella, Tandem Heart, VA-ECMO) each of which is associated with a different haemodynamic and complication profile. Whilst more effective in terms of haemodynamics, none has yet shown superior outcomes in terms of survival (2). In parallel, a number of percutaneous interventions for MR exist, however, recommendations for their use in CS are limited (1,3). In such high-risk cases, multidisciplinary discussions with the relevant teams (in this case shock and complex mitral) are essential for successful outcomes.

Size Matters; Can we improve intra-operative ventilation?

Lavanya Raman and Munita Grover

Northwick Park Hospital, London, UK

Abstract

Introduction: Lung protective ventilation (LPV) using a tidal volume (VT) of 6–8 mL/Kg ideal body weight (IBW) is recommended in the intensive care unit and theatres to reduce the incidence of pulmonary complications. The aim of this quality improvement project was to assess the extent to which LPV is used in theatres in a busy district general hospital and to implement measures to promote adherence to the recommendations.

Methods: Anaesthetists completed questionnaires for all patients undergoing general anaesthesia at Northwick Park Hospital over 1 week. Demographics, actual body weight (ABW), height, American Society of Anesthesiologists (ASA) score, and procedural information were recorded. Ventilatory parameters included the ventilation mode, VT, and positive end expiratory pressure (PEEP). The body mass index (BMI), IBW and VT (expressed in mL/Kg of ABW and IBW) were calculated for each patient. A Mann Whitney U test was used to compare IBW and PBW and a Chi squared test was used to identify an association between VT and other variables. The results were presented at a clinical governance meeting and aide memoirs in the form of IBW and tidal volume charts were constructed and displayed in each theatre. Anaesthetic staff received education and advice regarding lung protective ventilation. Following this, the audit was repeated to ascertain whether the intervention had led to a change in practice.

Results: The initial audit included 129 patients; 65 males and 64 females. Mean age was 51. 73 patients were overweight (BMI ≥ 25). 88% patients received PEEP. IBW was calculated in 106 patients and was significantly lower than ABW (61Kg [54–71] vs 72Kg [62–85] p < 0.05). VT was higher when calculated from IBW than ABW (8.7 mL/Kg [7.1–9.3] vs 7.5 mL/Kg [5.8–7.9] p < 0.05). 52 patients (49%) received LPV with VT of <8 mL/Kg IBW in accordance with the recommendations. Significantly more females (75%) received VT ≥8 ml/kg than males (29%) (p < 0.01). VT was independent of age, ASA, BMI, ventilation mode, speciality, and patient position. The re-audit following the intervention included 96 patients; 47 males, 49 females. Mean age 64. 58 patients were overweight. 91% received PEEP. IBW was again significantly lower than ABW (65Kg vs 78Kg p < 0.05) and VT was significantly higher when calculated from IBW than ABW (7.9 mL/Kg vs 7.2 mL/Kg p < 0.05). 64 patients (67%) received LPV with VT of <8 mL/Kg IBW. This was significantly higher than in the initial audit (p < 0.05). The mean VT (Ml/kG) was significantly lower than in the initial audit (p < 0.05).

Conclusion: Over half of the patients received VT ≥8 ml/kg IBW in the initial audit. Females were more likely to be over ventilated. Following staff education and the provision of VT/IBW charts the proportion of patients being ventilated using LPV improved, however 43% of patients were still receiving VT’s above the recommended volume. Females were still more likely to be over ventilated. A likely contributing factor is the disparity between ABW and IBW in this cohort. Ongoing and repeated education about LPV, particularly targeting new doctors, perhaps in the form of online modules is necessary to promote awareness of and adherence to the recommendations.

Compartment Syndrome following Di-Nitrophenol (Slimming Drug) Overdose

Lucy Barnes, Kushmandinie Goonetilleke, Rachel Green, Caroline Moss and Michael Margarson

St Richard's Hospital, Chichester, UK

Abstract

Dinitro-Phenol (DNP) was used in the production of explosives during and after the first world war but noted to be extremely toxic, the mechanism later discovered to be triggering of malignant pyrexia through direct uncoupling of oxidative phosphorylation.¹ Today this chemical has found an underground role as a slimming pill, modest doses increasing basal metabolic rate and reducing body fat. However inadvertent or deliberate overdose can lead to a rapid hyperpyrexia with severe rhabdomyolysis, associated with a high early mortality rate^2,3.

A 21 year-old previously well male presented to our emergency department in an agitated and febrile state. He revealed that he had ingested 16 tablets (3200mg) of DNP. Initial management consisted of intravenous fluid resuscitation, paracetamol and diazepam. He became increasingly unwell, and over two hours his temperature rose to 38.6^oC. Urgent telephone advice was obtained from NPIS⁴, highlighting the severity of this large overdose.

Around three hours after initial presentation the patient was admitted to critical care and intubated for his escalating agitation. An infusion of muscle relaxant was initiated and minute volume ventilation in excess of 25l/min was required to keep PaCO2 below 7KPa.

Aggressive active cooling was started to address his temperature which had now reached 39.4^oC. All available cooling options were used including continuous veno-venous haemofiltration (CVVHF) via ice-chilled intravenous lines, oro-gastric and bladder lavage with chilled fluids, administration of cold intravenous fluids and application of ice packs. The environment was cooled with fans and windows opened to entrain cold winter air. All hospital ice stores were rapidly exhausted and more ice obtained from a nearby supermarket. The process was highly staff intensive.

Despite aggressive cooling a peak temperature of 40.5^oC was recorded at 5 hours (Fig 1). In line with NPIS advice, dantrolene (2 mg/kg) was administered on admission and repeated six hours later.

Initial creatine kinase (CK) levels were >4000 at 5 hours, but reached a delayed peak of >1.9 million at 40 hours after admission (Fig2). CVVHF was continued for 4 days.

On extubation the patient described extreme calf pain and tenderness on palpation. Orthopaedic review confirmed bilateral compartment syndrome. Fasciotomies were performed on day 6, but all muscle of the anterior compartments was non-viable and required surgical debridement. The patient was subsequently discharged to the ward and eventually home, but with significant loss of function due to bilateral foot-drop. He awaits surgical tendon reconstruction.

This case demonstrates the extremely aggressive nature of DNP hyperpyrexia. Key learning points are that that strategies for advanced cooling should be identified in anticipation of cases of malignant hyperpyrexia⁵, including a knowledge of the availability and location of large quantity sources of ice. The massive secondary surge in CK, coming 36 hours after the peak of temperature presumably reflected a late-developing compartment syndrome. This has not been previously reported in this setting, but should be suspected and actively sought in cases of such severe hyperpyrexia and a secondary rise in CK.

Fig. 1.

Patient temperature vs Time since presentation.

Fig. 2.

Creatine Kinase vs Time since presentation.

Hypernatraemia in critical care patients on very high protein enteral feed: results of a retrospective study

Sarah Gavin

Queen Alexandra Hospital, Portsmouth, UK

Abstract

Critical care patients receive large amounts of non-nutritional calories from lipid-based sedative medications, and glucose containing intravenous fluids. In order to meet protein needs without overfeeding calories, very high protein enteral feeds, with a low ratio of non-nitrogen energy to nitrogen (NNE:N) are required. One such product has recently been developed, however anecdotal evidence suggests it may increase the incidence of hypernatraemia. Research suggests that the incidence of critical care acquired hypernatraemia is 26%.¹

The study aim is to determine whether the incidence of hypernatraemia is greater for patients who were fed with a very high protein feed, than in patients receiving the standard high protein feed.

Retrospective data was collected in the Department of Critical Care, Queen Alexandra Hospital from 1^st January to April 2018. Patients were grouped as those receiving: very high protein feed (n = 28; NNE:N = 52:1), and standard high protein feed (n = 28; NNE:N = 100:1), respectively. Patients were included if they received the feed for >3 days, and were fully fed without other supplementary feeding products. Hypernatraemia episodes were defined as serum sodium of >145 mmol/L and reported on the electronic CIS (Computer Information System). Data were collected from electronic patient medical notes, and included: total days on feed, episodes and day of hypernatraemia, and occurrence of a negative fluid balance. Data are expressed as mean ± SD (range).

Patients receiving very high protein spent 8.4 ± 5.2 days (3 to 26 days) on feed; patients receiving high protein spent 7.4 ± 4.4 days (3 to 20 days) on feed. Patients with episodes of hypernatraemia were 17 (61%; mean occurrence = day 4) for very high protein, and 12 (43%; mean occurrence = day 5) for standard high protein feed, respectively. Of these hypernatraemic patients, 76% had episodes of negative fluid balance with very high protein feed; 67% had episodes of negative fluid balance with standard high protein feed.

Critical care patients receiving very high protein feed had a higher incidence of hypernatraemia, when compared to patients on the standard high protein feed. These observations from a single unit suggest that patients receiving very high protein feed are at greater risk of developing hypernatraemia. Negative fluid balance combined with a very high protein feed, may be a contributing factor. Sodium concentration of the very high protein feed (7.5 mmol/100 ml), when compared to the high protein standard feed (4.1 mmol/100 ml) must also be considered. Further research is required to determine the relationship between protein feed concentration and the hypernatraemia in critical care. Non-critical care acquired hypernatraemia was not differentiated in this patient group, which may account for the higher than expected incidence of hypernatraemia across our cohort.

Evaluation of reactive HLH-like sepsis in an ICU: a red herring?

Judit Sallai¹, Jeronimo Moreno Cuesta² and Jonathan Ainsworth²

¹Semmelweis University Dept. of Anaesthesiology and Intensive Therapy, Budapest, Hungary

²North Middlesex University Hospital, London, UK

Abstract

Background: Sepsis is an infection causing organ dysfunction in which inflammation is dysregulated (1) Hemophagocytosis occurs when macrophages engulf viable bloods cells. Analysis post-hoc of a randomised controlled trial showed 43 (5%) of 763 patients had DIC and hepatobiliary dysfunction (HBD) – markers of severity of Haemophagocytic Lymphohistiocytosis (HLH) but not severe sepsis, necessarily – among whom treatment with Anakinra (an interleukin-1 receptor antagonist) reduced mortality (2). Adult HLH is usually a secondary disorder. Its probability was defined by Fardet et al based on 9 criteria based on 13 parameters (Hscore) (3). Here, we wanted to assess the feasibility of hemophagocytosis research in the ICU using an eligibility algorithm comprising Hscore parameters in order to identify “reactive” (non-secondary) HLH-like sepsis, which may be an excessive pro-inflammatory sepsis.

Methods: We assessed Hscore parameters available to us (6 of 13) at presentation to ICU of patients meeting sepsis-2 criteria (4) (April 2015- April 2016), and evaluated the proportion who had reactive HLH-like sepsis according to our definition. (Not available were fibrinogen, triglycerides, ferritin, hepatomegaly, splenomegaly, hemophagocytosis on bone marrow aspirate; not used was known immunosuppression.) We compared the prevalence of perturbations in further parameters, and accordingly modified Hscore parameter thresholds for WBC, liver tests, and temperature. Final parameters were Haemoglobin ≤9.2 g/dl, Platelets ≤110 x10⁹/L, White Blood count (WBC) ≤8000 x10⁹/L, Serum ALT 1.5 times ULN (>45 IU/L women >75 men) or AST > 1.5 times ULN (60 IU/L men or women), and Temperature ≥38°C. At least 3 of these latter was HLH-like sepsis. Of note, an Hscore criteria (bicytopenia) was not used, since thrombocytopenia alone is a marker of inflammatory sepsis (5).

Results: We found that 147 (18%) of 800 admissions had sepsis: 48 were excluded due to suspected or proven cancer, lymphoproliferative disease, autoimmune disease, sickle cell crisis, massive blood transfusion, or chronic liver disease. Of the remaining 99, 12 (12%) had reactive HLH-like sepsis.

Conclusion: It is feasible to identify sepsis patients for research using an eligibility algorithm based on Hscore parameters.

National Air/Gas Embolism Guideline: Developed on request of the Joint Standards Committee of FICM/ICM

Pieter Bothma

James Paget University Hospital, Great Yarmouth, UK

Abstract

Dr P Bothma, Consultant Anaesthetist, James Paget University Hospital.

Key points in managing gas/air embolism:

• Consider Gas Embolism (GE) in ANY periprocedural stroke, neurological or unexpected haemodynamic event. It can happen during virtually any invasive procedure.

• Administer 100% Oxygen

• Early discussion with the nearest Hyperbaric oxygen treatment facility. Relapse can occur after initial ’recovery’.

Gas embolism (GE) is the entrainment of air (or exogenously delivered gas) from the environment into the venous or arterial vasculature, producing systemic effects. The outcome may be rapidly fatal or it may be so mild that the onset is missed and diagnosed retrospectively. In some animal studies a large amount of air was tolerated, but it has been found that as little as 0,02 ml/kg injected in the coronary vessels could be lethal(1).

The immediate management is to stop the gas entry, life support according to the situational need, stabilising the patient, then managing the removal of gas in the most appropriate way applicable to the patient’s condition and with the facilities and skills available.

Although there are case reports of these being aspirated percutaneously, the most effective treatment is urgent HBOT (2–4). The outcome is dependent on HBOT within 6–8 h, akin to early thrombolysis in thromboembolic stroke.

It is essential to keep in mind that many or possibly most episodes of GE are preventable and no effort should be spared to put the necessary preventative steps in place in the different clinical situations where gas embolism may occur.

When gas embolism is suspected, the algorithm above should be followed:

Stop any potential further gas entry into the vascular system

1. Immediately call for anaesthetic/critical care support, as patients can very quickly become unstable or require resuscitation.

i) 100% oxygen, breathing spontaneously or IPPV if required

ii) keep patient flat, supine, consider Durant position, but not good for CPR

iii) urgent ultrasound of the vessels in the neck and chest, transthoracic and transoesophageal echo are useful in making the diagnosis in many instances.

iv) arrange immediate CT to rule out haemorrhagic or thromboembolic stroke – consultant to consultant referral to emphasise urgency

v) early start of inotropes/vasopressors essential to maintain good perfusion pressure

vi) aspiration of gas with a central line could be attempted if line in place and insertion of a multiorifice line should only be considered if air in the heart is shown on ultrasound and can be inserted quickly, not interrupting CPR for more than a few seconds.

1. Contact nearest hyperbaric unit as soon as diagnosis is entertained

2. If patient not fit for transfer needs optimal resuscitation, maintaining perfusion pressure, IPPV in ICU and hourly re-evaluation for transfer.

Hyperbaric oxygen treatment at 2.8 ATA for 5 h or longer is most often used, and take place in a chamber and with staff able to treat critically ill patients.

VV-ECMO Resistant Hypoxia and RV Failure in a case of PVL-MSSA Community Acquired Pneumonia

James Wilson, Francisca Caetano, Richard Fisher, James Doyle, Shahana Uddin and Susanna Price

Royal Brompton and Harefield NHS Foundation Trust, UK, London, UK

Abstract

Introduction: PVL (Panton-Valentine leucocidin) is a cytotoxin produced by certain Staphylococcus aureus strains that causes severe tissue necrosis, including the lung [1]. Use of VV-ECMO in treatment of PVL pneumonia has been described previously [2,3], but we present a case where its severity proved resistant to VV-ECMO support and where acute RV dysfunction (cor-pulmonale) resulted.

Figure 1.

Saggital section of CT Thorax in case of PVL-MSSA demonstrating extensive consolidation and lung destruction. (Ao; aorta, LPA; left pulmonary artery, RPA; right pulmonary artery, C; cavity, ECMO; extracorporeal membrane oxygenation cannulae (arrowed)).

Figure 2.

Transthoracic echocardiogram (parasternal short axis view) in a patient with acute cor-pulmonale relating to PVL-MSSA. (LV; left ventricle, RV; right ventricle, IVS; inter-ventricular septum, coll; pericardial collection)

Case report: A 54-year-old man with a background of type 2 diabetes and beta-thalassemia trait was admitted to his local hospital with shortness of breath. Admission investigations were suggestive of community acquired pneumonia with sputum cultures positive for methicillin-sensitive Staphylococcus aureus (MSSA). He deteriorated rapidly requiring intubation and ventilation and the Staphylococcus was found to be PVL positive, therefore treatment was modified accordingly. Despite this he continued to deteriorate and so was referred to an ECMO centre for retrieval on VV-ECMO for which he was accepted.

On arrival, CT demonstrated multifocal, thick-walled pulmonary cavities and extensive consolidation of both lungs (Figure 1). Transthoracic echocardiogram (TTE) demonstrated normal biventricular function with features suggestive of significantly elevated pulmonary vascular resistance.

Despite adequate VV-ECMO flows he remained hypoxic (PaO₂ 8.7 kPa) with a low post-oxygenator PaO₂suggesting requirement for a circuit change. This was carried out but post-oxygenator PaO₂remained low (13.6 kPa) and so additional strategies (including a secondary oxygenator) to improve hypoxaemia were considered. He started to show improvement on standard VV-ECMO however, and so a target PaO₂ > 6 kPa (saturations >80%) was accepted.

On day 7 he developed elevation of serum lactate and his ECG demonstrated new RBBB. Repeated TTE showed a low cardiac output state due to acute cor-pulmonale with severe right ventricular (RV) dilatation (Figure 2) for which milrinone was started.

As his underlying lung pathology improved, VV-ECMO support was weaned and milrinone stopped. Due to persistent pulmonary hypertension sildenafil was started. Tracheostomy was inserted and on sedation hold, neurological examination was normal. Serial TTE showed partial resolution of the acute RV failure. He was decannulated from VV-ECMO on day 43 and was repatriated to his local hospital at day 47.

Conclusion: This case documents a severe case of PVL-MSSA community acquired pneumonia, where full medical treatment with VV-ECMO support was insufficient to achieve normal oxygenation, and relative hypoxaemia was accepted. Despite this the patient made a good clinical recovery with no gross signs of neurological deficit. Here, the degree of acute RV dysfunction reflects the severity of the underlying pulmonary pathology, highlighting the importance of regular TTE in patients with this rare condition.

Evaluating the extent of ceilings of care decision-making by the Critical Care Outreach (CCO) and Medical emergency teams: the elephant in the room?

Sarah Paterson, Joshua Singleton, Gabriella Bugg and Jamie Gross

Northwick Park Hospital, London, UK

Abstract

Introduction: Most hospitals, like ours, have a critical care outreach (CCO) team that reviews patients who are deteriorating on the wards – usually as part of a medical emergency call team (MET). The aim of these teams is to prevent further deterioration but also act as a link to higher levels of care. CCO teams are now becoming more involved in assisting with treatment escalation plans and ceilings of care particularly in the emergency setting.¹ It is highly questionable whether this is an appropriate use of CCO resources given the limited resources available and difficulty making such decisions in an emergency setting with limited knowledge of the patient and a time when patients may lack mental capacity and/or their wishes unknown.

Objectives:

1. To determine the extent of the role CCO takes with escalation and ceiling of care plans at the time of a patient’s deterioration during a MET call

2. To further characterise which types of patients undergo escalation and ceiling of care plans during a MET call

Method: Data was collected prospectively from all MET calls at Northwick Park hospital over a 5-week period between 6/6/17 and 10/7/17. Data was collected in the form of a written proforma that was filled in by a member of the CCO team. Data collected included patient demographics, comorbidity status (the number of comorbidities), functional status (as determined from social history in admission clerking), previous ceiling of care/escalation plan decisions and whether CCOT were involved in adjusting/instigating an escalation plan.

Results:

• Data was collected from 132 MET calls for 114 separate patients (54 Female, 59 Male, 1 missing)

• Mean age of the patients was 70.3 years (range 18–99)

• 41% (47) of the patients had a previously documented ceiling of care decision. 53% (61) had no documentation. There were 6 (6%) pieces of missing data.

• 31 patients of the 47 (68%) with a previously documented ceiling of care had a DNAR form put in place

• 10 of the patients with DNAR forms in place (32%) did not have any documentation regarding higher levels of care.

• The CCO team were involved in discussions regarding new or revised escalation plans and ceilings of care for 24 (21%) patients. Of these 18/24 (75%) had 1 or more comorbidity, 18/24 (75%) had a degree of functional impairment and 17/24 (71%) had 1 or more comorbidity coexisting with a degree of functional impairment.

Discussion: More than half of patients who required a MET had no ceiling of care or escalation plan in place during the study period. Approximately 1 in 5 MET calls involved some form of treatment escalation planning and these patients frequently had 1 or more comorbidity and a degree of functional impairment. Initiatives are needed to promote advance care planning before the onset of critical illness. This may reduce the burden of making such decisions in emergency situations. The use of functional limitation and comorbidity as criteria to initiate and promote advance care planning could be explored.

Predicting unplanned admissions to the intensive care unit using routinely collected blood tests and vital signs: validation of the LDTEWS:NEWS score

Oliver Redfern¹, Marco Pimentel², David Prytherch³, Paul Meredith⁴, David Clifton², Lionel Tarassenko², Gary Smith⁵ and Peter Watkinson^1,6

¹Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK

²Institute of Biomedical Engineering, Department of Engineering Science, University of Oxford, Oxford, UK

³Centre for Healthcare Modelling and Informatics, University of Portsmouth, Portsmouth, UK

⁴Portsmouth Hospitals NHS Trust, Portsmouth, UK

⁵Faculty of Health and Social Sciences, Bournemouth University, Bournemouth, UK

⁶Oxford University Hospitals NHS Foundation Trust, Oxford, UK

Abstract

Background and Aims: The Royal College of Physicians (RCP) advise that all patients with a National Early Warning Score (NEWS) value of seven or above [1] are considered for transfer to a high dependency or intensive care unit (HDU/ICU). We recently developed a novel risk score (LDTEWS:NEWS) by combining NEWS (vital signs) with the Laboratory Decision Tree Early Warning Score (LDT-EWS [2]), which uses the patient’s most recent laboratory results. This score was developed using data from one acute hospital to predict unplanned transfer to ICU or mortality within 24 hours. The aim of this study was to validate LDTEWS:NEWS using data from three acute hospitals, with a primary outcome of unplanned ICU admission (UICU).

Methods: The LDTEWS:NEWS score was originally developed on adult emergency medical admissions to Portsmouth Hospitals (PH) over five years. The score is a linear weighting of NEWS and LDTEWS, using patients’ most recent laboratory and vital signs measurements. For the current study, we used vital signs and laboratory data obtained from all emergency medical admissions to Oxford University Hospitals (OUH) in 2016. We excluded admissions where the patient was: under 16, discharged alive on the day of admission or had no vital signs observations in the 24 hours prior to the first of either ICU admission or discharge. For each admission, we calculated the maximum NEWS and LDTEWS scores. The discrimination of each score for the primary outcome was assessed using the c-statistic (area under the receiver operator curve). Confidence intervals were calculated using a stratified bootstrap (2000 samples). We identified an equivalent LDTEWS:NEWS “trigger score” with comparable specificity to a NEWS >=7, for which we calculated the sensitivity in the period prior to UICU admission.

Results: The validation cohort contained 16,309 admissions (median age = 73 years) of which 721 (4.4%) resulted in in-hospital death and 159 (1.0%) in unplanned ICU admission. The population was similar to the development cohort, which contained 97,933 admissions (median age = 73 years) of which 4,723 (4.8%) resulted in in-hospital death and 1,078 (1.1%) in unplanned ICU admission. 20.2% of patients had a NEWS of seven or greater at some point during their admission. At this threshold, NEWS had a sensitivity of 78.6% for the primary outcome, with a specificity of 80.6%. The LDTEWS:NEWS score with an equivalent specificity, had a sensitivity of 84.9%. At 12 hours prior to UICU admission, LDTEWS:NEWS identified 58% of ICU admissions compared to 48.4% for NEWS. The c-statistic was higher for LDTEWS:NEWS (0.874, 95% CI 0.848–0.904) than for NEWS (0.857, 0.828–0.889), although this difference was not significant.

Conclusions: The LDTEWS:NEWS risk index showed higher discrimination for UICU and correctly identified approximately 10% more patients with the primary outcome than NEWS in the 12 hours prior to unplanned transfer from the ward to the ICU, with an equivalent specificity. However, these differences in performance were not statistically significant, possibly due to the low number of UICU events in the validation cohort.

Managing challenging behaviours in visitors to Critical Care

Lynne Barrass¹, Nicola Rudkin², Helen Hewitt², Polly Fitch³, Marie Healy¹, Sanet Marais⁴ and Catherine Smith⁵

¹Consultant in Anaesthesia and Critical Care Medicine, Royal London Hospital, London, UK

²Matron in Critical Care, Royal London Hospital, London, UK

³Clinical Psychologist in Critical Care, Royal London Hospital, London, UK

⁴Senior SIster/Psychology Support, Royal London Hospital, London, UK

⁵Senior Sister in Critical Care, Royal London Hospital, London, UK

Abstract

Background: The Adult Critical Care Unit (ACCU) at the Royal London Hospital is a 44-bedded area, covering a range of specialities including polytrauma, neurosurgery, complex vascular, hepatology and renal medicine. In 2017, there was a significant increase in verbal and physical aggression towards our staff. This resulted in low morale, anxiety and stress amongst members of the entire ACCU team. At exit interview, several staff cited this as a key reason for leaving the Unit.

There was a clear need to develop a management strategy via a coordinated multi-disciplinary approach, with the aim of addressing the causes of challenging behaviours, and support the team in addressing them.

The Staff Survey Results also highlighted some concerning trends:

• Experience of physical violence from patients/visitors: Bart's Health 12% -ACCU 57%

• Harassment, bullying or verbal abuse from patients or visitors: Bart's Health 30 % -ACCU 63%

Method: We formed a triumvirate of senior nurse, clinical psychologist and consultant, with support from Trust management team. We reviewed the following sources to identify key concerns:

• Datix incident reporting system

• Patient and relative feedback via ‘IWantGreatCare’ data

• Recent complaints

• Information from a multi-disciplinary ‘Listening in Action’ event

• Exit interviews

From this analysis, we summarised recurring themes and possible methods of resolution. We met every fortnight to review our progress. We used a ‘driver diagram’ to organise our action plan and to maintain a timeline.

Results: The triumvirate worked with the senior management team to achieve the following improvements, with assistance from the service manager to secure financial support when necessary.

Escalation Policy

• Alongside Trust policy, we formulated a specialist local policy for managing challenging behaviours of visitors. This was more flexible than the Trust ‘red card’ policy as it allowed responses which could be adapted to individual needs.

Volunteers Trial

• We introduced trained volunteers to the ACCU. Their role is to assist relatives; helping them in accessing facilities, liaising with bed-space nurses to identify optimal timings for visiting and highlighting escalating situations early so clinical staff can address concerns promptly.

Flexible Visiting Pass Pilot Scheme

• We piloted a scheme for flexiblevisiting hours in extenuating circumstances

Training

• All staff were offered bespoke conflict resolution training conducted by the Security Lead.

Way Finding/Signage

• We improved directions between our waiting area and our unit.

• We have displayed posters which outline that no violence or abuse will be tolerated.

Personal Alarms

• We have completed a business case for personal alarms for areas of lone working.

Visitor Information

A new, improved leaflet has been introduced which provides insight for visiting guidelines within the Unit.

Conclusion: This was a novel process for ACCU where, through utilising the varied skills of a multidisciplinary triumvirate, we were able to consult widely and ensure that there was a significant improvement in both relative and staff well-being. Following a peak (=10) of related incidents in July 2017, numbers have reduced, to on average 3 per month, staff report feeling more equipped to managing challenging situations and annualised staff turnover has reduced.

Improving antimicrobial stewardship using a procalcitonin-based algorithm: a single centre before and after comparison

Benjamin Millette¹, Katy Jeffery¹, Ferenc Los¹, Daniel Day¹, Sarah Millette², Nishan Guha¹, Tim James¹ and Stuart McKechnie¹

¹Oxford University Hospitals NHS Foundation Trust, Oxford, UK

²Buckingamshire Hospitals NHS Foundation Trust, Aylesbury, UK

Abstract

Procalcitonin (PCT) is a biomarker that can be used for diagnosis and monitoring of bacterial infection in intensive care. Algorithms using PCT to assist with antimicrobial stewardship have been shown to decrease consumption of antibiotics and may decrease mortality (1). We hypothesised that introduction of procalcitonin testing in our tertiary medical/surgical/trauma intensive care unit (ICU) would decrease the average duration of antibiotic therapy.

We conducted a single centre investigation comparing a baseline patient cohort to a patient cohort after introduction of routine PCT testing. We conducted a retrospective audit of all patients meeting the inclusion criteria (see Table 1) over a two month period. We then prospectively identified consecutive patients meeting the same inclusion criteria and measured PCT routinely. The results were incorporated into daily multidisciplinary microbiology assessment. When PCT reached <0.5 µg/L or <20% of peak value, a recommendation was made to consider discontinuation of antibiotics. Data collected for both cohorts of patients included demographics, length of stay, type of admission (surgical, medical or trauma), initial antibiotic duration, total daily defined doses (DDD) of antibiotics received, hospital mortality and microbiology lab data.

Our local study classification group determined that ethical approval was not needed and classed the project as a service development because the PCT test has already been determined to be safe and effective, is in routine use in other ICUs and there was no randomisation of patients. Abbott Diagnostics UK supplied the PCT assays free of charge but were not involved in study design, execution, data collection or analysis.

76 patients were analysed in the baseline group and 102 patients in the PCT group. Groups were well matched with the exception of a higher number of trauma patients in the PCT group (3 vs 19 p = .0028). Median initial antibiotic duration was significantly reduced in the PCT group (6.5 days vs 5.5 days p = .0036 by Mann –Whitney U test). Median DDD of antibiotics received was also significantly reduced in the PCT group (13.6 vs 11.2 doses P = .043). There were no differences in mortality or drug resistant infection.

In our single centre study, a PCT-based algorithm was associated with significantly reduced initial antibiotic duration as well as total antibiotic consumption in comparison to a baseline cohort. Limitations include the before and after study design, the potential for the Hawthorne effect and some differences in the patient cohorts. While we cannot definitively establish causality, these results encourage the use of PCT to improve antimicrobial stewardship in our patient population, especially as it may also be associated with improved patient outcome.

Table 1.

Inclusion criteria:
Patient aged ≥18 admitted to ICU with confirmed or presumed bacterial infection being treated with antibiotics
Antibiotics not prophylactic
No specific plan for prolonged antibiotic therapy
Expected ICU stay >24 h
Immunocompetent
No severe infection due to viruses/fungi/parasites/tuberculosis

Clinician and patient consensus on the definition of an adverse event during mobilisation of intensive care unit patients: A Delphi process

Huw R Woodbridge^1,2, Christopher J McCarthy³, Mandy Jones⁴, Matthew Willis^1,2, Caroline M Alexander^1,2 and Anthony C Gordon^1,2

¹Imperial College London, London, UK

²Imperial College Healthcare NHS Trust, London, UK

³Manchester Metropolitan University, Manchester, UK

⁴Brunel University London, Uxbridge, UK

Abstract

Introduction: The mobilisation of patients on an intensive care unit (ICU) is associated with a low number of adverse events. However, studies to date have used different criteria when measuring these adverse events. Further, it is important to include the patient perspective when defining what is unsafe. A uniform adverse event definition used internationally would allow studies to be combined for greater power. This study therefore aimed to develop a clinician and patient consensus agreement on the definition of an adverse event whilst a patient is being mobilised on an ICU.

Method: A three stage, international Delphi study was carried out involving a purposive sample of international ICU clinicians and clinician researchers, as well as former ICU patients and relatives from the UK. Ethical approval was gained and informed consent received from all participants. In round one, participants listed the events they thought were important to include in an adverse event definition, which then underwent content analysis. In round two, participants rated their agreement with whether different events should be included in the final definition on a five-point Likert scale. Events reached consensus for inclusion if 70% or more of participants rated ‘agree’ or ‘strongly agree’. Those events that did not reach consensus in round two, were re-rated in stage three along with anonymised feedback of participants’ previous responses.

Results: The expert panel comprised 10 doctors, five nurses and nine physiotherapists/physical therapists recruited from 10 countries across five continents. In addition, five former ICU patients and one patient relative were recruited from the U.K. Each of the three rounds of the Delphi process were then returned with a 97% response rate. After analysis of round one, 87 potential adverse events were generated for rating in rounds two and three. Following round three, 54 events (62%) reached consensus for inclusion in the adverse event definition. These events fall into categories which include airway, respiratory, cardiovascular, neurological, attachments, falls and injuries.

Discussion: This adverse event definition has been generated using a robust consensus process and represents the agreement of an international group of multidisciplinary ICU clinicians and includes the opinions of former ICU patients. Further work is required to test the feasibility of using this tool to allow standardisation of measurement of adverse events during mobilisation of ICU patients in future research.

Funding acknowledgements: This report is independent research arising from a Clinical Doctoral Research Fellowship, awarded to Huw Woodbridge, (ICA-CDRF-2015-01-026), supported by the National Institute for Health Research and Health Education England. The views expressed in this publication are those of the authors and not necessarily those of the NHS, the National Institute for Health Research, Health Education England or the Department of Health. The authors also acknowledge the support of the National Institute for Health Research Clinical Research Network (NIHR CRN) and infrastructure support for this work was provided by the NIHR Imperial Biomedical Research Centre. ACG is an NIHR Research Professor (RP-2015-06-018).

The ROX index identifies the need for escalation in patients receiving Nasal High Flow Oxygen in level 1 wards – an Outreach Team pilot

Esther Price, Linda Cooper, Emma Prower, Louisa Chrisman, Kathleen Daly, Paul Burdett, Luigi Camporota and Guy Glover

Department of Critical Care, Guy's and St Thomas' NHS Foundation Trust, London, UK

Abstract

Introduction: Nasal High Flow Oxygen (NHFO₂) is a supportive treatment for acute hypoxaemic respiratory failure which may reduce intubation and mortality, however there is limited data on effectiveness outside Critical Care. Local audit identified a cohort with a low mortality, admitted to Critical Care soley for NHFO₂, suggesting that NHFO₂ may be feasible in level 1 wards if supported by the Outreach team. The aim of the study was to analyse characteristics and outcomes of patients receiving NHFO₂ outside of Critical Care.

Method: A pilot project to deliver NHFO₂ in level 1 wards, supported by Outreach (The Clinical Response Team) at Guy’s and St Thomas’ Hospital; December 2017 to August 2018. Inclusion criteria were hypoxic respiratory failure (PaO₂:FiO₂ (PF) ratio <26 kPa). Exclusions included PF <10 kPa, multiple organ failure and altered consciousness. Physiological variables and HFNO₂ settings were recorded and the requirement for Critical Care admission and for respiratory escalation (CPAP/NIV, intubation or ECMO) were studied. Data is presented as n,%, mean ± SD or median[IQR] and comparisons made using Fisher’s exact test, independent-sample T-test or Wilcoxon rank-sum test, using ‘R’.

Results: N = 28; age 65 ± 15yrs; male 17(60.7%). Diagnoses were community-acquired pneumonia (14), aspiration (3), hospital-acquired pneumonia (1), pneumonia related-to-immunosuppression (2), pulmonary oedema (2), post-operative respiratory failure (3) and other (3). 10(35.7%) had bilateral infiltrates. APACHE II was 20.4 ± 5.6 and SOFA 4.2 ± 1.5. Hypoxaemia was moderate severity (PF 20.5 ± 6.9 kPa) and respiratory rate 26 ± 7bpm. 14(50%) patients were admitted to Critical Care, however only 6(21%) required CPAP/NIV (2), intubation (3) or ECMO (1). Time from initiating HFNC to intubation was 5.5[3.9–20.5] hrs. Comparing Critical Care admission versus none, baseline characteristics were similar except male gender (42.9% vs. 78.6%, p = 0.05) and HCO3⁻ (21.5 ± 6.3 vs. 26.2 ± 3.2, p = 0.02). However, the change in ROX index [(SpO₂/FiO₂)/RR] (DeltaROX) after 2 hours was significantly different between the groups (Table 1; Fig 1). After 6 hours, the difference in DeltaROX also reached significance for the respiratory escalation comparison. Similar results were obtained for the change in National Early Warning Score. One patient died after escalation, unrelated to respiratory failure.

Conclusion: In a pilot study, NHFO₂ for selected patients with moderate hypoxaemic respiratory failure appears to be safe in level 1 wards when supported by the Outreach team. The absence of an early improvement in the ROX index may be useful for the early identification of patients who require Critical Care escalation.

Fig 1.

ROX index (A) and NEWS (B) over the first 6 hours; Critical Care admission versus none.

Table 1.

Post-HFNO₂ characteristics.

	No Critical Care Admission	Critical Care Admission	p
Max FiO2	0.44 ± 0.10	0.58 ± 0.07	<0.01
Max Flow(l/min)	46.8 ± 5.4	49.6 ± 7.5	0.26
DeltaROX-2 hrs	1.9 ± 3.7	−1.3 ± 2.8	0.02
DeltaROX-6 hrs	4.8 ± 2.9	−1.4 ± 1.1	<0.01
Delta NEWS-2 hrs	−1.2 ± 3.0	−0.9 ± 2.9	0.80
Delta NEWS-6 hrs	−3.1 ± 2.7	0.1 ± 2.0	<0.01
NHFO2 duration(hrs)	48.6[31.4–61.5]	7.6[3.8–9.9]	<0.01

The role of the Critical Care Outreach Team in determining ceilings of care in acutely deteriorating patients

Lisa Wood, Karen Cotton and Dhupal Patel

Addenbrooke's Hospital, Cambridge, UK On behalf of the East of England Critical Care Network Outreach Group

Abstract

Introduction: There is increasing evidence to suggest that rapid response teams (RRT) and critical care outreach teams (CCOT) are involved in the provision of end of life care (EoLC), and determining limits of escalation of treatment.¹ This has been anecdotally noted across the East of England critical care network. The RRT and CCOT in the East of England have become increasingly consulted in determining ceilings of treatment and end of life care planning.

Aims: We undertook a survey across the network aiming to understand the involvement of CCOT in EoLC planning.

Method: We asked all the hospitals in the East of England via the Critical Care Network Outreach Group the following 6 questions about EoLC using a paper-based survey: 1) Does your hospital have a formal escalation of care proforma? 2) Was the proforma competed within your hospital’s guidelines? 3) Does your patient have a ceiling of care documented? 4) Does your patient have a do not attempt cardiopulmonary resuscitation (DNACPR) decision documented? 5) Was the CCOT invited to advise on a ceiling of care? 6) Did the CCOT advise a ceiling of care discussion following review? Data was collected on all acute referrals to the CCOT over a seven day period.

Results: We received data from 71% (12/17) hospitals within the network. 75% (n = 9) of the hospitals that responded, reported having a formal escalation proforma. The percentage of patients with a documented ceiling of care varied widely between hospitals, and this was reflected accordingly in the percentage of patients with a documented DNACPR decision. CCOT were invited to advise on decision making in up to 34% of total referrals, and recommended decision making in up to 75% of all referrals.

Discussion: This survey showed that CCOT are increasingly involved in EoLC planning within our deanery. We noted that where hospitals did not have an escalation proforma, CCOT had a higher involvement in EoLC discussions, highlighting the importance of documented formal escalation plans prior to deterioration. These discussions are time-consuming and have resource implications for CCOT who are already engaged in managing deteriorating patients on the wards. The changing patient demographic and varying requirements suggests that a different skill set is required in addition to managing deteriorating patients on the wards.

Subarachnoid Haemorrhage: A Case Report and Review of the Literature

Daniel Cunningham

Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK

Abstract

Subarachnoid haemorrhage (SAH) accounts for 5% of all strokes in the UK each year and can be subdivided into traumatic and non-traumatic. The majority of SAH being aneurysmal. Several well known risk factors that increase the likelihood of developing a SAH, including: uncontrolled hypertension, atherosclerosis, illegal drug use, alcohol abuse, and smoking. Several other conditions such as: polycystic kidneys, connective tissue disorders, neurofibromatosis have an increased risk of developing an aneurysmal SAH. The mean age for a non-traumatic SAH is between 40 and 60 years of age and is more common in women than men. Common sites for rupture are the posterior communicating artery (PCA), internal carotid artery and the anterior communicating artery (ACA). SAH results in bleeding into the subarachnoid space with or without intraventricular extension. Patient Demographics: Demographics: A 45 year Caucasian British woman from the East of England. Past Medical History: Depression. No history of operations or other co-morbidities. Drug History: Olanzapine. Family History: Hypertension. Social History: Married with two children. Works as a social worker. Sedentary lifestyle but healthy diet.

Presenting Complaint: Admitted following a witnessed collapse, subsequent seizure activity while walking to her car. On arrival was intubated and ventilated, GCS reported as 8 at the scene with likely aspiration.

Clinical Findings: Pupils equal and reactive. No tonal changes. No on-going signs of seizure activity. Hypertensive to 160/52, MAP 70, Heart Rate 110. Normal FAST scan. No signs of bleeding, head trauma or bony injury. Bloods unremarkable. CT head and CT angiogram: extensive subarachnoid haemorrhage, small volume intraventricular haemorrhage and a small haematoma in the inferior medial part of the left frontal lobe. There was mild hydrocephalus and no large artery territory infarct seen, with an aneurysm of the anterior communicating artery (ACA).

Results: Taken to theatre where she had the aneurysm coiled and an extraventricular drain (EVD) inserted. Drain was removed after the 5th day with a concomitant drop in GCS and right sided weakness. Repeat CT angiogram showed vasospasm in both ACA, moderate vasospasm of the supraclinoid left ICA just proximal to its bifurcation and mild vasospasm of the M1 segment of the left middle cerebral artery. Commenced MAP augmentation to 100 mmHg and a lumbar drain inserted with resolution of right sided weakness and return of speech. Resolution of the weakness and return of speech was maintained after the removal of the lumbar drain and de-escalation of MAP augmentation. CNS antibiotics were started following drain removal. Repeat CT angiogram did not show any further hydrocephalus, vasospasm or signs of bleeding. Right sided inattention remained. The patient was discharged from the unit after 21 days to continue rehabilitation at her local hospital.

Conclusion: Although based on a single case report, it is felt that through timely intervention and management at a specialist neurosurgical centre the outcome from a substantial aneurysmal subarachnoid haemorrhage is optimised, and the best outcome for the patient achieved, mirroring the conclusions highlighted in previous research, demonstrating lower in hospital mortality, with higher numbers of patients discharged home.

An Intensive care units experience of citrate Renal Replacement therapy (RRT); Audit data comparing heparin & citrate RRT and citrate monitoring in RRT

Fiona Lakin, Daniel Keating and Christina Biela

University Hospitals Southampton NHSFT, Southampton, UK

Abstract

In patients admitted to intensive care, those with an AKI have a higher mortality, this rises to 57%-80% in those who require renal replacement therapy (RRT).¹ Citrate is now the recommended anticoagulation for RRT in the intensive care setting requiring accurate and regular monitoring. Data comparison from the introduction of citrate to present time allowed us to look at the changes within the unit and comparison of RRT data. Over the years, data was collected regarding both heparin and citrate anticoagulation in RRT allowing comparison to previous practice. Analysis looked at filter length both with heparin and then citrate, citrate monitoring, compliance with guidelines, filter length and serum sodium changes. Average length of filter using citrate was 28.2 hrs (94 filter events) in comparison to heparin filters which lasted an average of 24.0 hrs (155 filter events). Results showed a poor general understanding of citrate monitoring further analysed through questionnaires completed by medical staff. 23% of patients had initial calcium checks performed correctly and any subsequent changes documented. In only 29% of filter episodes were Serum Calcium and ionised calcium checked and subsequently Calcium:ionised Calcium ratio was not calculated or documented for any filter episodes audited. 3% of filter episodes had a rise in calcium compensation above recommended in the guidelines, in these cases this was not escalated appropriately. Serum sodium was also audited due to the sodium component of sodium citrate, only 1% of filter episodes audited were found to have serum sodium outside of normal range.

Results from this audit show that filters last longer with citrate RRT compared to Heparin RRT and are therefore a positive change in RRT therapy. However, results surrounding monitoring of citrate have led to unit wide change due to concerns regarding patient safety and inadequacy of citrate monitoring. This has led to increased training and education for nursing staff and junior medical staff. It encouraged us to think about ways to improve access to the guidelines at the bedside and led to creation of an educational PDF that could be attached to the filters.

In conclusion, it was found that compliance with monitoring protocols for patients on renal replacement therapy using citrate was poor with results falling below set standards. Due to both the high mortality and complexities of patients requiring renal replacement therapy on the intensive care unit it has been important to instigate further education on the unit and promote guidelines throughout all members of staff.

Should mucoactives be used in acute respiratory failure in the critically ill? Results from a systematic review

Rohan Anand¹, Judy Bradley¹, Mark Borthwick², Bronwen Connolly³, Bronagh Blackwood¹, John Warburton⁴, Brenda O’Neill⁵, Murali Shyamsundar¹, Chee Yap¹ and Danny McAuley¹

¹Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

²Oxford University Hospitals NHS Foundation Trust, Oxford, UK

³Lane Fox Respiratory Service, Guy's and St Thomas' NHS Foundation Trust, London, UK

⁴University Hospitals Bristol NHS Foundation Trust, Bristol, UK

⁵Centre for Health and Rehabilitation Technologies, INHR, Ulster University, Newtownabbey, UK

Abstract

Acute respiratory failure (ARF) is a common cause of admission to the intensive care unit. Mucoactives are a class of medications that aid mucus clearance from the lung and are used in patients with ARF, despite a lack of published guidance (1). The aim of this systematic review was to determine whether the use of mucoactive agents in ARF improves clinical outcomes.

We conducted a systematic review to assess the evidence from clinical trials that had investigated the efficacy of mucoactives in patients with ARF. Our protocol was prospectively published on PROSPERO (Registration number: CRD42018095408). Searches were conducted on MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials (CENTRAL), EU Clinical Trial Register, ClinicalTrials.gov, the World Health Organisation (WHO) trial registry and a grey literature search on OpenGrey. Studies which recruited adults receiving ventilatory support or high flow nasal oxygen for ARF were eligible. Patients with cystic fibrosis and bronchiectasis were excluded. Trials involving at least one mucoactive compared to placebo or standard care were included. The primary outcome was the duration of ventilation as defined in the study publication and secondary outcomes included mortality. Two reviewers independently screened and extracted data. Where a median and interquartile range were reported, data were converted to mean and standard deviation. Results were reported according to PRISMA guidelines.

Our search identified 499 papers. After title and abstract screening, 38 papers were full text screened of which 11 trial reports were eligible. Studies included 4 different mucoactive agents, given by various routes of administration. The majority of the studies were small. The usual standard of care used was also inconsistent across the studies. Based on these findings the quality of the evidence was assessed to be low.

The primary outcome for this review was reported in 5 of the 11 eligible papers. There was a reduction in the duration of ventilation with the use of mucoactive agents, however this was not significant (Figure 1).

Mortality was reported in 10 of the 11 papers. There was a reduction in mortality in favour of mucoactive agents, however this was not significant (Figure 2). There was considerable heterogeneity present across the studies.

Our systematic review of the literature up to June 2018 found evidence that the use of mucoactive agents improved outcomes in critically ill patients with ARF, however this was not significant. Furthermore, this conclusion is limited by the low quality of the evidence. High quality randomised controlled trials are needed to determine the role of specific mucoactive agents in critically ill patients with ARF.

Figure 1.

Forest plot for duration of ventilation.

Figure 2.

Forest plot for mortality.

Abbreviations; ITS: Isotonic saline, NAC: N-acetylcysteine, HTS: Hypertonic Saline, IV: Intravenous, Neb: Nebuliser.

Assessing the risk of mobilising intensive care unit patients receiving vasoactive drugs: A Delphi study

Huw R Woodbridge^1,2, Christopher J McCarthy³, Mandy Jones⁴, Matthew Willis^1,2, Caroline M Alexander^1,2 and Anthony C Gordon^1,2

¹Imperial College London, London, UK

²Imperial College Healthcare NHS Trust, London, UK

³Manchester Metropolitan University, Manchester, UK

⁴Brunel University London, Uxbridge, UK

Abstract

Introduction: Mobilisation of intensive care unit (ICU) patients is recommended to address the physical problems resulting from an ICU stay. The use of vasoactive drugs is a factor in determining a patient’s stability for safe mobilisation. However there is a lack of agreement over exactly when it is safe to mobilise patients receiving vasoactive drugs. This study aimed to develop an expert clinician consensus on the defining characteristics of adult ICU patient groups receiving vasoactive drugs, considered to have a low or a higher risk of adverse events when being mobilised or where mobilisation should be considered contraindicated.

Methods: Following ethical approval, an international, three stage Delphi process was completed. A group of ICU clinicians were purposively sampled from a range of professions and countries and all participants provided informed consent. In round one, clinicians listed how they would assess whether a patient receiving vasoactive drugs was low risk (would be mobilised routinely), higher risk (would be mobilised with more caution, for instance by consulting a lead/senior clinician), or contraindicated for mobilisation. In round two, clinicians rated each indicator of risk on a five point Likert scale. The threshold for including indicators of risk in the final risk group definitions was if 70% or more of participants agreed with an indicator of risk. Those indicators of risk that did not reach agreement were then re-rated in round three after viewing a summary of clinician’s previous ratings from round two.

Results: Twenty-four expert ICU clinicians (10 doctors, five nurses and nine physiotherapists/physical therapists) from 10 countries in five continents were recruited. All three rounds of the Delphi process were then completed with a 97% response rate. Content analysis of the responses to round one yielded 114 different indicators of risk for rating in rounds two and three. After round three was complete, 50 (44%) of these indicators reached consensus for inclusion in the risk group definitions. This firstly included general statements on what to consider when assessing the risk of mobilising patients receiving vasoactive drugs. Secondly, indicators of different levels of risk could be categorised as vasoactive drug specific, cardiovascular specific, plus other indicators such as additional organ support.

Discussion: These results represent the consensus of an international group of expert multidisciplinary clinicians and builds on previous work on judging when it is safe to mobilise an ICU patient. The process highlighted the difficulty in defining risk specifically when there are individual patient factors to take into account. Further work is required to provide empirical evidence to inform risk assessment when mobilising this patient group.

Funding acknowledgements: This report is independent research arising from a Clinical Doctoral Research Fellowship, awarded to Huw Woodbridge, (ICA-CDRF-2015-01-026), supported by the National Institute for Health Research (NIHR) and Health Education England. The authors also acknowledge the support of the National Institute for Health Research Clinical Research Network (NIHR CRN) and infrastructure support for this work was provided by the NIHR Imperial Biomedical Research Centre. ACG is an NIHR Research Professor (RP-2015-06-018).

Simulation of Emergency Evacuation

Ben Griffiths and Alun Roberts

Jersey General Hospital, St. Helier, Jersey

Abstract

Background:

• Emergency evacuation of intensive care patients in the event of fire is an emergency with low likelihood but very high impact

• Fire evacuation training is often theoretical or delivered in a low fidelity environment in the context of statutory and mandatory training

• Evacuation should be timely and follow a defined method to ensure safety of staff and patients

Aims:

• Design and deliver a fire evacuation training drill

• Incorporate this drill into in situ intensive care simulation sessions

• Provide statutory fire evacuation training to staff

• Audit of evacuation times and adherence to policy

• Utilise the unique situational challenges of emergency evacuation to explore human factors and team dynamics

Method:

• Teams of 4 ICU nurses and 1 doctor

• Scenario unknown to participants beforehand

• Response times measured

• Repeat scenario after debrief

• Response times compared with initial scenario

• Monthly repeat of scenario with different simulation team to test improved institutional performance

Conclusions:

• In situ high fidelity ICU simulation can be a highly effective medium for improving response times and policy adherence in emergency evacuation drills

• Participants reported high satisfaction with this mode of training

• Statutory training can be a useful vehicle for human factors training in a simulation environment

Distance between the tips of central venous catheters inserted via the internal jugular vein does not depend on whether the catheters are inserted on same site or opposite sites

Sophia Butt and Luigi Camporota

Guy's and St Thomas' NHS Foundation Trust, London, UK

Abstract

Background: Central venous catheters (CVC) are routinely required in intensive care for infusion of drugs, often concurrently with central venous dialysis catheters (CVDC) for renal replacement therapy (RRT). Increased clearance of drugs, due to direct aspiration via the CVDC, has been reported when both catheters were inserted into the same internal jugular vein. This prompted a call to place catheters in different veins.

PICO question: In patients requiring a CVC and CVDC for renal replacement therapy, does insertion of the CVDC in the opposite internal jugular vein to the CVC increase distance between the line tips?

Methods: We conducted a review of the 23 chest x-rays performed between 30/5/18 and 30/7/18 of patients with a CVC and CVDC within the same or opposite internal jugular vein. Distance between the tips of the CVDC and CVC were measured using PACS. Insertion of lines were checked on x-ray and clinically. Results were entered in EXCEL. Data were analysed in STATA v 15

Results: Among 23 patients, 60.87% had catheters inserted in the same vessel and 39.13% had catheters inserted in opposite sites.

The mean distance (SD), [95%CI] between tips was 31.4 mm (18.4), [17.3 to 45.6 mm] for catheters inserted in opposite sites; while for catheter inserted on the same sites it was 37.3 mm (24.7), [23.0 to 51.5 mm].

The mean difference was −5.8 mm [95% CI −25.8 to 14.2]; p = 0.55. (figure 1).

Fig 1

Conclusions: There is no difference between the tip distance of central venous catheters inserted on the same site or opposite sites.

Recommendation: Attention needs to be paid to the distance between the line tips – site of insertion is less relevant and does not influence final distance between the two line tips

The use of proton pump inhibitors in critically ill patients: a clinical audit

Ana Spataru, Thomas Jones, Skye Perkins, Reena Mehta and Tom Hurst

King's College Hospital, Critical Care Department, London, UK

Abstract

Background: Gastrointestinal bleeding secondary to stress ulcers is associated with significant mortality and morbidity in critically ill patients. Although proton pump inhibitors (PPIs) have significantly reduced the incidence of stress ulcers in this population, their use is not harmless and has been associated with potential complications and additional costs.

Objective: To describe the pattern of PPI prescription and to establish the compliance with local stress ulcer prophylaxis guideline in patients admitted to Intensive Care Unit (ICU) in a large urban tertiary care hospital over a period of 3 months.

Methods: One hundred twenty-two adult patients admitted to General ICU between April 2017 and June 2017 were included in the audit. Medical records were reviewed retrospectively for information on PPI prescription during admission and at discharge from ICU. The local stress ulcer prophylaxis guideline for adult Critical Care was used to assess the appropriateness of PPI prescriptions.

Results: Eighty-four per cent of patients were prescribed PPIs on admission to ICU. Invasive mechanical ventilation, sepsis, and acute kidney injury were the main indications for initiating stress ulcer prophylaxis. The most frequently encountered indications for continuing PPIs in ICU were intolerance to oral feeding and acute kidney injury. In 14% of patients, PPIs were continued without a justified indication. In 64% of cases, PPIs were converted from the intravenous to the oral form during patients’ ICU stay. In only 44% of cases, the timing of the conversion was appropriate. In half of the patients, the intravenous PPI form was switched into the oral form later than 24 hours after conversion criteria were met. An appropriate PPI prescription strategy was followed in 65% of the patients discharged from ICU. In 25% of cases the prescription was not indicated and in 10% of cases a PPI was indicated but was not prescribed.

Conclusion: The results of this audit highlight a limited compliance with local stress ulcer prophylaxis guideline resulting in a large proportion of inappropriate PPI prescriptions in critically ill patients. Interventions aimed at increasing prescribers’ awareness of clinical guidelines are needed in order to reduce the risks associated with PPI misuse.

Dear SIRS, we are sorry to say you are ousted but NEWS is… we have made an IMPACT!

Mahendra Kakollu and Ranga Mothukuri

ED Morriston Hospital, Swansea, UK

Abstract

Aim: To increase awareness and recognition about sepsis.

Introduction: Sepsis is the leading cause of death and is responsible for nearly 44,000 deaths in the UK alone. Therefore, improving compliance of sepsis screening and delivery of sepsis six within one hour was one of our aims in this quality improvement project. Currently, in most hospitals in Wales, NEWS of 3 is being used as an early warning system to identify acutely unwell patients who could potentially be unwell with sepsis. There is evidence to show that NEWS of 6 is associated with more likely rapid deterioration if they are septic. Therefore, we wanted to show improvement in this group of patients.

Methods: This study was conducted in a single centre University Teaching Hospital Emergency Department that sees approximately 100,000 patients annually. SeGMED (Sepsis Group Morriston Emergency Department) was formed in 2016. Our mission statement was, “To inspire, educate and promote early recognition, treatment and management of sepsis”. The group incorporates receptionists, a data analyst, Health Care Assistants, Staff nurses, Advanced Nurse Practitioners, junior doctors, a registrar and a consultant. Sepsis has now become a part of the departmental mandatory teaching. We have developed a new sepsis screening tool, created a staff notice board with weekly updates, compliance charts, star of the week and pictures of the SeGMED group for staff to communicate. Since the groups’ creation in 2016, we have done four sepsis awareness days. We had prominent guest speakers and patients who have had their lives affected by sepsis come and narrate their side of the story. The whole process has helped us improve our screening tool and also improve compliance.

Results: The results are from patients registered between June 2017 to February 2018, a total of 9 months. Our data showed steady progress. There was a dramatic increase in screening tools being commenced from 305 per month in June 2017 to nearly 700 a month in December 2017. For patients with NEWS ≥6, there was a 27% increase in patients receiving all elements of sepsis six. Focussing on the number of septic patients who were given antibiotics within the hour, there was an increase from 35% in June 2017 to 40% in January 2018. In the group with NEWS >6, our 1-hour antibiotic compliance rose from 42% in June 2017 to 47% in February 2018.

Conclusion: For nearly 80% of the patient’s sepsis journey begins outside the hospital. Whenever a patient or relative comes to us and asks if one of their relatives or friends could have sepsis because of the visual prompts and awareness around sepsis, we know we have made an impact. Our results are proof that our busy staffs are more aware of sepsis and conscientious of commencing and completing screening forms for one main cause – TO SAVE SOMEONE’S LIFE.

Mobile extracorporeal membrane oxygenation (ECMO) – A Review of 6 years of experience

Ada Maria Krzak¹, Jo-anne Fowles² and Alain Vuylsteke²

¹University of Cambridge, Cambridge, UK

²Royal Papworth Hospital, Papworth Everard, UK

Abstract

Introduction: Adult respiratory extracorporeal membrane oxygenation (ECMO) is indicated as part of treatment escalation in acute and severe but potentially reversible respiratory failure.¹ Due to limited demand, high risk and resource-intensive nature of the procedure five centres are commissioned in England to provide this service using a specialist retrieval team. Our aim was to evaluate the service provided by one of the 5 centres using the clinical reports written by the consultant leading the retrieval team.

Methods: We conducted a single-centre retrospective qualitative study of consecutive ECMO retrieval clinical reports written between October 2013 and April 2018. The reports described clinical details of the adult patients with severe acute respiratory failure assessed by the team, as well as the process of care including complications and reflections about the retrieval and lessons learned.²

Results: A total of 176 patients from 48 hospitals were referred for assessment. Three patients were referred for mechanical cardiac support and excluded from the analysis. Six (3.5%) patients died before, during or after cannulation; 31 (17.9%) were declined either due to improvement (22) or futility (11); 124 (71.7%) had ECMO started at the referring site; 6 of the 12 retrieved without ECMO required ECMO after transfer to the ECMO centre; and 3 patients were transferred to a different nationally commissioned ECMO centre. Training staff was present on 40 (23.1%) retrievals.

Recorded issues included ambulance delays (2.3%) or malfunction (0.6%), navigation issues (0.6%), blood product unavailability (3.5%), theatre delays (3.5%) and delays at arrival to the referring hospital (3.5%) and back to ECMO centre (3.5%). There were 11 (6.4%) reports of unavailable equipment, 5 (2.9%) of equipment failure and 3 (1.7%) O₂supply issues during patient transfer – none of them were reported to have caused patient’s harm. Forty-three (24.9%) reported excellent communication and team work, while eight (4.6%) reported communication issues between the ECMO centre and the referring team, such as lack of consultant handover at arrival (1.7%), medication administration issues (0.6%) and lack of interpreter (0.6%). Twenty-one (16.9%) reports mentioned issues related to patients’ large body habitus such as difficulties securing patient on the trolley (5.6%) and cannulation difficulties (12.9%).

Conclusion: In this evaluation the prevalence of critical events during retrievals was relatively low.³ The issues identified were discussed during team debrief and addressed in subsequent practice.

Initial Measurement of Ideal Body Weight in Ventilated Patients to Guide Tidal Volumes

Michael Edwards and Ascanio Tridente

St Helens and Knowsley NHS Trust, Whiston, UK

Abstract

Introduction: Lung hyperinflation is a known contributing factor to ventilator-induced lung injury. Tidal volume (Vt) strategies utilising 6 ml/kg targets have shown decreased mortality and fewer days of ventilation when compared to more traditional Vt approaches^1,2. Early measurement of Ideal Body Weight (IBW) on admission to critical care is essential to allow accurate tidal volume ventilation which will subsequently improve patient outcomes.

Objectives: To identify ventilated patients who were managed with a lung protective strategy upon arrival to critical care based on their recorded ideal body weight (IBW), comparing accuracy of ventilation against patients where IBW was not calculated on admission.

Methods: Retrospective note analysis of all patients admitted to critical care over a 2-month period (25/8/17–25/10/17). The patient’s actual Vt was compared to their ideal Vt (6 ml/kg) as calculated from their IBW. IBW was calculated using the Devine formula. Patient height was calculated using the British Association for Parenteral and Enteral Nutrition (BAPEN) method that utilises measurement of ulna length.

Results: 52 patients requiring invasive ventilation were identified. 16 (30.7%) of patients had their IBW recorded on arrival with 36 (69.2%) having their IBW recorded during their stay. The mean difference between IBW and actual body weight (ABW) was 8.5 kg (11.1% of ABW). The average discrepancy between ideal tidal volume and actual tidal volume was 58.7 ml/min; if IBW was recorded on admission this was 49.1 ml/min but increased to 70.4 ml/min in those whose IBW was not recorded at all. Logistic regression analysis of mortality with adjustment for APACHE-II score, gender and age however identified no significant difference in patient outcome.

Conclusions: Patients who have their IBW recorded on admission are more likely to be ventilated nearer to the ideal 6 ml/kg Vt. Whilst our study was likely underpowered due to insufficient patient numbers, it is important to maximise the number of patients whose IBW is recorded upon arrival at critical care.

Severe Traumatic Brain Injury: A case Study and Review of Literature

Orlando Buhay

Cambridge University Hospitals NHS FT, Cambridge, UK

Abstract

Case Study

Traumatic brain injury (TBI) is a major cause of death and disability worldwide, with more than 13 million people estimated to live with disabilities related to severe brain injury. It is the largest single cause of death in those aged under 40 years in the developed world with enormous economic consequences globally. All patients with severe grades of TBI are commonly managed in the intensive care unit (ICU) or neuro intensive care unitwith a combined medical and surgical approach.

Background: A 40-year-old male was presented to the emergency department after he was involved in a road traffic accident, pedestrian hit by a car. He was GCS 6 at scene with both pupils equally reactive to light size 4mm. Sustained multiple scalp and an extensive left facial laceration extending to lateral left eye lid. He was intubated, ventilated and sedated at scene. Patient was taken to CT Scan wherein they found that he had a right frontoparietal subdural haematoma, traumatic subarachnoid haemorrhage, scattered petechial haemorrhages- suspicious of diffuse axonal injury and has a feature of raised intracranial pressure.

Interventions: He was taken to theatre for insertion of triple bolt for multimodal monitoring devices and suturing of his facial lacerations. Multi monitoring devices used were ICP monitor, Lycox machine measuring the PbtO2 and Microdialysis machine in combination of ICM+ software to individualise management of brain injury. Patient was also put on therapeutic hypothermia management to a low temperature of 33 degrees. The initial ICP was 50 mmHg upon arrival to ICU post operation. Hypertonic saline and sedation boluses were given to stabilise intracranial pressure. CPP was augmented and maintained between 60–70mmhg with the use of inotropic agent. Management was escalated with therapeutic hypothermia and despite all the medical interventions given, patient’s ICP remained difficult to manage hence the decision to perform craniectomy and external ventricular drain (EVD) was inserted on his fifth day of hospitalisation. On the 31^stday of admission, patient was discharged to a rehabilitation ward with a GCS of 15.

Conclusion: Based on the case study presented, a protocol based management of TBI when timely executed in a specialised unit, best possible outcome can be achieved.

Intra-tidal PaO₂, PaCO₂ and pH oscillations in lung-lavaged pigs: implications for clinical arterial blood gas sampling

John Cronin¹, Rongsheng Chen², Douglas Crockett², Andrew Farmery², Goüran Hedenstierna³, Anders Larsson⁴, Luigi Camporota^1,5 and Federico Formenti^1,2

¹Centre for Human and Applied Physiological Sciences, King’s College London, London, UK

²Nuffield Division of Anaesthetics, University of Oxford, Oxford, UK

³Hedenstierna Laboratory, Department of Medical Sciences, Uppsala University, Uppsala, Sweden

⁴Hedenstierna Laboratory, Department of Surgical Sciences, Uppsala University, Uppsala, Sweden

⁵Department of Adult Critical Care, St Thomas' Hospital, Guy's and St Thomas' NHS Foundation Trust, King's Health Partners, London, UK

Abstract

Introduction: Arterial blood gas (ABG) measurements are frequently used to titrate mechanical ventilation settings in the acute respiratory distress syndrome (ARDS).¹ Recently, however, large intra-tidal swings in PaO₂ (up to 58 kPa in amplitude²) have been measured using fast-acting intra-arterial oxygen sensors in animal models of ARDS.³ We hypothesized that spot ABG samples, sampled in rapid succession (satisfying Nyquist criteria) in lung-lavaged pigs would demonstrate oscillations at the respiratory frequency similar to those seen by intra-arterial sensors, with implications for the clinical utility of single ABG measurements in titrating ventilator settings.

Methods: The study received ethics committee approval (Uppsala REC ref C98/16). Eight domestic pigs, weight 28.8 (1.9) kg (mean (SD)) were studied under general anaesthesia. A saline-lavage surfactant-depletion lung injury model was used. Tidal volume was 20 mL/kg using pressure control ventilation, PEEP 5 cmH₂O, inspiratory:expiratory (I:E) ratio 1:2 and respiratory rate 6 bpm in 4 animals (RR6 group) and 12 bpm in the other 4 (RR12 group). PaO₂ was monitored continuously using an intra-arterial sensor. 10 ABG samples were rapidly obtained via a 20G arterial line (aiming for a sampling frequency of at least twice the respiratory rate), analysed in a random order and the amplitude and dominant frequency of a cubic spline interpolation of PaO₂, PaCO₂ and pH values determined.

Results: Intra-tidal PaO₂ oscillations were observed at the respiratory rate by the intra-arterial sensor in all animals (amplitude 4.8 (2.2) kPa). Spot ABG sampling frequency was 4.0 (0.3) and 2.9 (0.2) times faster than the respiratory rate at RR6 and RR12 respectively. The dominant frequency of interpolated spot PaO₂ measurements was 5.3 (0.6) bpm at RR6 and 11.1 (0.4) bpm at RR12, with similar frequency oscillations seen in PaCO₂ and pH. Amplitude of interpolated spot ABG PaO₂ samples was 4.3 (3.1) kPa representing a mean increase of 0.5 [95% limits of agreement −7.3 to 8.3] kPa from intra-arterial sensor measurements. Interpolated spot ABG PaCO₂ oscillation amplitude was 0.7 (0.2) kPa and pH oscillation amplitude 0.025 (0.005).

Conclusions: Rapid sampling of ABGs identifies oscillations in PaO₂, PaCO₂ and pH at the respiratory frequency in lung-injured pigs. Mean amplitudes of PaO₂ oscillations were similar to those measured by a rapid acting intra-arterial sensor however a large degree of variability was seen. The oscillations observed were clinically significant and if demonstrated in ARDS patients imply that management decisions based upon single values could be incorrect if the sample was taken too rapidly. It is recommended that ABG samples should be slowly aspirated over several respiratory cycles in order to minimise variability.

Ultrasound Guided Peripheral Cannulation in the Emergency Department: A Cheap and Effective Multi-Disciplinary Approach to Improve Patient Experience of Cannulation

Thomas Smith¹, Christopher Hodgkinson¹, Alistair Wood² and Anu Mitra¹

¹Imperial College Healthcare NHS Trust, London, UK

²GE Healthcare, London, UK

Abstract

Background: Patients who are difficult-to-cannulate may need repeated painful attempts before successful access is gained as well as experiencing delays in receiving treatment such as fluids, antibiotics or intravenous analgesia. Traditionally, in the ED, ultrasound has been used exclusively by doctors. Other professions have felt unable to develop knowledge and experience of ultrasound. Ultrasound guided peripheral cannulation (UGPC) has been shown to improve cannulation success rates in difficult-to-cannulate patients. However, most cannulation in the ED is performed by professionals other than doctors and it is usually only registrars and consultants who can cannulate under ultrasound guidance. Registrars and consultants often have high workloads meaning that difficult-to-cannulate patients may have to wait for a member of staff capable of performing UGPC and therefore suffer delays to their treatment.

Methods: Over a period of 6 months we arranged 6 teaching sessions which were advertised to staff of all grades and roles within the EDs of 2 London teaching hospitals. Participants were taught how to perform UGPC on a homemade ultrasound phantom during a 2-hour teaching session. A competency framework was then designed to further increase the confidence of those who attended a course, compile a logbook of cases and to have a final sign-off of their newly gained competency. Patients who experienced UGPC were then surveyed to get feedback of their experience.

Results: Patients who were surveyed felt that UGPC was a better experience compared to normal cannulation and was associated with reduced pain, reduced inconvenience and increased effectiveness. Most patients also stated they would recommend UGPC to friends and family as well as saying they would prefer it over normal cannulation in the future. Attending a course markedly increased the confidence of all staff who attended and made them feel more likely to use UGPC for difficult-to-cannulate patients in the future. Other hospital departments also showed interest in the project. The cost of a phantom was reduced to 78 pence. Over 95% of patients said they preferred UGPC to normal cannulation and would recommend UGPC to a family member. 79% of patients believed UGPC to be more effective and 52% felt it was less painful when compared to normal cannulation.

Conclusions: Difficult-to-cannulate patients have better experiences of cannulation because of UGPC compared to normal methods of cannulation. Teaching UGPC to all healthcare professionals in the ED is a cheap and effective way of improving patient experience of cannulation and has the potential to be rolled out to other departments within the hospital.

Evaluating the documentation of nasogastric tube insertion and adherence to safety checking

Lowri Roberts¹, David Hepburn², David Penney¹ and Jonathan Pugh¹

¹Cardiff University, Cardiff, UK

²Royal Gwent Hospital, Aneurin Bevan Health Board, Newport, UK

Abstract

Aims: Enteral feeding into a misplaced nasogastric tube is recognised by the National Patient Safety Agency as a Never Event (1). These invasive tubes are commonly indicated amongst level 2 or 3 patients thus we set out to evaluate the current practice in critical care.

The aim was to evaluate: the documentation of insertion, adherence to safety guidance pertaining to checking safe use and assessment of correct chest radiograph interpretation.

Methods: This prospective cohort study was based on inpatients in critical care who had insertion of nasogastric tubes over a four-week period; there were 57 insertions. Data was analysed from patients’ medical and nursing notes as well as the hospital’s imaging system.

Results: 65% of insertions were documented using proforma stickers. 97.1% of these proforma documentations included 4 or more of the following details: type of tube, tube length at the nostril (cm), NEX measurement, aspirate adequacy, chest x-ray adequacy, whether it was safe to feed. Only 13.6% of the hand-written documentation included 4 or more of these details.

Initial aspirates were only obtained on insertion in 51% of cases, of these, 57% had an appropriate pH between 1 and 5.5. This lead to 74% of patients having chest radiographs to confirm initial placement of nasogastric tube on insertion. Only 54% of chest radiographs adequately met the anatomical requirements of the four criteria (2). Written documentation in medical notes stating whether it was safe or unsafe to feed was only completed in 67% of cases.

Conclusion: There were large inconsistencies in the amount of detail documented on insertions of nasogastric tubes. It is shown that proforma stickers ensure a higher level of detail and better uniformity in the notes. There was a high incidence of chest radiographs being performed to confirm correct placement of tubes; this was due to difficulties in obtaining aspirates and failures to follow guidelines, by using this as a first line rather than, correctly, as a second line method of confirmation (3). A need for a uniform, ward-specific proforma on nasogastric tube insertion has been identified. It has also been recognised that there is a need for a teaching session on chest radiograph interpretation and on techniques to aid obtaining successful aspirates.

Using technology to understand the symptoms associated with conditions presenting to critical care units and introducing pre-emptive strategies to treat them

Timothy Baker¹, Katy Surman² and Vilas Navapurkar¹

¹Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK

²University of Cambridge, Cambridge, UK

Abstract

25–65% of patients who survive critical illness suffer from Post Traumatic Stress Disorder.¹ An inability to effectively communicate is the most stressful aspect of being an ICU patient^2,3 often because of a delay or failure to treat uncommunicated symptoms. It has been shown that nurses underestimate the frequency and severity of symptoms^4,5. Early interventions have been shown to have the greatest impact on reducing psychological morbidity.ⁱ We have started a quality improvement project intended to reduce the symptoms experienced by ICU patients. A software tool (“myICUvoice”) was introduced to enable patients to communicate with nurses and to consequently allow clinicians to better treat the symptom burden of individual patients. Investigations into the efficacy of myICUvoice are on-going. We report an early observation.

All patients admitted to a single ICU were provided continuous access a tablet computer with a reporting tool (myICUvoice) loaded onto it. myICUvoice^Ó is a ICU specific symptom communication tool, that allows all patients to continuously self-report symptoms. Nurses were encouraged to ask the patients to complete a focused daily survey to understand how they felt. Questions focused on physical, emotional, ICU intervention and pain related symptoms. Patients were also able to self-report at any time they wanted during their admission and could select any one of the 84 symptoms listed. The responses were automatically uploaded into a database and then compared to the ICNARC dataset for these patients. As above

278 patients used the reporting tool in the first 6 month period of use. There was a wide spread of types of symptoms being reported across patients and conditions (n = 84). Due to small sample sizes for specific conditions we focused our attention on our top 3 admission diagnoses (acute kidney injury (AKI) n = 32, liver ransplant n = 31, and pneumonia n = 54) and those patients whose outcome was ‘worse or dead’ (n = 46). The most frequently reported symptoms were similar irrespective of the condition of the patient. High levels of anxiety, dry mouths, being too hot and too uncomfortable were consistently reported. ‘My mouth is dry’ is the most frequently reported symptom irrespective of the patient’s conditions (48% of all patients). The symptom profile of patients who died on ICU does not differ significantly from those who do not.

In this early analysis patient numbers are not sufficiently powered to attain any statistical significance. However, this evaluation has provided us a better understanding of the symptoms most frequently occurring for our patients. Our current methods of reducing xerostomia, temperature control and general comfort need to be adapted. Continuing to use this tool as part of our routine practice we will allow us to assess how effective our interventions are at reducing these symptoms. As our data set grows we hope to develop early and pre-emptive strategies to prevent the negative symptoms associated with specific critical care conditions.

Mobile EEG in ICU, a quality improvement project

Sara Leddy, Mark Sargent and Kate Regan

Brighton and Sussex University Hospitals, Brighton, UK

Abstract

In 2013 guidance from the European Society of Intensive Care Medicine recommended the use of EEG to investigate several clinical states commonly encountered in critical care. These include generalized convulsive status epilepticus, non-convulsive status epilepticus prognostication for hypoxic brain injury in patients post cardiac arrest, and investigation of unconsciousness of unknown aetiology.¹ A survey of 55 UK trusts in 2015, demonstrated that only 18 had access to mobile EEG despite the current guideline.² Brighton and Sussex University Hospital NHS Trust, has a total of 39 critical care beds across two geographical locations. It is a Major Trauma Centre, tertiary referral for major vascular, cardiology and cardiothoracic surgery and regional neurosurgical services. The 7 bedded neuro-critical care beds are located at the Royal Sussex County Hospital (RSCH), in Brighton. The neurophysiology department is located at a smaller hospital within the trust, the Princess Royal Hospital (PRH). This department provides services to four other local hospitals spanning a wide geographical area, with 2 neurophysiology technicians available. Following a concern regarding delays in access to EEG services for ICU patients, we carried out a retrospective data collection of EEG referrals across the RSCH and PRH sites. We investigated the number of referrals, indication, time to obtaining EEG and EEG report. We also looked at the outcomes of patients who had an EEG and the ICU and hospital length of stay.

Total number of patients	23
Age range	25–75
Average delay at RSCH Range	3 days 1–10 days
Average delay at PRH Range	<1 day 0–3 days
Average time in critical care Range	6 days 1–69
Out of hospital cardiac arrest	10
Reduced GCS	10
Seizure	2
Intracerebral haemorrhage	1

This data highlighted the increased demand for EEG at RSCH, which is likely due to it being the regional trauma centre and a much larger ICU. It also highlighted a significant delay for patients at RSCH compared with PRH, which in the context of OOHCA could lead to increased emotional stress for families. These delays may also contribute to unnecessarily lengthy and costly ICU stays for patients with other indications for EEG. Following PDSA quality improvement methodology, a novel mobile EEG device is being piloted prior to introduction of new equipment. Further studies will be undertaken to assess the impact of this equipment on reducing times to obtaining EEG, and this effect on patient outcomes.

An ethnological assessment to evaluate how nursing staff responded to the introduction of a new software tool to improve the understanding the symptoms experienced by ICU patients

Vilas Navapurkar¹, Timothy Baker¹, Nadya Pohran²

¹Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK

²University of Cambridge, Cambridge, UK

Abstract

25–65% of patients who survive critical illness suffer from Post Traumatic Stress Disorder.¹ An inability to effectively communicate is the most stressful aspect of being an ICU patient^2,3 often because of a delay or failure to treat uncommunicated symptoms. It has been shown that nurses underestimate the frequency and severity of symptoms^4,5 Early interventions have been shown to have the greatest impact on reducing psychological morbidity.ⁱ We have started a quality improvement project intended to reduce the symptoms experienced by ICU patients. A software tool (“myICUvoice”) was introduced to enable patients to communicate with nurses and to consequently allow clinicians to better treat the symptom burden of individual patients. Investigations into the clinical efficacy of myICUvoice are on-going. Our initial focus has concentrated on understanding how nursing staff respond to incorporating myICUvoice, and to develop strategies to improve its introduction to new nursing staff.

All patients admitted to a single ICU were provided continuous access to myICUvoice^Ó. myICUvoice^Ó is an ICU-specific communication tool that allows patients to continuously self-report symptoms. Nurses were encouraged to ask patients to complete a daily survey focused on physical, emotional, ICU intervention, and pain related symptoms. Patients were also able to self-report at any time during their admission. 5 months after its introduction, we conducted qualitative research with nurses, which consisted of 25 interviews (purposive sampling techniques were used to ensure equal representation in seniority), observation (20 hours), and a survey (30 respondents.) Grounded-theory was used to develop codes, concepts, and categories which cumulatively revealed significant trends revealing how and why myICUvoice was being used.

Nurses used myICUvoice equally across all age groups and showed a desire to consistently use the tool. However, although encouraged to use the tool with every patient, 88% of nurses interviewed explicitly or implicitly revealed that they predominantly focused their use of myICUvoice to a particular type of patient. This was a CAM ICU negative, RASS 0, ventilated patient. Nurses tended to use the tool less on patients who were able to phonate. Only 1 of 25 interviewed nurses expressed an understanding of the need to ask focused and holistic questions to better understand all symptoms experienced by our patients. Qualitative feedback has demonstrated enormous benefit for individual patients.

The introduction of this tool has raised awareness of the need to better understand the needs of our patients. Most poignantly it has changed the dynamic between our nurses and ventilated patients. The importance of providing augmented communication tools is now a priority because the benefit is easy to demonstrate. This study has demonstrated that with verbally capable patients, nurses still prefer the routine question ‘how are you?’ to which an automatic ‘ok’ response predominates. We are now developing a training program to facilitate investigating whether routinely and systematically asking patients to self-report physical, emotion, intervention and pain related symptoms improves our understanding of the personal burden of critical illness. Concurrently we are also developing strategies to effectively address problematic symptoms.

A Strategy for Tackling the Reproducility Crisis in Big Data Critical Care Research

Edward Palmer, Niall MacCallum, Dave Brealey, Mervyn Singer and Steve Harris

UCL Dept. Medicine, London, UK

Abstract

Introduction: The rise in electronic health records (EHR) for secondary research presents new challenges. To protect patient confidentiality, it is rarely possible to publish patient level data with all processing and analytic decisions alongside a study manuscript. This lack of transparency limits the reproducibility of clinical informatics; at best slowing translational research and at worst contributing to the reproducibility crisis in medicine[1]. The Critical Care Health Informatics Collaboration (CC-HIC) is a large multi centre research project, aggregating high fidelity time series data on critical care patients from 12 intensive care units across five biomedical research centres in the UK (UCL, Cambridge, Oxford, Guy's and St. Thomas' and Imperial). 263 covariates are available to researchers including: demographics, acute illness severity scores, high resolution bedside monitoring, drug infusions, microbiology and organ support. Data validation is performed prior to issuing anonymised data releases to external research collaborators. Building on concepts introduced in the RECORD statement[2], and adapted from Khan et. al.[3], we present a method for the publication of data provenance and validation, detailed in a public “table zero”. Table zero provides an immutable description of upstream statistical decision making, necessary to maintain the scientific record and facilitate study replication and reproducibility.

Objectives: To showcase a principled method for the publication of data provenance and validation from a large confidential critical care research resource; the CC-HIC research database.

Methods: A number of validation checks are applied to raw CC-HIC data. These primarily aim to address fundamental issues of logical consistency; e.g. an episode end must occur after an episode start. Records that comply sufficiently with these checks are flagged for “safe” downstream research. This process is iterative; allowing contributing centres to correct errors, and evolving; adding new rules to improve the reliability of validation over time. Details of these checks, along with the code required to recreate the CC-HIC database in its current state are available online in a freely accessible format, and published with an immutable digital object identifier[4].

Results: CC-HIC has aggregated high fidelity data on 37,571 ICU episodes from 01 Jan 2014, with ongoing data collection. 35,419 (94%) episodes pass provisional validation checks and can reliably used by external collaborators for downstream research.

Conclusions: We have presented a robust and novel framework for ensuring data provenance whilst maintaining confidentiality at the patient level. This method provides sufficient information for external research teams to replicate research from CC-HIC. This method allows for research to be independently repeated and confirmed. These are essential materials and methods to facilitate translational learning from the CC-HIC resource.

Grant Acknowledgement: This research was funded by the National Institute for Health Research Health Informatics Collaborative and supported by the National Institute for Health Research University College London Hospitals Biomedical Research Centre.

Dr. Palmer is funded by the MRC

EVALUATING MANAGEMENT OF HYPOGLYCAEMIA IN THE INTENSIVE CARE UNIT

Catarina Orvalho, Ken CH Lee, Sara Serafino and Kyra Dingli

Guy's and St Thomas' NHS Foundation Trust, London, UK

Abstract

Background: Hypoglycaemia is well-recognized for its potential harmful effects in patients, which include problems such as autonomic instability, cardiac arrhythmia, neuronal injury and pro-inflammatory response. Its association with higher mortality, or prolonged length of stay in the intensive care unit (ICU) independent of the initial disease severity, have also been consistently demonstrated in multiple studies. The critically ill population is particularly vulnerable to hypoglycaemia because these patients often present with coincident problems of renal and liver dysfunction, malabsorption and intensive insulin therapy. Detection of hypoglycaemia may be further hampered in patients with reduced consciousness, so close laboratory monitoring becomes crucial for early recognition and treatment of hypoglycaemic episodes. This study was performed as part of an audit on glucose control management at our ICU to review the frequency of hypoglycaemic events, current practices in hypoglycaemia management, and identify pitfalls or possible areas for improvement.

Methods: We performed a retrospective observational study of hypoglycaemic episodes (serum glucose <4 mmol/L) based on data obtained from the electronic records at St Thomas’ Hospital Intensive Care Unit between April 2017 and April 2018. Pancreas transplant patients were excluded. Information on patient demographics, disease severity, length of stay, use of insulin therapy and hypoglycaemia treatment administered (bolus/maintenance glucose-infusion, maintenance enteral/parenteral feed, reinforced oral intake, titration of insulin infusion) were collected and analysed using descriptive statistics. Hypothesis testing was performed with student’s t-test and chi-square testing as appropriate. Linear regression model was used to correlate glycaemia with treatment intervention.

Results: One hundred and sixty-three patients with 410 hypoglycaemic episodes in total were included in the study. Mean age was 56.6 (±17.4) years and 60.1% were male. The mean APACHE II score was 20 (±6). Reasons for admission were as follow: cardiorespiratory failure (32%), sepsis (20%), gastrointestinal problems (13%), post-procedure (10%), neurologic (9%), post cardiac-arrest (8%) and others (8%). 54.6% patients had recurrent hypoglycaemia, with 31.9% having ≥3 hypoglycaemic episodes during their ICU stay. The length of stay did not significantly affect frequency of hypoglycaemia (p = 0.055). Mean number of hypoglycaemic episodes was lower among patients on insulin infusion compared to those not on insulin (4 vs 6, p = 0.0015). Insulin infusion was continued in 15% cases of hypoglycaemia. Forty-two patients (10%) received no intervention and were not on maintenance glucose or feeds during the hypoglycaemic episode. Decision for intervention was not associated with the level of glycemia. The frequency of hypoglycaemia among patients who received treatment was higher compared to those without intervention (1.6 vs 0.4, p = 0.0237).

Conclusion: Over half of patients with hypoglycaemia developed recurrent episodes during their ICU stay. A significant proportion of patients were continued on insulin infusion and received no intervention for hypoglycaemia. More work is required to protocolise hypoglycaemia monitoring and management to ensure that timely intervention is delivered. Measures to inform physicians about ongoing insulin infusion in patients with hypoglycaemia may be helpful.

Intensive care unit resource use by patients with spinal cord injury

Muhanad Adam¹, Alexander Rouse², Tahir Ali¹ and Rajkumar Rajendram¹

¹Stoke Mandeville Hospital, Aylesbury, UK

²National Spinal Injuries Centre, Stoke Mandeville, UK

Abstract

Spinal cord injury (SCI) results in devastating functional impairment, significant morbidity and mortality.[1] However there are few reported data on the use of intensive care unit (ICU) resources by adults with spinal cord injury (SCI) in relation to outcomes such as survival to ICU discharge, rate of readmission to ICU prior to hospital discharge and survival to discharge home. These data were obtained from a service evaluation.

All adult patients who had SCI who were admitted to the ICU at Stoke Mandeville Hospital, Aylesbury, UK from 1^st January 2012 until 31^st December 2016 were identified from the ICU database. The dates of admission and discharge from ICU, mortality and discharge destination were also obtained from the database.

Year	Admissions	ICU Days N; mean ± SD	Readmissions (N)	Deaths ICU (N)/ Hospital (N)
2012	27	6.8 ± 8.1	5	6/7
2013	20	14.0 ± 20.4	2	2/4
2014	12	12.3 ± 19.2	1	2/2
2015	21	5.0 ± 7.0	2	2/2
2016	18	3.4 ± 3.1	2	1/2
Total	98	10.2 ± 15.7	12	13/17

Over the five-year period assessed 78 patients with SCI were admitted to ICU (98 admission episodes). The mean duration of ICU admission was 7.9 days (SD 13.0 days; range 1–89 days). Thirteen patients were readmitted after discharge from ICU. Two patients were readmitted twice, one was readmitted three times and another was readmitted four times. Seventeen patients died in hospital, of whom 13 died in ICU. Sixty one (78%) survived to hospital discharge (54 Home; 7 Nursing Home).

Previous data have suggested that the standardised mortality rate of patients with SCI is high, mainly due to secondary complications.[1] However, in our adult cohort with SCI survival to ICU discharge was good (86.7%) as was survival to hospital discharge (78%). To achieve these outcomes patients with SCI use a significant amount of ICU resources. Refinement of the management of patients with SCI is required to reduce the duration of ICU admission and improve survival to hospital discharge. We hope that dissemination of our experience will open discussion on how to improve outcomes in this particularly complex and vulnerable cohort.

Pop Up CPR – a quality improvement intervention to enhance CPR efficacy in Critical Care, Emergency Department, and Ambulance clinicians

Tom Evens¹ and Lisa Shennan²

¹Imperial College Healthcare NHS Trust, London, UK

²London Ambulance Service NHS Trust, London, UK

Abstract

Introduction: Early, effective delivery of cardiopulmonary resuscitation (CPR) is widely accepted to be a prime determinant of resuscitation success in Cardiac Arrest. Several components of CPR quality are associated with adequate coronary blood flow: Compression rate; compression depth; chest-wall release; ventilation rate and ventilation volume.¹

CPR for healthcare providers is typically taught and assessed using simple manikins on formal resuscitation courses, with recertification intervals spanning several years. However, evidence suggests that skills and knowledge deteriorate after 1–6 months without ongoing practice.²

The AHA recently identified Deliberate Practice, Spaced Learning and Feedback as important components of educational efficiency, noting that “it is possible for a learner to practice multiple times with no observable improvement in performance”.³ Studies of a mastery-based approach have shown improved quality of CPR after 2–4 hour blocks of training.⁴ However, given the very large number of clinical staff who may be required to provide CPR in an acute trust, there is a considerable service cost to releasing all staff for regular training periods of this length.

Intention: We intended to demonstrate a brief, meaningful training intervention which could be delivered to a large number of Emergency Department, Critical Care and Ambulance staff during their working day with minimal operational impact.

Methods: The London Ambulance Service training team brought two training ambulances to the ambulance ramps at Charing Cross and St Marys Hospitals. The team attended quarterly over nine months (Oct 2017 – April 2018). Each visit lasted four hours. The vehicles contained an instrumented manikin (Laerdal Advanced Skills Trainer, Laerdal, Stavanger, Norway), a monitor-defibrillator, and a report printer.

Participants undertook a three minute skill cycle whereby they delivered chest compressions and ventilations (bag-squeeze) to a patient in simulated Ventricular Fibrillation. A trainer provided standard airway positioning and mask seal.

Participants received direct feedback, and individual measures of hand position; compression rate, depth and release; compression fraction; and ventilation rate and volume. These measures combined to give an overall score, with each component weighted along an S-curve.

The first visit aligned with World Restart a Heart Day, and was supported by a social media and poster campaign. An inter-speciality competition (the RCEM cup) was used to drive interest.

Results: 161 operational staff received training over three four-hour sessions, with each intervention taking less than five minutes.

Overall Score performance was not related to seniority or service. Compression rate (104 bpm, 95% CI 97–110) and ventilation volume (475 ml 95% CI 452–498) were well delivered. Principal variation was in hand positioning, ventilation rate, and chest release.

20 staff attended two sessions. All improved their Overall Score between sessions (mean improvement 38.3%, p < 0.001).

RCEM cup winners: Round 1: London Ambulance; Round 2: Critical Care; Round 3: Emergency Department.

Conclusion: Pop-Up CPR allows a large number of staff to receive spaced learning with actionable feedback without impacting service delivery.

Secondary benefits included improved interspeciality understanding; and effective messaging that training doesn’t always require time away from work. Delivering training in a place and time that is convenient to staff demonstrates that th

Determining the Factors Influencing the Duration of Global Influenza Epidemics

David Cook^1,2, Ada Yan¹, Kylie Ainslie¹ and Steven Riley¹

¹Imperial College London, London, UK

²Section of Anaesthetics, Pain Medicine and Intensive Care, Imperial College London, London, UK

Abstract

Influenza is a highly infectious respiratory viral illness that causes substantial morbidity and mortality and significant demand on Intensive Care Units.^1,2,3 We analysed the World Health Organisation’s FluNet database in order to better understand global influenza dynamics and investigate factors that affect duration of epidemics.⁴ We believe this is the first analysis of multiple predictors on the duration of epidemics and will enable us to develop more accurate forecasting models and better guide implementation of public health strategies and resource allocation.

Influenza surveillance data was extracted for 130 countries spanning 8 seasons from 2010 to 2018. We defined epidemics as periods which began when the weekly incidence persisted above the season mean for three consecutive weeks and ended when the incidence fell below the season mean for 3 consecutive weeks. Historical climatic and population data was obtained for all 8 seasons. Differences between temperate and tropical regions and countries of different income levels were compared using Student’s t-test. Associations between the duration of epidemics and environmental and geographical predictors were analysed by Pearson Correlation. We built a statistical model to summarise epidemic duration using linear regression and forwards variable selection. Multicollinearity was assessed using variance importance factors and adjustments to the model made accordingly. Model fit was assessed using R-squared and Akaike information criteria. A sensitivity analysis was conducted to ensure model accuracy using partial correlation and mixed models with random effects. All data analysis was performed on R, https://www.R-project.org/.⁵

96 countries met our inclusion criteria and were included in the analysis. 54 (56%) countries originated from temperate regions, 42 (44%) from the tropics. 388 distinct seasons were included covering 33% of the world’s population. The mean duration of epidemics globally was 17.7 weeks. This differed between temperate and tropical countries where the mean duration was 16.4 weeks, and 20.1 weeks respectively (p < 0.001, 95% CI: (2.18, 5.18)). 61.53% of tropical countries had more than one epidemic a season comparted to just 13.75% of temperate countries. The strongest correlations between predictors and duration was seen with temperature and humidity (positively correlated) and latitude (negatively correlated). Our final regression model demonstrated increases in epidemic duration with larger populations, higher temperature and humidity and a greater difference in temperature extremes within seasons when adjusting for all other factors. An increase in latitude and country size is associated with shorter epidemics. Sensitivity analysis did not result in improvements to model fit.

The duration of influenza epidemics has significant consequences on public health strategy and demands on health care resources. Here were define factors that affect the duration of epidemics in order to develop better forecasting models and aid policy design in future.

Baseline NT-proBNP predicts outcome and treatment response to statin therapy in patients with ARDS

Peter Ferris¹, Andrew Boyle^1,2, John Conlon¹, Anthony C Gordon³, Manu Shankar-Hari^4,5, Cecilia O'Kane¹ and Danny McAuley^1,2

¹Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

²Regional Intensive Care Unit, Royal Victoria Hospital, Belfast, UK

³Section of Anaesthetics, Pain Medicine and Intensive Care, Imperial College London, London, UK

⁴School of Immunology & Microbial Sciences, Kings College London, London, UK

⁵Guy's and St Thomas' NHS Foundation Trust, London, UK

Abstract

Introduction: Prognostic enrichment (identifying patients with a higher risk of death), and predictive enrichment (identifying patients with traits which determine treatment responsiveness) are two emerging strategies that aim to identify a patient population recruited to clinical trials with the aim of increasing the probability of positive trials. Previous small studies suggest that B-Type Natriuretic Peptide (BNP), and its precursor N-terminal BNP (NT-proBNP), may discriminate between ARDS and non-cardiogenic pulmonary oedema. However, the importance of BNP in determining prognosis in patients with ARDS is uncertain.

Hypothesis: In patients with ARDS, baseline NT-proBNP is a prognostic biomarker in ARDS. Furthermore, the treatment effect of simvastatin will be greater in patients with high baseline NT-proBNP.

Methods: A post-hoc analysis of the HARP-2 clinical trial was undertaken. HARP-2 was a randomised, controlled clinical trial evaluating simvastatin in 540 patients with ARDS according to the American European Consensus Conference definition [1]. This definition attempts to exclude patients with hydrostatic pulmonary oedema. Patients were recruited within 48-hours of meeting ARDS criteria. Clinical data and baseline plasma samples obtained were analysed. The primary outcome was 28-day mortality. An enzyme-linked immunosorbent assay was used to measure NT-proBNP. High NT-proBNP was defined as a value above the median of the studied population.

Results: 507 patients recruited in HARP-2 where baseline samples were available were included, and of these 245 (48%) were allocated simvastatin. In patients allocated to placebo, there was no difference in baseline PaO₂/FiO₂ ratio between survivors and non-survivors, however more non-survivors were receiving vasopressors at baseline compared to survivors (79% vs. 60%; p = 0.004). In these patients, baseline NT-proBNP was higher in non-survivors compared to survivors at day 28 (Figure 1; p = 0.007). Of the 253 patients with high baseline NT-proBNP, 116 (46%) were assigned to simvastatin. Patients with high baseline NT-proBNP who received simvastatin had lower 28-day mortality compared to placebo (simvastatin 22% vs. placebo 37%; Figure 2; p = 0.01). Similarly, patients with high baseline NT-proBNP allocated to simvastatin had more ventilator-free days compared to placebo (14.0 [0.0–22.0 vs. 8.0 [0.0–21.0] days; p = 0.03). Finally, in patients with low baseline NT-proBNP there was no difference in 28-day mortality (simvastatin 20% vs. placebo 17%; p = 0.52) or ventilator-free days (simvastatin 15.0 [0.0–22.0], placebo 17.0 [0.0–23.0] days; p = 0.69) between groups.

Conclusion: In a post-hoc analysis of patients with ARDS recruited to the HARP-2 trial, baseline NT-proBNP was associated with 28-day mortality. In patients with high NT-proBNP at baseline clinical outcomes were improved in patients who received simvastatin. These findings suggest that for future clinical trials in ARDS NT-proBNP may provide predictive enrichment and offer prognostic enrichment for treatment responsiveness to simvastatin in ARDS.

Figure 1.

NT-proBNP and mortality at day 28 in patients allocated to placebo.

Figure 2.

Mortality in patients with high baseline NT-proBNP by treatment group.

A retrospective study into sepsis related mortality in the Aneurin Bevan Health Board

Olesya Godsafe¹, Tamas Szakmany², Thomas Cozens² and Paul Mizen²

¹Cardiff University, Cardiff, UK

²Royal Gwent Hospital, Aneurin Bevan Health Board, Newport, UK

Abstract

Background: Sepsis is a major cause of preventable mortality worldwide. The incidence of sepsis is increasing, our recent studies estimated its point-prevalence around 4–5% in the Welsh hospitals. (Szakmany et al. 2016; Szakmany et al. 2018)

Sepsis is a potentially reversible cause of death and therefore it is imperative that it is identified and treated early to avoid unnecessary loss of life. Currently, NHS Wales utilises the Sepsis 6 protocol that aims to treat within the hour of identifying septic patients. (Daniels et al. 2011) Previous national studies have identified a number of additional factors that help point towards a diagnosis of sepsis, these factors included increasing age, NEWS scores above 3 and comorbidities such as diabetes, Chronic Obstructive Pulmonary Disease (COPD), heart failure, and malignancy. (Szakmany et al. 2016) (Szakmany et al. 2018) Recently, the attributable fraction of mortality to sepsis has been questioned on the intensive care unit, with significant implications for future study design. (Shankar-Hari et al. 2018) The need for prospective data collection as opposed to code-based approach to help evaluating whether sepsis is amongst the leading causes of death has been highlighted. (Knoop et al. 2017)

The aim of the study was to identify patients who scored for sepsis and determine the number who died from sepsis related disease. The secondary aims were to identify significant comorbidities which positively correlate with poor outcomes to help risk stratify the patient group.

Methods: Information was obtained from an existing database of patients scoring for sepsis on the Sepsis Screening Tool in the Aneurin Bevan Health Board between January 2017 and April 2018. This data was collected prospectively as part of an ongoing quality improvement project, approved by the local R&D from electronic and paper notes, and the Sepsis Screening Toolkit database. Anonymised data was exported and analysed in SPSS. Patients were categorised into three groups: “sepsis related”, “non-sepsis related” and “possibly sepsis related” cause of death.

Results: Of the 476 patients, 178 (37.4%) patients died of sepsis, 112 (23.5%) were possibly sepsis related and 186 (39.1%) deaths were non-sepsis related. The patients who died from sepsis were older: 80 (78–82) vs 77 (75–79) years (p = 0.004). A higher NEWS score correlated with sepsis related death compared to non-sepsis related death (p = 0.001). White cell count (WC) and C-reactive protein (CRP) levels also appeared to be higher in the sepsis-related death group compared to non-sepsis related death (p = 0.003, p = 0.000 respectively). COPD appeared to be more frequent in those that died of non-sepsis (62) rather than sepsis related (40) causes (p = 0.009).

Discussion: Currently, there is little evidence regarding how many patients scoring for sepsis ultimately die due to the disease, and how many succumb to the cumulative effect of their comorbidities and age with clinical signs of a potential infection. Our data suggests, that even in a cohort with high likelihood of clinically significant infection, more than third of the patients who die do so irrespective of their infectious insult.

This study received ethics committee approval.

Emergency Intubation Bags: A Quality Improvement Project

Ross Little, Hannah Headon and Peter Bamford

Royal Liverpool University Hospital, Liverpool, UK

Abstract

There are currently five emergency intubation bags in use within the Royal Liverpool Hospital. It was noted that significant variation existed between bag contents, with no universal checking system in place across the trust. We set out to rationalise and standardise the contents of these bags, and to introduce a formal checking process similar to the tagging system employed on cardiac arrest trolleys. Using audit and trainee surveys pre- and post- changes, we aimed to demonstrate quantitative and qualitative improvements in practice, with the overall goal of improving patient safety.

Audits were performed to retrospectively examine the compliance of emergency intubation bag checks, and to determine the consistency of bag contents. An additional survey was conducted to determine trainee experiences of using and checking the bags.

The initial survey of trainee experiences identified that there had been a number of occasions where essential intubation drugs had been missing from the bags when they were required in emergency situations. This was due to bags not being restocked appropriately.

This correlated with findings from the initial audit, which identified that compliance of checking the bags varied from 57.9% to 93%. There were a number of times where bags had been checked but were found to be incomplete. Drugs missing included propofol, rocuronium and metaraminol.

Audit revealed that 25 different drugs were present across the five bags, with significant heterogeneity in the contents of every bag. From the survey, 75% of trainees reported that there was too much drug variety within the bags (making it more difficult to find key drugs in an emergency), and >90% respondents agreed the bags would be easier to check and restock if the number of drugs in the bags was reduced.

In response to these findings, the following interventions were made. We rationalised the contents of the bags to the absolute necessities, and standardised these across the trust. A tagging system was introduced to indicate a fully stocked bag, coupled with a logbook to provide a record of accountability for mandatory daily checks.

The audit and survey were then repeated to examine the impact of the intervention. 100% of post-intervention survey respondents reported no instances of emergency drugs being missing from the bags. 100% also indicated that the introduction of a tagging system was a positive and beneficial change in practice. There has also been widespread endorsement of the simplification of the bag contents with 100% trainees reporting it easier to find essential drugs and restock bags. Daily checks are now being carried out between 80–90% of the time, with the bags found to be complete without any missing drugs 100% of the time.

Overall there has been demonstrable improvement in compliance with restocking of emergency drug bags and an improvement in trainee experiences. We feel this has wider implications for patient safety ensuring that critical care doctors have the appropriate drugs immediately to hand in emergency situations.

Is there enough Phentolamine in your cupboard?

Rebecca Taylor, Priti Gandre

North Middlesex University Hospital, London, UK

Abstract

Phaeochromocytoma is a catecholamine secreting tumour of chromaffin cells of the adrenal medulla or the extra adrenal paraganglions. It is an uncommon cause of severe hypertension. We present a case involving management of a hypertensive crisis in a previously fit man with an undiagnosed phaeochromocytoma. The management was exceptionally challenging due to the atypical presentation and difficulty in obtaining and administering appropriate non-selective alpha blockers.

A 47-year-old male, presented to our Emergency Department with a blood pressure of 240/100 mmHg, chest pain and headache. Routine blood results and further clinical examination were unremarkable. An urgent CT angiogram revealed an adrenal tumour (3.9 cm) confirming the clinical diagnosis of phaeochromocytoma. Urgent treatment was commenced with oral phenoxybenzamine and intravenous phentolamine, for immediate control of blood pressure in view of the symptoms. Due to severe nausea and vomiting refractory to antiemetics, the phenoxybenzamine tablet could not be retained in the stomach long enough to ensure its absorption. The blood pressure remained controlled with intravenous infusion of phentolamine in ICU awaiting the nausea to subside. Up to 0.7 mg/min of the drug was needed to keep the systolic blood pressure below 200 mmHg.

Nausea continued to be difficult to control and multiple attempts at nasogastric tube insertion failed, despite sedation and analgesia. Each failed attempt led to further surges in blood pressure. The hospital stocked only 5 ampoules of phentolamine (total 50 mg). Frantic efforts to get hold of more stock from neighbouring hospitals, including the nearby tertiary centres, ensued without success. Intravenous infusion of labetalol was carefully titrated alongside phentolamine and magnesium, in an attempt to reduce the phentolamine requirement. Aortic dissection was feared since the patient complained of severe lower backache which worsened with time. He was too restless and hypertensive to perform another urgent scan to confirm this.

Within four hours, further stocks of phentolamine arrived from a hospital 100 miles away. These were sufficient to ‘buy time’ for the blood pressure and nausea to settle. Phenoxybenzamine was then administered orally and the patient was transferred safely to a tertiary centre the next day for further management. The backache was found to be due to haemorrhage into the tumour, which is an unusual finding. The patient successfully underwent elective excision of the phaeochromocytoma a few weeks after discharge from that hospital on high dose oral phenoxybenzamine.

In an 18-month period, we have seen three patients in our Emergency Department with hypertensive crisis as the first presentation of phaeochromocytoma. Learning from these unusual cases we have now increased our emergency stock of phentolamine six-fold to 30 ampoules (300 mg). This case highlights the practical problems that may arise when treating phaeochromocytoma crisis with oral phenoxybenzamine and alert readers to the limited stock of phentolamine that might be available on site. Without the sufficient control of the hypertensive crisis, supportive interventions such as intubation and clinical transfer to a tertiary centre can risk further complications, rather than being beneficial.

The Experiences of Newly Qualified Nurses in Critical Care: A Qualitative Systematic Review

Charlotte Elias^1,2 and Tina Day¹

¹King's College London, London, UK

²University of Gloucestershire, Gloucester, UK

Abstract

The aim of this qualitative systematic review was to explore the experiences of Newly Qualified Nurses (NQNs) working in Critical Care, as their first job post-registration. Additional objectives were to identify their lived experiences, to understand the unique challenges that they face and make recommendations based upon the findings, to support their transition from student to registered nurse in this environment. For the purposes of this review the term Critical Care is synonymous with the Intensive Care Unit (ICU).

Recent nursing shortages have been particularly felt in specialist areas, like Critical Care, which require higher staffing quotas. This has led to increased recruitment of NQNs, in an area traditionally reserved for more experienced nurses. Literature on the challenges faced by NQNs transitioning from student to registered nurse in the acute setting is ample, but little is known about their transition into Critical Care, a gap in the evidence addressed by this review.

A systematic search was undertaken using six primary databases, grey literature platforms and additional sources; yielding eight studies for inclusion in this review. A graphical representation of the search yield can be viewed a PRISMA flow diagram¹ (see Figure 1). Eligibility criteria included NQNs with less than 2 years post-registration experience, working in an adult ICU. Included studies were set in the UK and other developed countries, with comparable healthcare systems and resources. The studies were appraised using the Joanna Briggs Institute Critical Appraisal Checklist for Qualitative Research² and the Consolidated Criteria for Reporting Qualitative Research (COREQ) Checklist.³ The findings were synthesised using the Thematic Synthesis approach.⁴

Five main themes emerged from the data: ICU Readiness, An Emotional Spectrum, Developing Relationships, The Journey to Self-Satisfaction and ICU Commitment.

The five themes were developed into a conceptual framework (see Figure 2), which following further validation, may inform future practice. The ICU readiness of NQNs was developed through preparation by undergraduate education, addressing knowledge deficits and inexperience, supernumerary time and the ICU orientation process. NQNs experienced a spectrum of emotions, from excitement to fear; often experiencing mixed emotions during emergency situations. Developing positive relationships with colleagues, service-users and peers was important in helping NQNs feel a sense of belonging; however confrontations could also affect the NQN’s confidence. The journey to self-satisfaction comprised a period of self-reflection and development of critical thinking skills. ICU commitment was established once the NQN accepted themselves as an ‘ICU nurse’ and expressed a desire to stay.

Recommendations of this review include the provision of an undergraduate placement in Critical Care to enhance ICU readiness; simulated emergency scenarios to increase exposure to adverse events and help alleviate anxiety; consistency in mentorship and commencing NQN employment in cohorts, to facilitate peer support. This review does not provide sufficient evidence to suggest whether NQNs perceive Critical Care to be a suitable place to begin their career, further research is required in this area.

Figure 1.

PRISMA Flow Diagram.

Figure 2.

The Newly Qualified Nurses’ Journey in Critical Care.

Family Satisfaction in the ICU: A questionnaire-based project

Lenster Marshal, Sabahat Zaidi, Daniel Harvey, Charmaine Buss and Lauren Chambers

Queen's Medical Centre, Nottingham, UK

Abstract

Introduction: Family satisfaction in critical care reflects the extent to which perceived needs and expectations of family members of critically ill patients were met by healthcare professionals.¹ The satisfaction of family members with the information provided, and their engagement in shared, decision-making is being recognised as an important aspect of the quality of ICU care.² As a component of a new and novel family liaison role dedicated to improving these aspects of care, we sought to assess family satisfaction with our large ICU at the Queens Medical Centre, Nottingham University Hospitals.

Methods: Our Family Liaison nurse provides support to the patient and their families during and following an episode of a critical illness. To understand baseline needs, and the potential utility of this approach we conducted a survey of relatives utilizing the Family Satisfaction in the Intensive Care Unit questionnaire (FS-ICU-24).² The study was performed over a six-month period in our 21-bed ICU, using a convenience sample. Likert scale responses for each questionnaire item were summarised into percentage scores and pooled to give overall scores for the two domains; satisfaction with care and satisfaction with decision making. The questionnaire also collates additional comments which are feedback to the unit to identify strengths and areas for improvement.

Results: A total of 51 questionnaires were returned from a total of 73 that were distributed to family members giving a 68% response rate.

There were high scores throughout for family satisfaction with care and family satisfaction with decision making around care. Overall score for family satisfaction with care was reported as very good or excellent in 85%. Overall family satisfaction with the decision-making process were rated as very good or excellent in 81%.

62% felt at least somewhat included in the decision-making process and 67% felt that they have enough support throughout the process. 87% felt that they have had adequate time to have their concerns addressed and questions answered.

Family Satisfaction with Care	Number of Responses	%
Excellent or Completely satisfied	434	65
Very Good or Very satisfied	135	20
Good or Mostly satisfied	60	9
Fair or Slightly satisfied	31	5
Poor or Very dissatisfied	10	1
Total responses	670	100

Family Satisfaction with Decision-Making	Number of Responses	%
Excellent	170	58
Very Good	66	23
Good	35	12
Fair	17	6
Poor	3	1
Total Responses	291	100

Conclusion: A large majority of family members were highly satisfied with both the care provided for patients and the decision-making process.

This is a pilot questionnaire-based study to evaluate family satisfaction in our unit and therefore no prior study of family satisfaction exists for comparison. Nonetheless, this study is useful to highlight that there still is a gap in communication between healthcare professionals and families

The FS-ICU-24 is a robust, well-validated tool and was a useful addition in our large UK ICU for studying family satisfaction and perceptions of care. This project will be used to continually reassess our performance and provide an important source of learning as a critical care unit.

The use of hand held ultrasound devices to measure muscle architecture – a comparison to standard ultrasound using healthy volunteers

Peter Turton¹, Richard Hay¹ and Ingeborg Welters²

¹Royal Liverpool University Hospital, Liverpool, UK

²Institute of Aging and Chronic Disease, University of Liverpool, Liverpool, UK

Abstract

Muscle ultrasound is a commonly used research technique in studies examining muscle loss in intensive care. Hand held ultrasound devices could potentially be used in such studies, or as a clinical tool to detect early muscle wasting. A previous study measuring diaphragm thickness found that hand held devices perform well against standard ultrasound machines [1]. We sought to examine whether such a device could accurately measure indices of lower limb muscle architecture, in comparison to a standard ultrasound device.

Healthy male volunteers (n = 11) were recruited to attend the Gait Lab at the University of Liverpool. Using a standard ultrasound device (Telemed EchoBlaster 128), images were taken from the right thigh to measure the total muscle thickness of the anterior thigh (rectus femoris and vastus intermedius, RFVI), and the muscle thickness (MT) and fascicle pennation angle (PA) fromvastus lateralis (VL). Using trigonometry, we dervied the fascicle length (FL) from MT and PA. This was then repeated using the hand held device (V-Scan Dual Probe).

ICC estimates and their 95% confident intervals were calculated using SPSS statistical package version 23 (SPSS Inc, Chicago, IL) based on a mean-rating (k = 2), absolute-agreement, 2-way mixed-effects model, and are shown in the table below.

Bland-Altman plots were constructed for all measured variables (VL MT and PA, RFVI MT), and the one dervied variable (FL). One-sample t-tests for the difference in observed measurements between each device were performed, and no significant differences were detected. Linear regression showed that there was no evidence of proportional bias in any of the measured or derived variables. Thickness of VL and RFVI were measured offlline twice, with the second set of measurements performed by a blinded investigator who had taken no part in the scanning. Inter-rater agreement between the blinded investigator and the first investigator was also high in both muscles and in both devices (VL MT: 0.991 (Telemed), 0.994 (V-Scan); RFVI MT: 0.985 (Telemed), 0.991 (V-Scan), p < 0.001 in all cases).

ICC Estimates

Measurement	ICC	95% Confidence Interval	P-value
VL MT	0.942	0.783–0.984	<0.001
VL PA	0.995	0.983–0.999	0.001
VL FL	0.979	0.922–0.994	<0.001
RFVI MT	0.880	0.555–0.968	0.001

This small study in healthy volunteers demonstrates that the V-Scan Dual Probe hand held device, using pre-determined settings, performs as well as a standard ultrasound machine for the purposes of measuring and deriving indices of muscle architecture. Further work is required to assess if similar results can be obtained in critically ill patients, but this may give an alternative and convenient method of point of care ultrasound for intensive care research studies.

A survey of current attitudes to diaphragm ultrasound in the prediction of successful respiratory weaning

Peter Turton and Ingeborg Welters

Institute of Aging and Chronic Disease, University of Liverpool, Liverpool, UK

Abstract

There is an extensive body of literature examining how diaphragm ultrasound could be utilised to predict successful liberation from mechanical ventilation, so much so that in the last two years, there have been two systematic reviews [1][2] and one meta-analysis [3] performed on the subject. We wished to survey the current attitudes to this technique, to understand if it has been adopted since the publication of these reviews.

An online survey was created and distributed via social media and e-mail. The survey was open to any trainee or consultant in intensive care medicine, and to any interested allied health professionals, in any country. Questions within the survey assessed current use of ultrasound, whether they were using ultrasound to assess likelihood of successful extubation, and whether they would be interested in further training in the topic.

We received 41 responses from 13 consultants, 18 trainees. and 10 AHPs. All Health Education regions of the UK provided at least one response to the survey. Whislt 64% of respondents agreed or strongly agreed that they could visualise the diaphragm using ultrasound, only 20% felt they could assess thickness or mobility. 53% said that they were not aware of the current literature on using diaphragm ultrasound to predict successful weaning. No respondents were routinely using diaphragm ultrasound to assess if a patient was ready for a spontaneous breathing trial, nor was anybody using it to assess if they were ready to be extubated. 78% of respondents were however interested in learning more about the technique to potentially use it in a clinical setting.

Whilst all respondants said that their units possessed a linear probe, and 98% said they had a curved probe (to assess thickness and mobility, respectively), only 82% had a machine that could perform M-mode ultrasonography, which is usually required for assesment of mobility.

This small survey suggests that knowledge and usage of the current literature on using diaphragm ultrasound to assist with weaning is poor, but there is interest amongst the survey respondents to learn more about the technique. Most respondents have the neccessary equipment in their units to learn and perform the technique.

A qualitative synthesis regarding the factors related to recruitment to critical care trialsacross England and Wales

Natalie Pattison¹, Nishkantha Arulkumaran², Sally Humphreys³, TIm Walsh⁴, Philip Hopkins⁵ and Paul Dark⁶

¹Uni of Herts/East & North Herts NHS Trust, Hertfordshire, UK

²Critical Care Department, University College London Hospital, London, UK

³West Suffolk NHS Foundation Trust, Bury St Edmunds, UK

⁴Royal Infirmary of Edinburgh, Edinburgh, UK

⁵Kings College Hospital, London, UK

⁶University of Manchester, Manchester, UK

Abstract

Background: Conducting clinical trials in critical care is integral to improving patient care. There are unique practical and ethical challenges in this patient population that make patient recruitment challenging, including narrow recruitment timeframes and obtaining patient consent [1]. Units currently vary significantly in their ability to recruit according to their infrastructure and level of research activity.

Aim: Our primary objective was to identify variability in the research infrastructure of UK intensive care units (ICUs) and their ability to recruit patients into clinical trials.

Design: We evaluated factors related to intensive care patient enrolment into clinical trials in the UK. This consisted of a qualitative synthesis [2] carried out with two datasets of in-depth interviews conducted with 27 intensive care consultants, research nurses and trial coordinators from across the UK. Primary and secondary analysis of two datasets was undertaken in the thematic analysis.

Findings: The synthesis yielded the following themes: Organisational, Human, Study, Practical resources, Clinician, and Patient/family factors. The overarching core theme of Normalising Research was characterised by motivations for promoting research and fostering a research-active culture within resource constraints. There was a strong sense of integrating research in routine practice.

Conclusions: This qualitative synthesis yielded the concept of Normalising Research; the importance of fostering a culture where research is seen as part of routine patient care and clinical practice. This is required for everyone working from the organisational level to direct patient contact.

A feasibility study of a randomised intervention to enhance sleep in the ICU using aromatherapy massage

Natalie Pattison¹, Geraldine O'Gara², Jeannie Dyer², Karen Thomas² and TIm Wigmore²

¹Uni of Herts/East & North Herts NHS Trust, Hertfordshire, UK

²Royal Marsden NHS Foundation Trust, London, UK

Abstract

Background: Sleep in ICU is known to be problematic and can lead to adverse consequences, including delirium and protracted recovery. Massage to enhance onset and quality of sleep is one solution put forward in the literature, although little evidence on efficacy exists.

Aim: To assess the feasibility of conducting a randomised clinical trial exploring the effect of aromatherapy massage on sleep in critically ill patients.

Methods: A randomised feasibility trial was conducted in a large cancer hospital in the UK. Patients were randomised to receive aromatherapy massage or usual care. Depth of sleep was monitored using Bispectral Index monitoring. The sample included critically ill patients who were likely to stay in ICU for more than 3 days. Feasibility outcomes included: achieving recruitment targets (n = 50), drop-out rates, ability to collect outcome data. Secondary exploratory outcomes included the self/nurse-report Richards-Campbell Sleep Questionnaire, and the Sleep in the ICU Questionnaire. Patients also reported on the massage experience.

Results: 34 patients consented to participate, 17 (50%) were randomised to the intervention arm (aromatherapy massage) and 17 to the control arm. 7 (41%) received the intervention, 12 (35%) were withdrawn due to change in clinical condition or early discharge, 9 (52.9%) completed questionnaire data, 9 (26.4%) tolerated Bispectral index monitoring. All patients who received the intervention reported enjoying it. Inclusion criteria were challenging, demonstrating it would be unfeasible to conduct a larger study in this population. Outcome measures on the questionnaires were completed, however Bispectral index monitoring was unfeasible due to frequent missing values.

Conclusion: While the principles of conducting a study in this population appeared reasonable, the feasibility study demonstrated it would be unfeasible to conduct a larger study. This study revealed important limitations in such studies, highlighting the value of piloting complex interventions.

Prospective audit of direct return to ward policy for low risk hepatectomy patients

Behrad Baharlo¹, Usman Nizamani², Robert Sutcliffe², Rachel Moore¹, Davinia Bennett¹ and Catherine Snelson¹

¹University Hospitals Birmingham National Health Service Foundation Trust, Department of Anaesthesia and Intensive Care Medicine, Queen Elizabeth Hospital Birmingham, Birmingham, UK

²University Hospitals Birmingham NHS Foundation Trust, Liver and Hepato-Pancreato-Biliary Unit, Queen Elizabeth Hospital Birmingham, Birmingham, UK

Abstract

Background: Patients who undergo hepatectomy are typically transferred to a high dependency or critical care unit (CCU) for invasive monitoring after surgery. However, critical care is an expensive and limited resource accommodating for various clinical burdens.¹ Data from our unit suggests that the probability of a major complication or need for organ support within 24 hours after hepatectomy is low, and may be predicted by patient age and extent of hepatectomy.² A pathway was developed in which low risk patients were identified and transferred directly to an acute surgical ward post-hepatectomy. The aim of this study was to evaluate the safety of the direct return to ward pathway.

Patients and Methods: All patients identified as suitable for a direct return to ward (RTW) pathway were entered into a prospective database over a six-month period (Feb-July 2018). Inclusion criteria were (1) less than 70 years old and (2) resection of up to two liver segments. Patients with medical comorbidity were excluded. Short-term patient outcomes were evaluated prospectively.

Results: During the study period, 26 patients were considered suitable for the RTW pathway. However, only 11 patients were successfully transferred to the ward after surgery. 15 patients were transferred to CCU immediately postoperatively due to logistics (CCU bed being available). Although the CCU cohort were older than the RTW cohort (62 vs. 51 years; p = 0.002), there was no differences in operative approach, mode of analgesia, operating time, or acid-base values at the end of surgery. Escalation of care (anaesthetic input) was required for 1 RTW patient (9%) and 5 CCU patients (19%) for pain control only. One CCU patient required vasopressor support. For CCU patients, the median times for clinical decision to step down and actual time of ward transfer were 15 (range 11–64) and 40 hours (18–163), respectively. The median hospital stay was significantly longer for CCU patients compared to RTW patients (8 vs. 5 days; p = 0.029).

Conclusion: For low-risk hepatectomy patients, a return to ward pathway can be implemented safely and may improve patient recovery, shortening hospital stay and potentially reducing costs. Further work is required to embed the pathway, and to explore the option of expanding the indications without compromising patient safety.

Predictors of survival following cardiac arrest in a British district general hospital

James Heggie and Michael Reay

The Dudley Group NHS Foundation Trust, Dudley, UK

Abstract

Out-of-hospital cardiac arrest is a common occurrence in the UK. The ambulance service attempts cardiopulmonary resuscitation (CPR) in approximately 46% of pulseless patients. Of these, less than 1 in 10 will survive to hospital discharge(1). Variables suggested to affect survival include bystander CPR(2) use of an AED(3) (automated external defibrillator), time to start CPR(4) (no flow time), duration of resuscitation (low flow time) and presenting rhythm.

Russells Hall Hospital is a District General Hospital in Dudley in the UK. The Intensive Care Unit (ICU) frequently deals with immediate survivors of cardiac arrest. We investigated which variables affected survival and whether our experience with cardiac arrest survivors matched that described in the literature.

We performed a retrospective cohort analysis and service evaluation of all patients admitted to the Intensive Care unit in the years 2013 to 2017 (the duration of our electronic record) who had received CPR within the 24 hours prior to admission whilst not admitted to the hospital. Pre-admission events were obtained by locating electronic emergency department records and ambulance sheets. Data regarding their post-ICU hospital course was obtained by manually reviewed the paper record. The General practitioners of survivors were contacted for information regarding health status following discharge.

125 cases were identified. 61% were male. The mean age was 62 (SD 16). Overall 48% survived to discharge from the ICU. 34% survived to hospital discharge. Data of survival following discharge is pending. 34% had an initial rhythm that was shockable. 25% were in pulseless electrical activity and 21% had asystole. Rhythm data was absent in 13%.

Ambulance timings were available in 34% of cases. The mean time to arrival on scene was 8 minutes and 25 seconds. 45% received bystander CPR. An AED was used in 8% of cases. Mean no-flow time was 2.7 minutes but was unknown in 25% of cases. Mean low-flow time was 19.3 minutes.

The change of surviving to ICU discharge was 76% (95%CI 61–87%) for shockable rhythms, 42% (26–59%) for PEA, and 8% (1–25%) for asystole. The chance of survival to hospital discharge was 57% (42–71%) for shockable rhythms, 23% (11–40%) for PEA, and 4% (0–21%) for asystole. 52% (40–64%) of patients surviving to ICU discharge had a shockable presenting rhythm. 15% (8–26%) of non-survivors had a shockable rhythm. No-flow time did not differ between groups being 2.6 minutes for ICU survivors and 2.8 minutes for non-survivors. Low-flow time was 14 minutes (12–17mins) for survivors, and 24 minutes for non-survivors (20–28mins).

For patients surviving to reach ICU the prognosis remains poor but is significantly better than quoted for unselected patients. This suggests our patient selection is not unreasonable. Presenting rhythm and low-flow time appear to be the most significant predictors of survival in our population. Use of an AED showed a trend towards significance but did not reach threshold. No-flow time, and the presence of absence of bystander CPR was not significant in our analysis. This data will be used to inform conversations with relatives of patients suffering this devastating event.

Severe acute hyponatraemia following hysteroscopy – a case report

Li Lin Hong¹ and Ben Holst²

¹University Hospital of Wales, Cardiff, UK

²University Hospital Wales, Cardiff, UK

Abstract

Introduction: Severe acute hyponatraemia from irrigation fluid intravasation during hysteroscopy is a rare complication with potentially life-threatening sequalae.

Case report: We report on a 50 year old patient, American Society of Anaesthesiology physical status II, who underwent hysteroscopic transcervical resection of endometrium for menorrhagia. Preoperative tests showed normal urea and electrolytes. Anaesthesia was induced intravenously, a supraglottic airway inserted, and general anaesthesia maintained with sevoflurane. Around 30 minutes intraoperatively, a large irrigation fluid deficit was noted – 6L of 1.5% glycine solution had been administered, with 2L suctioned and 400 ml collected in drapes, corresponding to a disrepancy of 3.6 L. Cardiopulmonary signs including oxygen saturation, ventilation pressures, blood pressure and heart rate were stable. However the patient was assessed to have periorbital and facial oedema. The procedure was terminated, and endotracheal intubation performed, during which significant laryngeal oedema was noted. Blood gas analysis revealed severe hyponatraemia (Na 110 mmol/L). As the fluid deficit developed so rapidly, intraperitoneal infusion via the Fallopian tubes was suspected. Hence laparoscopy was performed, which revealed minimal intraperitoneal fluid.

The patient was transferred to the intensive care unit for further management. Hyponatraemia was treated initially with two 100 ml boluses of 5% sodium chloride (NaCl), followed by isotonic 0.9% NaCl. Serial blood gas and electrolyte analyses revealed a rapid improvement in serum sodium to 126 mmol/L after 3 hours. At postoperative day 1, serum Na was 128 mmol/L. The patient was extubated, and neurological assessment revealed no deficit. At the second postoperative day, serum Na was within the normal range, and the patient was discharged.

Discussion: A non-conductive solution is required to facilitate monopolar electrosurgery during hysteroscopy. 1.5% glycine, a hypoosmotic solution (200mOsm/L), is commonly used. Glycine absorption can lead to hyponatraemia and fluid overload, a complication that occurs in up to 0.2% of hysteroscopic procedures.¹ Clinical manifestations occur as a result of cerebral and pulmonary oedema, and cardiovascular collapse. However, general anaesthesia may mask neurological symptoms, and cardiopulmonary compromise may not be evident in healthy patients even after significant fluid absorption, making early clinical detection of this complication difficult.

Acute hyponatraemia carries a mortality rate of up to 40%. Treatment options include hypertonic saline, loop diuretics and haemodialysis. The risk of osmotic demyelination syndrome (ODS) from rapid increases serum sodium is thought to be low in this context. However, the rate of correction of serum sodium remains a subject of debate. European (ESICM) guidelines recommend an initial increase in serum sodium by 5 mmol/L using boluses of hypertonic 3% NaCl, followed by a limit of further increases to 10 mmol/L in 24 hours.² In contrast, a United States expert panel recommends that no restriction is required for the rate of correction of sodium in acute hyponatraemia.³ To date, there are no reports of ODS developing from rapid correction of serum sodium in acute hyponatraemia. In our patient, serum sodium was increased very rapidly by 16 mmol/L within 3 hours, with no neurological consequences. This case adds to the growing body of evidence supporting rapid correction of serum sodium in acute hyponatraemia.

Insertion of tracheostomy practices in polytrauma patients from a major trauma centre cohort – a review of current practice

Joanna Davy, Peter Russell, Tunde Szegi and Ahilanandan Dushianthan

University Hospitals Southampton NHSFT, Southampton, UK

Abstract

Polytrauma is often associated with severe injuries and prolonged ventilation. Consequently, patients may develop complications related to prolonged sedation, health care associated infections, and acquired neuromuscular weakness. Although early tracheostomy may reduce sedation requirements and duration of ventilation, mortality benefit is not substantiated [1]. Preference for tracheostomy weaning and its ideal timing is not fully defined. We reviewed our tracheostomy practices to identify trauma patients at-risk of prolonged ventilation.

A retrospective review of trauma patients admitted to the Southampton regional trauma centre from 01/01/2016 to 30/03/2018 was performed. Data obtained from TARN database, CIS system and electronic records. Comparisons analysed by Student’s T-test on GraphPad Prism 7.0.

337 trauma patients were admitted during the period evaluated. 197 patients (58%) were intubated (M: F ratio of 2.9:1 – mean age = 50 years). Mean Glasgow Coma Scale (GCS) at presentation was 12 with an Injury Severity Score (ISS) of 27. On average patients were ventilated for 7 days with associated ICU length of stay (LOS) of 11 days. 12% of intubated patients were ventilated for ≥10 days and 5% had a tracheostomy. There were no significant differences in age, gender, GCS at presentation between all ventilated patients. However, higher ISS score and increased number of surgeries were associated with tracheostomy and prolonged ventilation (Table 1). In general, tracheostomy was performed late (mean 18 days range 13–26 days).

Polytrauma is associated with prolonged ventilation and hospital stay. Severity of injuries and/or the number of surgeries patients require may be associated with this. Early identification of such patients may prevent prolonged ventilation and associated complications. Further targeted studies are essential to explore the benefits of early tracheostomy in this patient group.

An institutional review of prone position ventilation

Ashleigh Taylor, Christopher Kelly and Robert Docking

Queen Elizabeth University Hospital, Glasgow, UK

Abstract

Introduction: Acute respiratory distress syndrome (ARDS) has a spectrum of severity with variable incidence in UK intensive care units estimated at 64/100,000 (1). Despite a predicted mortality of up to 45% for severe ARDS](1) treatment options are limited. Lung protective ventilation with tidal volumes <6 ml/kg, airway pressures <30 cmH₂0 and permissive hypercarbia in combination with optimal fluid balance and neuromuscular blockade is advocated. Prone position ventilation has been used since 1970, (2) and following the PROSEVA trial (3) which showed a 51% relative reduction in mortality with its use has become more established in modern practice.

Background: Prone ventilation is widely accepted as forming best practice; but as a technique can be labour intensive and potentially dangerous for the patient. As such it may not be performed at the appropriate opportunity. It requires planning, ensuring appropriately skilled staff are present with relevant monitoring and equipment to minimise risk of injury to patient or staff. As part of our institution’s severe respiratory failure protocol we suggest using prone position when the PF ratio is <13.3 kPa.

Methods: This is a retrospective cohort study based in a tertiary intensive care unit (ICU) with 18 Level 3 beds and an average of 800 ventilated admissions per annum. Over a 2-year period patients with severe ARDS as defined by PF ratio <13.3 kPa (n = 263) was analysed and data collected for those who underwent prone ventilation (n = 30). This represents 11.4% of potential patients. Data was collected from Carevue and Wardwatcher and analysed using Microsoft Excel.

Results: Of the 30 patients 12 died in ICU (40%). All ICU survivors were alive at hospital discharge, and at one year 10 were alive (66% mortality). Of the 18 hospital survivors, 5 were re-admitted to hospital within 6 months of discharge. There was a statistically significant increase in PF ratio after proning patients (P < 0.05).

Table 1.

Characteristics of all intubated trauma patients.

	All intubated patients	Patients intubated <10 days	Patients intubated ≥10 days	Tracheostomised patients
Number of patients (% of total)	197 (58%)	139 (41%)	41 (12%)	17 (5%)
Age - years (±SD)	50 ± 20	48 ± 20	54 ± 19	52 ± 20
Gender (M:F)	2.9: 1	2.76: 1	3.1: 1	3.3: 1
GCS at presentation (±SD)	12 ± 4	12.3 ± 4.1	12.4 ± 4	11.5 ± 4.3
ISS score	27.3 ± 13.2	25.1 ± 12.3	32.9 ± 13.9*	32.8 ± 14.7$
Total number of surgeries (number ± SD)	2.4 ± 2.0	2.1 ± 1.9	3.1 ± 2.1*	3.7 ± 2.3$
Ventilation duration (days ± SD)	6.8 ± 7.2	3.0 ± 2.3	14.2 ± 5.3*	19.6 ± 8.8$
ICU LOS (days ± SD)	11 ± 9.9	5.9 ± 5.7	18.8 ± 6.7*	28.8 ± 10.3$
Hospital LOS (days ± SD)	25.8 ± 31.1	21.7 ± 34.1	32.8 ± 17.4*	42.2 ± 22.8$
Predicted probability of survival (PS) (%)	84.4 ± 22.4	85.5 ± 21.4	81.9 ± 25.3	82.1 ± 23.7
Actual hospital survival (%)	88%	88%	85%	100%

PS: Predicted Survival; *, P < 0.05 for comparison between patients intubated <10days and ≥10days. $, P < 0.05 for comparison between tracheostomised patients and patients intubated for <10days.

	SURVIVORS (N = 18)	NON-SURVIVORS (N = 12)
Age	52	46
SEX	61% MALE	58% MALE
PRIMARY DIAGNOSIS	72% PULMONARY ARDS	100% PULMONARY ARDS
APACHE II	Median 19 (IQR 15–27)	Median 23 (IQR 14–31)
PREDICTED MORTALITY	Median 33% (IQR 21–60%)	Median 50.9% (IQR 16–78%)
USE OF NEUROMUSCULAR BLOCKADE	94%	92%
VASOACTIVE SUPPORT	90%	67%
RENAL REPLACEMENT THERAPY	17%	8%
STEROIDS	50%	75%
FLUID BALANCE PRE-PRONE	Median +780 ml (IQR −32–2951)	+1402 ml (IQR 215-2853)
TIDAL VOLUME	Median 6.3 ml/kg (IQR 5.7–6.8)	Median 6.7 ml/kg (IQR 5–6.6)
PEAK AIRWAY PRESSURE	Median 30 CMH20 (IQR 28–31)	Median 30 CMH20 (IQR 28–35)
TIME FROM PF <13.3KPA TO PRONE	Median 11 hrs (IQR 4–28)	Median 11.2 hrs (IQR 5–32)
COMPLICATION INCIDENCE	16.6%	25%

Conclusions: We have studied a cohort of critically hypoxic patients undergoing prone position ventilation. Prone positioning is safe when performed by appropriately trained staff. There was an increase in all post-proning PF ratios, but no significant difference between survivors and non-survivors. None of the complications were deemed clinically serious. Despite local guidelines we only used prone ventilation in 11.4% of potentially appropriate patients. These findings support ongoing staff education in proning and increased utilisation in line with our severe respiratory failure guideline.

Intubation Drug Checklist: Reducing stress for out of theatre intubations

Amanda Skingle and Anwen Williams

Wrexham Maelor Hospital, Wrexham, UK

Abstract

Out of theatre intubation is a common occurrence in hospitals throughout the country. It is well documented that out of theatre intubation has a higher complication rate with regards to adverse events and potential difficult airways and there are national guidelines available to minimise this risk of adverse events. Most of these guidelines focus on the availability of intubation checklists ensuring airway equipment, personnel and planning for difficult airways prior to intubation have been planned for and arranged.

Intubation drugs are a part of an intubation checklist however are often summarised by a box which simply says drugs are available or that drugs from categories such as induction agents are available, but don’t specify which drugs are available.

Out of theatre intubations can by their nature, occur anywhere in the hospital, but more commonly occur in the emergency department (ED) or one of the acute medical wards. Staff performing these intubations are called from the intensive care or anaesthetic department and are often the solo representation from their speciality. In situations requiring out of theatre intubations such as trauma cases, the priority is often airway management and stabilising the patient before intubation drugs can be drawn up, delaying the intubation process.

Nursing staff asked to sign out controlled drugs and collect drugs required for intubation are often unfamiliar with the drugs requested and don’t know how to prepare drugs before administration. Following a survey of ED nurses, 90% responded they weren’t confident with drugs requested for intubation and 95% said they didn’t know how to prepare the drugs.

In order to reduce the stress for both the intubator and ward nurses, an out of theatre intubation drug checklist was designed.

This checklist comprised of a clear table with common intubation drugs, split into categories of induction agents, muscle relaxants, opioids, benzodiazepines, emergency drugs and others. Each category was colour coded with the same colour as the corresponding drug syringe sticker used everyday in theatres. This had the aim to make labelling the syringes easier, as in an emergency, colours are easier to match than unfamiliar words.

In addition to the drug name, the dose and preparation was also included and instructions provided for the staff of how to dilute the medication and in what solution. Included in the table was a column to tick if the drug was requested and a further column to tick and sign when the drug was prepared. This was to ensure that in an emergency situation with lots of information being conveyed, no drugs requested were forgotten.

The checklists were distributed throughout the hospital in common areas they might be used, such as the ED and a supply was also added to an intubation trolley taken to wards by anaesthetic nurses when they assist at out of theatre intubations.

Provisional feedback has shown 80% of ED nurses are now confident with preparing drugs for intubation and 100% of trainees find the checklist useful and reduce the stress of out of theatre intubations.

The Introduction of a Family Liaison Lead Nurse in Adult Critical Care at Nottingham University Hospitals

Charmaine Buss, Thearina De Beer, Rebecca Selwyn and Daniel Harvey

Nottingham University Hospitals, Nottingham, UK

Abstract

Admission to Critical Care often is frequently unplanned and unexpected. This results in a highly stressful experience for patients and their relatives, especially when the outcome is uncertain, or likely to be poor. Families have to manage their life outside of critical care, resulting in caregiver burden and this impacts their ability to understand medical information (1).

Critical Care is a busy environment and clinical staff must prioritise therapy and patient safety. We hypothesised that patient and family expectations combined with complex decision making and communication makes responding to family needs difficult to achieve reliably by clinical staff, and that a role dedicated to family support would ease the stress and anxiety experienced by relatives of critically ill patients.

For a one-year period, we implemented a “Family Liaison Lead (FLL)” nurse to work with relatives in Critical Care, with the vision “Patients and families within Critical Care are given resources and support to empower them to cope with the short and long term psychological impact of critical illness”.

The FLL aims to speak to all relatives within Critical Care, but focuses on identifying patients and relatives that may require additional support.

Consultations with relatives took place independently at the bedside, or in collaboration with the medical team for planned discussions. This has ensured concerns can be addressed early to improve the experience for relatives and reduce workload for staff.

The FLL updates the medical and nursing teams of any relevant issues raised. They facilitate and encourage planning, prior to discussions with relatives to and reduce consultation time and improve the quality of communication. By utilising the FLL, more staff time can be committed to clinical care.

We assessed family satisfaction with their overall experience in critical care by using the FS-ICU survey (2) throughout. This is a widely validated tool to assess family satisfaction in relation to care and decision making (3).

Feedback on the FLL from patients, relatives and staff has been overwhelmingly positive. One family stated “…she was my constant throughout this whole nightmare and I don’t think mentally I would have coped without her… her role is so important for family members, someone to be able to talk to, to let off steam without any judgement.”

Further exploration of the impact of the FLL within Critical Care is required and there are opportunities to expand this role. These include mediation between relatives, the multi-disciplinary team, patients and the clinical team to ensure the best outcome in decision making. There is also an opportunity to utilise the FLL in the care and support of longer term patients, as part of their rehabilitation.

The introduction of the FLL in Critical Care has been successful and we plan to develop the role and metrics to assess its impact. Whilst the long- term psychological impact of the FLL role on critical illness for patients and relatives is difficult to measure, we hypothesise that the overall care and communication experience is enhanced (4).

Functional outcomes following cardiopulmonary resuscitation in a British district general hospital

James Heggie and Michael Reay

The Dudley Group NHS Foundation Trust, Dudley, UK

Abstract

It is estimated that less than 1 in 10 patients suffering an out-of-hospital cardiac arrest will survive to hospital discharge(1). Russells Hall Hospital is a District General Hospital in Dudley in the UK. The Intensive Care Unit (ICU) frequently deals with immediate survivors of cardiac arrest. We evaluated outcomes in patients surviving to receive ICU care in order to better counsel patients and relatives suffering this devastating event.

We performed a retrospective cohort analysis and service evaluation of all patients admitted to the Intensive Care unit in the years 2013 to 2017 (the duration of our electronic record) who had received CPR within the 24 hours prior to admission whilst not admitted to the hospital. Pre-admission events were obtained by locating electronic emergency department records and ambulance sheets. Data regarding their post-ICU hospital course was obtained by manually reviewed the paper record. The General practitioners of survivors were contacted for information regarding health status following discharge.

125 patients were identified. 60 patients (48%) survived to be discharged from ICU. 42 patients (34%) survived to be discharged from the hospital. Of the 18 patients who died in hospital after discharge from ICU, 14 (78%) had been identified as dying prior to discharge and received palliative care elsewhere in the hospital. Of these patients, 8 (57%) were seen by the specialist palliative care team. Data regarding long-term outcomes is still being collected.

The mean age was 62. For survivors to hospital discharge it was 56. 61% of patients were male; 64% of survivors were male. The median length of hospital stay for survivors was 12 days (range 3–129).

For patients surviving to hospital discharge, 26 (62%) had a Glasgow coma score (GCS) or 14 or better 72 hours after their cardiac arrest. 5 case (12%) had a GCS of 3 at this point. The remaining cases were evenly distributed between scores 4–13.

Survivors frequently required specialist input from cardiology (40%), psychiatry (17%) or the drug and alcohol liaison team (10%).

24 patients (57%) were able to walk by the time of hospital discharge. 20 of these (48%) were able to do so independently.

In the subset of patients who survive resuscitation and reach ICU, the outlook remains poor with only 34% surviving to discharge. Survival to ICU discharge was not a helpful outcome as a further 14% of patients died subsequently. However, most (78%) of these patients were identified in ICU. In the subset of patients who survive ICU and are discharged without palliative intent, the prognosis is reasonably good with 81% surviving to hospital discharge. Of these, over half were able to walk, and around half were able to do so independently. Such patients and their family members can be advised to be cautiously optimistic when leaving the ICU. The small numbers of patients who had a low GCS at 72 hours and yet survived merit further investigation.

Pneumatosis Coli between emergent surgery and successful conservative management in critical care

Amr Youssef¹, Vamshi Jagadesham² and Ingi Elsayed¹

¹University Hospitals of North Midlands, Dept. Critical Care, Stoke-on-Trent, UK

²University Hospitals of North Midlans, Dept. General Surgery, Stoke-on-Trent, UK

Abstract

Pneumatosis intestinalis (PI) is defined as presence of extra-luminal bowel gas that is confined within bowel wall. The small intestine is most commonly involved (42%), followed by colon (36%), with involvement of both in 22%.¹ There is a wide spectrum of causes of PI ranging from benign to life-threatening. PI may be caused by bowel ischemia, mechanical trauma, inflammatory/autoimmune bowel disease, intestinal neoplasms, bowel infection, obstructive pulmonary disease, or drug-induced, including immunosuppression therapy^1–2. Bowel necrosis requiring surgery was traditionally predicted by the presence of acute abdomen, lactic acidemia, an raised amylase, and presence of CT findings such as dilated bowel and portal or portomesenteric venous gas which is strongly associated with bowel ischemia and poor outcome ^3–4. Algorithms guiding management of patients presenting with PI; particularly those with extensive portal venous gas (PVG) have been proposed.⁵ These algorithms would suggest emergency laparotomies for symptomatic patients with extensive PVG; with or without evidence of lactic academia; as these findings were highly suggestive of acute mesenteric ischemia. Here we describe a critically ill elderly patient, referred to us with acute abdomen, widespread small bowel PI & extensive PVG, treated conservatively and with successful outcome, on our critical care unit. He was a 75 year old gentleman, found to have T3N0M0 distal oesophagus adenocarcinoma, for which he received neoadjuvant chemotherapy. Following this, he was offered subtotal oesophegectomy & feeding jejunostomy, which was uneventful. Post-operatively, his progress was slow due to delayed conduit emptying, 2ry to herniation of small bowel through hiatus. On exploratory laparotomy, he had his small bowel, reduced with approximation of the hiatal defect. A further gastroscopy was performed two days later, confirming integrity of the anastomosis, and healthy conduit. The patient started on jejunostomy feeding with plans to be discharged over coming 24 hrs. The following morning he developed severe abdominal pain, distension with evidence of septic shock. CT scan showed extensive diffuse small bowel dilatation, PI throughout whole small bowel and conduit. The patient was referred to our critical care unit for further support. Upon discussions between MDT and patient and family; we felt that emergency laparotomy was not appropriate due to his predictably very poor outcome. However we agreed to offer organ support with frequent reviews mainly to allow family to come to terms with this development but also; as success was reported with supportive care alone for PI, albeit, not in cases with severe symptoms, septic shock and extensive PVG. The patient stayed on critical care for 9 days, during which he received mechanical ventilation, significant vasopressor support, TPN & CRRT. He was given broad spectrum antibiotics, for gram negative bacteria and anaerobes. The patient went on to recover fully and was discharged home, where he is currently being followed up in surgical clinic. With this case presentation, we aim to highlight that maximising organ support, without emergent surgical intervention may be suitable therapeutic option, in carefully selected patients. We would like to propose this question: Is extensive PVG really a predictor for poor outcome?

Medical Handover on Intensive Care – A Regional Survey of the East Midlands

Joel Swindin¹ and Zoë Whitman²

¹Queen's Medical Centre, Nottingham University Hospitals NHS Trust, Nottingham, UK

²University Hospitals of Leicester NHS Trust, Leicester, UK

Abstract

The handover on intensive care is a critical focus point for transfer of information, decision making and forward planning. As the care of patients is transferred between teams, there is both opportunity for excellence in patient care and the risk of error, miscommunication and threat to patient safety. Handover is also an opportune time to check healthcare workers’ levels of fatigue and wellbeing. Despite interest in critical care nursing handover there has been little work studying medical handover on adult intensive care units.

Method: Utilising the MERCAT network we conducted a region-wide survey across the East Midlands; capturing data from 13 intensive care units at morning and evening handover over a 7-day period (182 handovers), with 22 data points per handover. As part of our survey we collected quantitative data regarding handover timings, location and format. We recorded the frequency and nature of interruptions, and qualitative data regarding perceptions of the handover.

Results: On average across the East Midlands, handover started 3.5 minutes late. This equates to 10.6 hours of delay per week across all sites. Of the 47 reasons given for delay, 27 (57%) were for staff being late to handover and 15 (32%) were for clinical urgency (usually the outgoing team being busy). Every unit took verbal handover, 3 units recorded use of electronic handover, most of the others had verbal handover with a typed handover sheet.

All units had consultant presence at every morning handover, 4 units had consultant presence at every morning and evening handover. One unit had no consultant presence at any evening handover. Eleven units had up to six handovers per week interrupted (often multiple occasions within handover) 20% of all handovers were interrupted at least once. At a time when medical staff fatigue was being highlighted nationally, 31/182 (17%) handovers included a verbal fatigue check.

The qualitative data collected provided opportunity for medical staff to grade the quality of handover with 92% of responses indicating they felt handover was safe and effective. Suggestions for improvement included punctuality and succinctness (whilst maintaining detail), minimising interruptions, effective allocation of patients, and additional or improved resources.

Conclusion: Handover is a complex process that enables patient care to be passed from one team to another. Utilising this survey there is much that can be learned and improved; particularly strategies for minimising interruptions, encouraging timeliness and structure whilst fostering a culture of wellbeing.

Whilst recognising that the format and needs of individual ICUs will vary, we are in the process of creating an Excellence in ICU Handover document with the aim of region-wide knowledge sharing, and improvement of handover to the benefit of both patients and staff.

Table 1.

Time taken to handover.

	Morning Handover	Evening Handover
Total time taken for handover (mins)
Mean	28.9	24.3
Median	29	25
Range	7–60	5–75
Time taken to handover per patient (mins)
Mean	2.7	2.3
Median	2.3	1.8
Range	0.4–9	0.6–8

The development and introduction of the Leeds Post-Extubation Dysphagia Screen

Claire Mills¹, Kerry Hunter¹, Lucia Hudson-Evans¹, Helen Timmins¹, Emilia Michou^2,3 and Mark Bellamy¹

¹Leeds Teaching Hospitals NHS Trust, Leeds, UK

²Centre for Gastrointestinal Sciences, The University of Manchester, Manchester, UK

³Speech and Language Therapy, Technological Educational Institute of Western Greece, Patras, Greece

Abstract

Approximately 56% of patients intubated >48 hours develop post-extubation dysphagia (PED).¹ Almost 50% of these patients aspirate, a high proportion silently², owing to the impact of intubation on cough reflex sensitivity.³ Detecting and preventing aspiration in this population is of paramount importance. Stage one of this quality improvement project involved development of a PED screen and a training programme for nursing staff. We aim to maximise detection of silent aspiration, enable the commencement of oral intake as soon as safely possible and ensure patients with rapidly resolving dysphagia are not kept nil-by-mouth unnecessarily.

An online survey targeting nursing staff and Adult Critical Care Practitioners (ACCPs) was conducted to ascertain current dysphagia screening practice. The Leeds Post-Extubation Dysphagia Screen (L-PEDS) was subsequently developed by the multi-disciplinary team, based on survey findings and recent research. L-PEDS includes a water swallow test, as continuously drinking a large volume has a high sensitivity for the detection of silent aspiration.⁴ It also incorporates both early, 1 hour post-extubation, and repeat screening prior to speech and language therapy assessment. A training programme was developed for nurses and ACCPs which includes: 2.5 hour training session (covering PED theory, instruction in L-PEDS use and training videos) and a minimum of one observed screen with a patient. Staff were required to complete a pre- and post-training quiz.

Twenty-four members of ACCP and nursing staff responded to an online survey. 83.3% of staff only try patients with oral intake for the first time post-extubation ≤1–2 times per month. 90% of staff reported waiting for ≥2 hours post-extubation before trying their patients with something to eat and drink. 54.2% of staff vary their approach with patients when screening for dysphagia. Staff suggested a standardised approach to screening with guidelines and training. Eighteen staff members underwent our trial training programme. The evaluation quiz showed staff confidence in giving patients oral intake post-extubation improved from 7.2 to 8.9 (on a scale of 0 to 10; 0 being very unconfident, and 10 being very confident). Staff self-reported knowledge of PED improved from 4.6 to 8.4 (on a scale of 0 to 10; 0 being very poor, and 10 being very good). Staff were also asked a range of specific questions related to screening patients for PED; the average score increased from 23.5% to 88.0% correct.

Our survey reveals that without a formal PED screen, there is variation in approach to dysphagia screening both within and between members of staff. Furthermore, infrequency of screening may lead to reduced confidence and a delay in patients being trialled with oral intake. Stage one of this quality improvement project demonstrates that L-PEDS is well received by staff and improves staff knowledge of PED and confidence in screening.

We acknowledge the Intensive Care Foundation Nurse and AHP Foundation Fellowship for funding this quality improvement project.

De- Bugging ITU; What to do with a positive blood culture

Rachel Saunders, Sam Clark, Ascanio Tridente and Yvonne Stubbington

Whiston Hospital, Liverpool, UK

Abstract

Positive blood cultures form a cornerstone in the diagnosis and treatment of sepsis. Falsely positive blood cultures due to contamination pose a significant clinical conundrum for the treating clinician.¹

We aimed to develop a tool to aid decision making in the process of distinguishing the postive from the false positive blood cultures. The aid also guides the clinician through difficult decisions, including at what point to remove central venous catheters, repeat cultures or treat the source of presumed infection.

In collaboration with the medical microbiology department we produced a Blood Culture Decision Algorithm, integrating knowledge of local microbial pathogens prevalence with evidence from current research, to help inform clinician decisions when faced with a positive blood culture. The alogrithm covers common ITU blood culture pathogens and suggests management plans depending on variables such as the site of culture sample, number of positive cultures, time to positivity, central and peripheral catheters in situ and the clinical picture.

A service evaluation of the frequency of arterial blood gas analysis and the incidence of anaemia and RBC transfusions in critically ill ITU patients

Aniket Paranjape¹ and David Hepburn²

¹Cardiff and Vale University Health Board, Cardiff, UK

²Royal Gwent Hospital, Aneurin Bevan Health Board, Newport, UK

Abstract

Introduction: Anaemia is a frequent complication in ITUs; partly speculated to be due to iatrogenic phlebotomy, especially high frequency arterial blood gas analysis (ABGA). Anaemia is corrected by red cell transfusions, but recent studies have shown a correlation between increasing RBC transfusions and higher patient mortality.

Objective: To explore the existence of a link between the frequency and volume of blood taken for ABGA and the incidence of anaemia and RBC transfusions in ITU patients.

Methods: This was a retrospective service evaluation based in Royal Gwent ITU involving 21 patients. Data was collected from 14 March —>4 April 2018. I measured the number of ABGs taken over total patient stay and calculated ABGA ‘discard’ and ‘testing’ volumes by observing ABGs being taken. The current system is an ‘open arterial line' system which requires blood to be drawn initially to clear the line (discard volume), followed by a second draw which is entered into the analyser (test volume). From this I extrapolated the average blood volume lost per draw and total daily blood loss per patient. I tracked RBC transfusion data and patient haemoglobin levels and assessed for a correlation. I designed and conducted a survey designed to provide insight into the reasons ICTU staff take an ABG.

Results:

Average volume per ABG (mL)	5.6 (4.8 discard volume, 0.76 test volume)
Average number of ABGs/day	5.8 (0.1–17)
Average blood loss/day (mL)	32.2 (0.5–94.6)
Blood loss over entire  ength of stay (mL)	236.1 (11.1–673.2)
Transfusion trigger Hb baseline (g/L)	70.6 ± 9.8

Significant correlation (p = 0.00023) between the ‘no. ABGs per day’ and the ‘% decrease in patient Hb levels’, suggesting a possible causation effect; the greater the number of ABGs/day a patient in ITU receives, the faster the rate of decay of their Hb levels (g/L).

Significant inverse correlation (p = 0.0000054) between duration of stay in the ITU and average patient Hb levels (g/L); the rate of haemoglobin loss = 4.64 g/L/day from day 0–12 in ITU; most patients become anaemic by day 2 of ITU admission despite normal admission Hb levels.

Survey showed that most staff were aware of the high blood volumes lost via ABGA, and interestingly predicted a total daily blood loss (36.6 mL) higher than the true value (32.2 mL). Answers showed the primary reasons for excess blood loss were the lack of a written ABGA guideline, and the use of an open arterial line system which required a high discard volume to clear the line.

Conclusion: The results agree with referenced studies and emphasise the omnipresence of ITU anaemia and its strong correlation, and possible causation, with phlebotomy testing. Staff should be aware of the dangers of excessive ABGA, and methods of reducing anaemia incidence include the implementation of strict written guidelines for ABGA and the introduction of a closed arterial line system, which would reduce blood loss by 86.3% by eliminating the need for a discard volume thus reducing the total ABG volume from 5.60 mL to only 0.76 mL.

High Flow Nasal Oxygen Therapy: Can we relieve some of the pressures on the critical care unit by offering th

Amanda Skingle, Janakan Anandarajah and David George

Wrexham Maelor Hospital, Wrexham, UK

Abstract

Bed occupancy in Welsh intensive care units is at an all time high, with many units reporting over 100% occupancy at times (1). Many hospitals have utilised the expertise of acute teams to manage the deteriorating patient on the ward and avoid admission to the high dependancy unit if possible to conserve bed numbers.

The benefit of high flow nasal oxygen is well documented in critical care, however it’s widespread use on acute medical and surgical wards isn’t as well established. To try and reduce the bed pressures facing critical care units, The Acute Intervention Team (AIT) at the Wrexham Maelor Hospital, a district general hospital in North Wales introduced the provision of high flow oxygen to ward patients.

Patients with increasing oxygen requirements on acute medical and surgical wards were identified to the AIT via escalation procedures already in place. Hypoxic patients, who could be managed on the ward by increasing oxygen delivery but didn’t require non invasive or invasive ventilation, were commenced on nasal high flow oxygen therapy by the AIT.

The AIT provided 24 hour support for the high flow therapy and undertook daily reviews of patients to ensure clinical improvement and escalate care if required. In order to evaluate the potential benefit of the provision of high flow nasal oxygen in the ward setting, data was collected during the daily review by the AIT and reviewed after 4 months of use.

Provisional results showed 18 patients were commenced on high flow nasal oxygen therapy, 14 of which were on acute medical wards. The age of patients receiving therapy ranged from 17–86, with the mean age of 66. The most common indication for commencing high flow therapy was pneumonia, representing 11 of the cases.

With regards to reducing bed pressures on the intensive care unit, 11 out of the 18 patients commenced on therapy were potential candidates for escalation. Out of these 11, only 2 were admitted to critical care, therefore avoiding 9 admissions if high flow were to be only offered in the critical care unit. Out of the 7 patients who weren’t not considered suitable for escalation of care, 3 died during their inpatient stay, showing a potential mortality benefit in this group. The overall mortality for all patients commenced on high flow oxygen therapy was 22%, representing 4 of the 18 patients.

Provisional data has shown an improvement in Pa02 in pre- and post- high flow oxygen therapy (p-value 0.041) and has already reduced pressures on the critical care unit by avoiding 9 admissions to the critical care unit, representing numerous bed days and financial savings.

Ongoing data is being collected on further patients commenced on high flow oxygen therapy in order to produce a local guideline for ongoing use on the ward and to further evaluate benefits of its use.

Serratus anterior compartment block: managing pain in traumatic rib fractures

Shahid Khan, Jonathan Leung and Ruth Tighe

EKHUFT, Ashford, UK

Abstract

Trauma is one of the leading presentations in the Emergency department.also one of the leading causes of mortality among ED patients.not just trauma the sequelae of the injuries and how they are subsequently managed has a significant impact on the mortality and morbidity of this group of patients. With an increase in the ageing population, prone to recurrent falls and trivial trauma, sustaining rib injuries, most of these patients end up in the icu with pneumonia; respiratory problems and persistent pain.

Managing pain in rib fractures is a prime concern and adequate analgesia And pain control has proven to reduce the morbidity in these patients with decreased number of ICU stay days and decreased complications.

the serratus anterior compartment forms a potential space similar to fascia iliaca, with the anatomy supporting pain control adequately via a local infiltration volume block. Under usg guidance it has a very good success rate and less prone to complications associated with nerve blocks and other forms of pain control.we performed a study on ten patients in our icu with good results, achieving better pain control, loss need for mechanical ventilation and decreased duration of icu stay.

Serratus anterior compartment block definitely holds great potential as a definite modality for pain control in traumatic rib fractures.

Predicting outcome in liver patients admitted to intensive care: dual-centre district general hospital external validation of the LiFE score

Steve Fry, Eleanor De Sausmarez, Paul Crowest, James Hayward and Luke Hodgson

Western Sussex Hospitals NHS FT, Worthing, UK

Abstract

Background: Acute hepatic dysfunction in the critically ill is associated with high mortality. Several prediction models have been developed to help risk stratify patients with liver disease.¹ This study externally validated the LiFE score², which incorporates lactate, bilirubin and INR, whilst also comparing performance with two other complex general ICU prediction models (ICNARC and APACHE II scores).

Methods: Dual-centre district general hospital UK retrospective study (2015-18). Inclusion criteria matched a recent UK-wide study¹ including at least one of: biopsy proven cirrhosis, imaging suggestive of cirrhosis, hepatic encephalopathy or portal hypertension. TRIPOD (transparent reporting of a multivariable prediction model for individual prognosis or diagnosis) recommendations for reporting was followed including discrimination (area under the receiver operating characteristic curves (AUROC)) and calibration (Hosmer-Lemeshow (H-L) test) performance of the models.

Results: Of a total of 4,690 admissions, 199 episodes (4%) met inclusion criteria (n = 169), with a mean age 57 (SD 13) and hospital length of stay of 13 (16) days. In-hospital mortality was 39% in this cohort compared to 18% of all admissions. In those patients with 1-year follow-up from time of inclusion (n = 153) survival was 35%.

Variceal bleeding (n = 55) was associated with a lower mortality (18% vs 47%, P < 0.001, OR 0.3 (95% CI 0.1–0.5)). Patients (n = 38) requiring renal replacement therapy had significantly higher in-patient mortality (82% vs 29%, P < 0.001, OR 11.1 (4.6–26.9)). Of those with 1-year follow-up only 3.2% (n = 1) were still alive. Requirement for ventilation was associated with increased in-hospital mortality (47% vs 32%, P = 0.03, OR 1.9 (1.1–3.4)), but not with a significant decrease in 1-year survival (29% vs 39%, P = 0.24, OR 0.7 (0.3–1.3)).

Mean scores were: LiFE 6.6 (5.2), ICNARC 20.3 (10.6) and APACHE II 16.0 (9.6). For in-patient mortality AUROCs were: LiFE 0.694 (95% CI 0.618–0.769), ICNARC 0.802 (0.741–0.862) and APACHE II 0.729 (0.648–0.810). The H-L test P-value for calibration was LiFE P = 0.764, ICNARC P = 0.377 and APACHE II P < 0.001. AUROCs for 1-year survival were: LiFE 0.706 (0.620–0.792), ICNARC 0.773 (0.694–0.852) and APACHE II 0.695 (0.609–0.780).

Conclusions: This first validation of the LiFE score in a DGH setting suggests this parsimonious and easy to calculate model may have utility in prediction of short-term and 1-year mortality in this cohort. A limitation of the presented study is the relatively small sample size.

Table 1.

comparison between survivors and non-survivors.

	Survivors (n = 122)	Non-survivors (n = 77)	P-value, OR (95% CI)
Age	57.0 (SD 13.7)	57.7 (12.4)	0.714
LiFE score	5.3 (4.8)	8.7 (5.1)	<0.001
ICNARC score	16.1 (8.3)	26.9 (10.6)	<0.001
APACHE II score	13.5 (7.5)	20.1 (11.0)	<0.001
Lactate mmol/L	3.2 (3.0)	6.2 (5.7)	<0.001
Bilirubin μmol/l	68 (77)	105 (108)	0.006
INR	1.7 (0.8)	2.0 (0.9)	0.010
Renal replacement therapy	6%	40%	<0.001, 11 (4.6–26.9)
Ventilation	39%	55%	0.029, 1.9 (1.1–3.4)
Surgery	16%	7%	0.073, 0.4 (0.1–1.1)
Variceal bleed	37%	13%	<0.001, 0.3 (0.1–0.5)
Hepatic encephalopathy	16%	40%	<0.001, 3.7 (1.9–7.1)

Setting treatment limits in ITU patients

Henry Douglas Robb¹, Alona Courtney², Maria Schofield Legorburo² and Monica Popescu²

¹Imperial College Healthcare NHS Trust, London, UK

²West Middlesex University Hospital NHS Trust, London, UK

Abstract

Introduction: It is a given that every patient admitted to ITU is in critical condition. According to INARC Case Mix report, the mean age on admission to ITU is 61 years old, with 58% of patients admitted with non-surgical issues, including Sepsis and AKI [1]. About 20% of all patients admitted to ITU do not survive to discharge following acute admission to hospital [1]. Of those that are discharged, another one third die within 3 years [2]. NICE, GMC, the Royal College of Physicians and the Royal College of Anaesthetists advocate the importance of setting appropriate treatment limits for all acutely ill patients [3], especially if the risk of cardiac or respiratory arrest is foreseeable.

Methods: Data was collected prospectively between October and November 2016. All patients admitted to the ITU department of West Middlesex University Hospital, London, during this period were included. The Royal College of Anaesthetists standard 10.4 from “Raising the Standard: a compendium of audit recipes” (2012) was used as a gold standard [3]. Data collection included patient’s age, sex, co-morbidities and documentation of the treatment limits by the ITU team. Following initial audit, two new proformas were introduced and the data was re-audited in May 2018.

Results: Forty patients were included in this audit. The mean age on admission was 62, of whom 58% were male. 75% of patients had significant co-morbidity, for example COPD, congestive cardiac failure and previous MI. Only 15% (n = 6) of patients had a treatment limit set either on admission or prior to discharge. Of those discharged from ITU (n = 37), only 8% had a treatment limit documented in their discharge summary. At least 15% of patients discharged without a treatment limit were re-referred back to ITU by the ward teams and their admission was declined. Following introduction of the proforma, 95% of patients (n = 19) had treatment limits set on admission and 50% on discharge.

Discussion: Despite clear standards from multiple regulatory bodies, the compliance with the guidelines in our unit was poor. Although our unit was small and the lack of documented treatment limit may have had little impact on the patient care. When it came to transfer of care back to the wards, lack of documented treatment limit may have resulted in confusion and inappropriate patient care by the ward teams, as most patients had ‘Full Escalation’ documented on their discharge summary. Initially implementation of a simple tick-box ‘Escalation Plan’ form was unsuccessful; however, it led to increased awareness of this issue amongst the management team. This instigated subsequent introduction of a formal Treatment Escalation Proforma, which resulted in a significant improvement in the documentation of treatment limits.

Prognostication of Hypoxic Brain Injury Patients in the Intensive Care Unit

Pia Zaldua¹ and David Hepburn²

¹Cardiff University, Cardiff, UK

²Royal Gwent Hospital, Aneurin Bevan Health Board, Newport, UK

Abstract

Objective: Due to the complicated and severe nature of devastating and hypoxic brain injury, it is difficult to determine the prognosis of these patients. This audit will assess the prognostication modalities used in brain injury patients both due to anoxia following cardiac or respiratory arrest and due to suspected devastating brain injury from an intracranial event retrospectively in critical care patients in the Royal Gwent using the current recommended guidelines.

Method: Using the ICNARC database, 29 patients were identified to be diagnosed with significant brain injury. The study aimed to identify the investigations performed and the length of time between admission and diagnosis of suspected brain injury. The project involved analysing patients’ notes, clinical letters and radiological reports to determine the investigations done and the length of time between admission and diagnosis of suspected brain injury. Other factors such as neurology or neurosurgical referral, organ donation, outcomes and delays in modalities were also considered.

Results: Severe brain injury was diagnosed with an average of 5.91 days (141.91 hours), meeting the standards for the management of devastating brain injury and hypoxic brain injury after cardiac arrest. All of the patients received CT scans, 66% of the scans were found to have no abnormalities. Two CT scans and 40% of MRIs were reviewed by neuroradiologists and found to have early signs of brain injury. Cases such as this would be avoided if there was a dedicated neuroradiologist in the critical care unit. Only a small number of MRIs were performed compared to CT scans and this could be due to the fact that patients require sedation, ventilation and transfer. However, MRI is very useful at portraying the grey and white matter differentiation compared to CT scans. Only 7 EEGs were performed, 5 of the EEGs determined the diagnosis. The critical care unit does not have good availability for EEGs, hence the reason why they may be infrequent and formal EEG reports can take days to arrive.

Conclusions: In conclusion, the guidelines for both types of brain injury were largely met. Most diagnoses happened after the 72 hour period, as suggested by the recommendations. Few patients had access to multimodal prognostication and the critical care access to other forms of investigations should be improved. Accurate prognostication can help determine the outcome of patients with both types of brain injury and availability of other diagnostic modalities would assist in early and definite diagnosis, allowing families to cope and understand the reality of their relatives’ poor prognosis.

Reducing morbidity through surgical morbidity and mortality discussions

Calum Worsley, Stephen Webb and Sarah Powell

Royal Papworth Hospital, Papworth Everard, UK

Abstract

Morbidity after major cardiac surgery is associated with increased duration of ITU stay, impaired return to function, and is a major cause of perioperative mortality.¹ In recent years there has been a national focus from the society of cardiothoracic surgeons on perioperative mobidity, as well as locally at Royal Papworth Hospital through a Sign up to Safety initiative.

The focus of the project at large was the monthly surgical morbidity and mortality (M&M) meetings. Initially the baseline rate of strokes was determined, through retrospective analysis of the past year’s CT head reports and was calculated at 1.8%, within expectations nationally and internationally.^2,3 Further data about patients’ longer term inpatient progress was then obtained through liaison with the speech and language therapy and physiotherapy teams.

The aim of the project was to facilitate improvement work, so the data presented was regularly varied in form and content according to what the morbidity and mortality meeting attendees thought would be useful.

General data was regularly presented at every meeting subsequently, then over time, as the meetings got accustomed to the presentations, more specific data was introduced. Individual consultants began to bring discussion about patients of theirs who had suffered perioperative strokes to the meetings. Subsequently, such discussion became standard, with updates on those patients’ progress from physiotherapy and speech and language therapy teams as well as from the medical and surgical perspective.

Once the stroke discussions were well-established, data about other perioperative morbidity markers were presented. Senior registrars pointed out that they were frequently the performing surgeon in operations that were being rolled up into the consultant’s figures; at their request their figures were separated out and presented alongside the consultants’.

Potential solutions dreamed up from behind a spreadsheet could not hope to be the most effective – problems and potential solutions are often most apparent to those working most closely with them. We didn’t come at the project with solutions in mind, nor did we seek to generate them ourselves, but the discussion that became commonplace in the meetings began to generate ideas of its own accord. A registrar found a checklist for reducing post-operative bleeding and adapted it for local use. Discussion of patients requiring renal replacement therapy led to a group of consultants pushing for a more standardised approach to use of pulsatile vs non-pulsatile bypass. A business case was written to trial epiaortic ultrasound to reduce watershed strokes caused by inadvertant cannulation of aortic plaques.

Data collection requires time and effort, so much so that frequently the fact of having collected and processed that data is an end in and of itself. This project established continuity, requiring the same outcomes to be measured on a monthly basis and presented in consistently familiar formats. The regular turnover of trainees can have a stifling effect on long term projects, but this continuity facilitates the creation of a culture of improvement as well as allowing others to build on this work.

The application of Machine Learning techniques to patient outreach data to predict admission to the Intensive Care Unit

Thomas Rodwell

Cardiff University, Cardiff, UK

Abstract

I report on the application of Machine Learning and ‘Big Data’ techniques to the patient outreach data collected at University Hospital Wales, Cardiff to predict admission to the hospital’s Intensive Care Unit (“ICU”). Specifically, I describe the application of logistic regression, ensemble classifiers and neural networks to analyse a large data set of anonymised patient outreach data, including physiological data and written observations from hospital staff.

I find that it is possible to predict with a high degree of accuracy (>80%) those patients that will be admitted to the ICU. I also report on my initial findings to make accurate predictions on when particular patients will be admitted to the ICU, as well as predictions in respect of those patients' Length of Stay on admission, with the aim of better forecasting admissions over a given (short-term) time period. Such forecasts could have a wide range of applications. In particular, I am interested in assisting hospital staff make better scheduling decisions. This could allow scheduled nursing levels to more accurately reflect the predicted utilisation of the ICU and, therefore, reduce the number of cancelled elective surgeries.

Procalcitonin in the ICU: Early increases in PCT are associated with increased hospital mortality

Paul Hayden, Rahul Sarkar and Karen Williams

Medway NHS Foundation Trust, Gillingham, UK

Abstract

Introduction: Procalcitonin (PCT) is a widely investigated validated biomarker that is a valuable aid to guide decision making with regard to the optimal management of patients admitted with sepsis in critical care units. Specifically, the assay provides objective data to guide the duration of antibiotic regimes (1), but may also be useful as a prognostic tool. The MOSES trial, a recent US multicentre study by Shuetz and colleagues demonstrated an increase in mortality in those patients who initially presented with a low PCT value that subsequently rose. We have used PCT as a biomarker on our ICU for the past ten years and sought to identify whether this finding would be similar in our cohort.

Methods: We conducted a retrospective analysis of laboratory data for 617 patients admitted to the intensive care unit between 2008 and 2018 using admission and outcome data from the Wardwatcher™ system with a clinical diagnosis of sepsis and an initial PCT value of at least 0.25 ng/ml and at least one subsequent PCT measurement. A logistic regression was performed on the crude data to adjust for severity of illness (ICNARC score) using the STATA® statistical software package. We considered the sample in terms of centiles of percentage change in PCT value from the initial PCT measurement (PCT1) to the next PCT measurement 48 hours later (PCT3).

Results: Our data demonstrates a trend towards increased mortality for any rise in PCT value from PCT1 to PCT3 but this is statistically significant if the increase in PCT is 50% or greater (odds ratio 1.55, CI 1.06–2.26; p = 0.02). Conversely, an early drop in PCT value of at least 60% is associated with improved outcome (odds ratio 0.52, CI 0.29–0.92; p = 0.03).

PCT gradient	n	Hospital mortality (%)	Odds ratio	P value	CI
>10% rise	222	105 (47%)	1.45	0/04	1.02–2.06
>20% rise	208	98 (47%)	1.37	0.08	0.96–1.97
>30% rise	188	90 (48%)	1.41	0.07	0.97–2.03
>40% rise	177	84 (47%)	1.39	0.08	0.96–2.02
>50% rise	168	83 (49%)	1.55	0.02	1.06–2.26
>60% rise	156	78 (50%)	1.59	0.02	1.08–2.35
>70% rise	146	73 (50%)	1.60	0.02	1.08–2.38
>80% rise	142	70 (49%)	1.56	0.03	1.04–2.32
>90% rise	134	68 (51%)	1.75	0.007	1.16–2.62
>100% rise	126	65 (52%)	1.84	0.004	1.21–2.80
>10% drop	296	103 (35%)	0.77	0.15	0.55–1.1
>20% drop	252	54 (33%)	0.71	0.06	0.5–1.01
>30% drop	206	67 (33%)	0.72	0.08	0.5–1.05
>40% drop	161	57 (35%)	0.86	0.45	0.58–1.27
>50% drop	120	37 (31%)	0.67	0.08	0.43–1.05
>60% drop	75	20 (27%)	0.52	0.03	0.29–0.92
>70% drop	45	12 (27%)	0.51	0.07	0.25–1.05
>80% drop	20	4 (20%)	0.32	0.058	0.09–1.04
>90% drop	7	2 (29%)	0.45	0.38	0.08–2.66

Conclusion: Early changes in PCT value predict hospital mortality in patients admitted to the ICU with sepsis. These data could be incorporated into future risk scores to improve the accuracy of outcome prediction and are clinically useful to identify high risk patients.

Peri-operative support for an adult with severe ischaemic intestinal malrotation

Ben Griffiths

Jersey General Hospital, St. Helier, Jersey

Abstract

Introduction: Intestinal malrotation is a congenital condition resulting from arrested embryonic rotation of the gut in utero. Most cases present in early childhood but malrotation can remain asymptomatic throughout life. Malrotation with mid gut volvulus is a life threatening emergency that requires prompt resuscitation, emergency laparotomy and peri-operative critical care involvement.

Case History: We describe a 60 year old patient of excellent general health who presented following 4 years of intermittent constipation, 6 months of weight loss and 12 hours of abdominal pain, vomiting and distension. She became hypotensive in the Emergency Department and required pre-operative resuscitation with crystalloids and phenylephrine en route to theatre.

Exploratory laparotomy showed a malrotated small bowel with profound ischaemia. The rotation was reduced and the patient was transferred to Intensive Care with wound packs to allow time for reperfusion. A repeat laparotomy was performed at 6 hours, where small sections of clearly ischaemic bowel were excised. At 18 hours post reduction the proximal colon, entire ileum and most of the jejunum were excised. A right hemicolectomy and proximal end jejunostomy was performed with only 30 cm of jejunum remaining.

Critical Care: The patient suffered a profound metabolic acidosis and remained intubated and ventilated for three days following surgical closure. After a 48 hour period of post extubation delirium she made a full physiological recovery.

Clean and clear central vascular access was afforded critical importance due to her anticipated reliance on total parenteral nutrition. A PICC line was sited and the patient was restricted to 1,500 ml oral fluid intake with a low residue diet. Loperamide was started to reduce transit time. The patient was transferred to an intestinal failure unit for continuing care and is expected to have good prospects for rehabilitation.

Discussion: Although rare, small bowel malrotation should be considered as a cause of severe acute abdomen in adults with a history of intermittent constipation and weight loss.

Mid gut volvulus is a reversible critical ischaemia, the old adage prevails: time is bowel.

Intestinal failure often necessitates the use of long-term or lifelong total parenteral nutrition. In this context, avoiding phlebitis becomes imperative.

Early phase multidisciplinary management of intestinal failure can improve patient outcomes and should be instituted whilst awaiting transfer to an intestinal failure unit.

Association between hospital volume and mortality in status epilepticus: a national cohort study

Robert Goulden, Tony Whitehouse, Zahid Khan, Tom Hayton, Nick Murphy, Catherine Snelson, Julian Bion and Tonny Veenith

Queen Elizabeth Hospital, University Hospital Birmingham NHS Trust, Birmingham, UK

Abstract

Objective: In various medical and surgical conditions, research has found that centres with higher patient volumes have better outcomes. This relationship has not previously been explored for status epilepticus (SE). This study sought to examine whether centres that see higher volumes of patients with SE have lower in-hospital mortality than low volume centres.

Design

Cohort study, using 2010 to 2015 data from the nationwide Case Mix Programme (CMP) database of the UK’s Intensive Care National Audit and Research Centre (ICNARC).

Setting

>90% of intensive care units in England, Wales, and Northern Ireland.

Patients

20,922 adult critical care admissions with a primary or secondary diagnosis of status epilepticus or prolonged seizure.

Interventions

Annual hospital SE admission volume.

Measurements and main results: We used multiple logistic regression to evaluate the association between hospital annual SE admission volume and in-hospital mortality. Hospital volume was modelled as a non-linear variable using restricted cubic splines, and generalised estimating equations with robust standard errors were used to account for clustering by institution. There were 2,462 in-hospital deaths (11.8%). There was no significant association between treatment volume and in-hospital mortality for SE (p = 0.54). This conclusion was unchanged across a number of subgroup and sensitivity analyses, although we lacked data on seizure duration and medication use. Secondary analyses suggest that many high-risk patients were already transferred from low to high volume centres.

Conclusions and relevance: We find no evidence that higher volume centres are associated with lower mortality in SE overall. It is likely that national guidelines and local pathways in the UK allow efficient patient transfer from smaller centres like district general hospitals to provide satisfactory patient care in SE. Future research using more granular data should explore this association for the sub-group of patients with refractory and super-refractory SE.

Airway Management in the Intensive Care Unit (ICU) Quality Improvement Project (QIP)

Yun Mei Lau^1,2, Lynne Barrass² and Julia Hadley²

¹The Royal Free NHS Foundation Trust, London, UK

²Royal London Hospital, Barts Health NHS Trust, London, UK

Abstract

Introduction: Airway management in the ICU is a challenging process as it presents issues unique to the setting. ICU patients have pre-existing deranged physiological reserve and acute haemodynamic instability making airway assessments difficult to perform to the usual standards. Despite these circumstances, prompt and successful airway management is required to halt a patient's rapid deterioration.

The 4th National Audit Project (NAP4) identified that 25% of major airway events occur in the ICU/ED setting with the rate of ICU airway-related complications being 50 times higher than the operating theatre. NAP4 stressed that avoidable airway deaths occurred.

Objectives: We designed an Airway Management QIP looking into three main stems: 1) Documentation of intubation, 2) Difficult airway identification and rescue plan and 3) Identification of tracheostomy patients and rescue plan.

Method: Prospective review of airway documentation compliance in our Adult Critical Care Unit, a busy 44-bed unit in a major trauma and tertiary referral centre. Daily Observation Charts (DOC), Daily Review sheets (DR) and Admission Proformas (AP) were reviewed for patients who have/have had an airway intervention for a period of three weeks in October 2016.

Results: 56 cases including 7 tracheostomy patients were examined. 31 were in Day 1–5 of admission, 10 in Day 6–10, 7 in Day 11–15, 5 in Day 16–20, 1 in Day 21–25, 1 in Day 26–30 and 1 in more than 30 days.

Table: Percentage of airway documentation compliance

n = 56	Daily Observation Chart (DOS)	Daily Review (DR)	Admission Proforma
Grade of intubation	39.3%	10.7%	33.9%
Size of ETT/Trache	42.9%	N/A	16.1%
cm at lips	30.4%	N/A	N/A
Cuff Pressure	37.5%	N/A	N/A

100% of patients with an identified difficult airway had no airway information recorded on the DOC and DR. 50% of those patients had their grade of intubation documented on the AP. Size of ETT was not documented. None of these patients had a rescue plan in place.

In tracheostomy patients, 57.1% had no grade of intubation, size of tracheostomy or cuff pressure documented on the DOC. Size of tracheostomy was the most commonly completed value followed by grade of intubation. None of the patients’ initial grade of intubation was noted in the DR sheets.

Importantly, 37.5% of all cases examined did not have their Grade of intubation documented in any of those three main ACCU documents.

There was no demonstrable relationship between the day of admission and documentation compliance.

Conclusion: Airway documentation plays an essential role in airway management. It is currently a hit-and-miss phenomenon amongst doctors and nurses. This may be explained by ICM becoming a stand-alone specialty with trainees and nurses coming from different acute medical backgrounds with varying airway experience.

With these results, we have introduced an Airway Documentation sticker which includes the above airway information, identification of a Difficult Airway and the subsequent individualised Rescue Airway Plan. We aim to assess awareness and subsequent behavioural change with further data collection and user questionnaire.

Could a checklist allow any doctor or nurse to safely extubate?

Fei Lin and Rony Berrebi

Northwick Park Hospital, London, UK

Abstract

Background: Prior to extubation, a patient’s ability to maintain sufficient oxygenation, ventilation, airway protection and airway patency should be evaluated. This usually involves assessing a patient’s general condition and respiratory status. There are frequently applied standards, for example a patient should be able to maintain oxygen saturations of >90% on an inspired oxygen concentration (FiO2) of <40%. However, there are no universally accepted threshold values or published guidance on extubation criteria. Often the final decision to proceed with or postpone extubation is made on an individual case basis and can vary from practitioner to practitioner.

Pre-extubation assessment aims to prevent major complications, such as extubation failure and the subsequent need for re-intubation. Re-intubation is associated with higher rates of hospital mortality, hospital acquired pneumonia, a longer stay in the intensive care unit (ICU) and prolonged hospitalisation.

Methods: In a large, acute general hospital an evidence-based extubation checklist was created for use in the ICU. The checklist assessed 19 criteria: alertness, pain control, respiratory pattern, tidal volume, cough strength, airway secretions, FiO2, arterial partial pressure of oxygen, ventilator positive end-expiratory pressure or pressure support, base excess, temperature, signs of sepsis, blood pressure, inotropic requirement, haemoglobin level, nutritional status and time of last feed.

We conducted a three month retrospective study comparing the following patient outcomes after extubation with or without using the checklist:

• Rate of extubation failure, defined as re-intubation within 48 hours of extubation

• Length of ICU stay

• Length of hospital stay

• Mechanical ventilation duration

Results: We have collected data from 24 patients thus far (the number is expected to exceed 100 by the time of presentation). 13 were assessed with the extubation checklist and 11 without. No patients failed extubation in either cohort. Medical patients in the checklist cohort had shorter ICU and hospital stays (6.75 vs. 9.67 and 12.13 vs. 18.33 days, respectively), as did emergency intubation patients (7.83 vs. 8.11 and 17.58 vs. 18.00 days, respectively). Duration of mechanical ventilation was shorter in the checklist cohort (2.92 vs. 3.27 days). This was most marked for medical patients (2.75 vs. 4.67 days) and emergency intubation patients (3.08 vs. 4.00 days).

Discussion: Our preliminary data suggest that using an extubation checklist provides non-inferior rates of extubation failure compared to traditional approaches. It also demonstrates the potential benefits of using a checklist in standardising and supporting the safe extubation of ICU patients. It was associated with a shorter mechanical ventilation duration, ICU stay and hospital stay for both medical and emergency intubation patients.

The use of a non-randomised, small sample size and a single centre are limitations of this study. A larger, randomised trial is required before we can change practice. Additionally, the parameters assessed in the extubation checklist may benefit from further refinement.

A validated extubation checklist could allow safe junior- or nurse-led extubation, which could reduce extubation delays, costs and complications associated with prolonged mechanical ventilation.

Failure to correct Procalcitonin in ICU survivors is associated with increased hospital mortality

Paul Hayden¹, Rahul Sarkar¹ and Karen Williams²

¹Critical Care Department, Medway Maritime Hospital, Gillingham, Kent, UK

²Medway NHS Foundation Trust, Gillingham, UK

Abstract

Introduction: Procalcitonin is a valuable biomarker in the intensive care unit to optimise the use of antibiotic therapy in the management of sepsis (1). There are an increasing amount of data to demonstrate a potential role for procalcitonin as part of a predictive strategy to identify patients at increased risk of poor outcome, thereby enabling clinicians to focus attention on this high risk cohort. A recent US multicentre trial of PCT in patients admitted to the ED and ICU demonstrated an increased mortality in those patients who were unable to correct an elevated PCT value by at least 80% (2). This work has been demonstrated elsewhere but to our knowledge has not been identified in a UK cohort and we were unable to find published literature evaluating the effect of a failure to reduce PCT value in patients following ICU discharge. We sought to identify whether a failure to correct PCT value in patients surviving ICU admission is associated with an increase in hospital mortality.

Methods: We conducted a retrospective analysis of 444 patients admitted to our ICU with a clinical diagnosis of sepsis from 2008–2018 with a PCT value >0.25 ng/ml who survived to ICU discharge. We performed logistic regression to account for variation in severity of illness (ICNARC score) at admission to ICU using the STATA statistical analysis package. We then evaluated hospital survival following ICU discharge based on the percentage change from maximal PCT value (PCTmax) to the final PCT value (PCTfin) for patients prior to their discharge from the ICU.

Results: There were 444 patients who survived ICU admission (crude ICU mortality 28.1%). Failure to correct PCT from maximal levels on the ICU was associated with higher hospital mortality. The optimal odds ratio for hospital survival was 0.33 (CI 0.18–0.58; P < 0.0001) for patients whose PCT values dropped by at least 70% prior to ICU discharge compared with those who were unable to correct PCT by this amount.

PCT drop	n	deceased (%)	Odds Ratio	CI	p value	AUROC
>10%	345	51 (15)	0.46	0.25–0.84	0.011	0.62
>20%	330	49 (15)	0.51	0.29–0.91	0.023	0.62
>30%	314	46 (15)	0.53	0.31–0.92	0.025	0.61
>40%	295	43 (15)	0.54	0.32–0.93	0.026	0.61
>50%	272	36 (13)	0.46	0.27–0.77	0.003	0.63
>60%	228	26 (11)	0.38	0.22–0.65	<0.0001	0.65
>70%	198	20 (10)	0.33	0.18–0.58	<0.0001	0.67
>80%	155	17 (11)	0.42	0.23–0.77	0.005	0.64
>90%	94	11 (12)	0.52	0.26–1.05	0.07	0.61

Conclusion: Our data suggest that patients who do not reduce their PCT value prior to ICU discharge at at elevated risk of hospital mortality. This is potentially useful in the UK healthcare system where critical care bed availability is always at a premium and there is considerable pressure to step patients down to ward level care. The inability to correct PCT value could be used as a factor in decision-making regarding who to keep longer on the critical care ward.

Outcomes following acute poor grade aneurysmal subarachnoid bleed – Is early definitive treatment better than delayed management?

Adam Gittins¹, Ahmed Gilani¹, Edward White¹, Hang Seng Chew², Alan Thomas², Fang Gao-smith¹, Catherine Snelson¹, Nick Murphy¹, Tom Gallacher³, Tony Whitehouse¹, Murali Shyamsundar⁴, Antonio Belli⁵, Tonny Veenith¹

¹Birmingham Acute Care Research, University of Birmingham, Birmingham, UK

²Department of Neuroradiology, University of Birmingham hospitals NHS trust, Birmingham, UK

³Department of Critical Care medicine, University Hospital of Birmingham NHS trust, Birmingham, UK

⁴Consultant in Intensive Care Medicine Queen's University, Belfast, Belfast, UK

⁵Department of Neurosurgery, Birmingham Acute Care Research, University of Birmingham, Birmingham, UK

Abstract

We present the largest cohort of patients with poor grade subarachnoid bleed with long term outcome data comparing immediate management within 24 hours (IM) to delayed management after 24 hours (DM) in patients with poor-grade (WFNS grades IV-V) aneurysmal subarachnoid haemorrhage (pSAH). 254 patients with pSAH admitted in our tertiary referral centre with pSAH between 2011 and 2016 were included in our analysis. The primary outcome was disability following discharge up to 12 months, as measured by the Glasgow Outcome Scale (GOS). The secondary outcome was mortality. 111 patients received aneurysm treatment (AT) and 143 patients received conservative management (CM). Of the 111 AT patients, 53 patients received IM and 58 received DM. At discharge, GOS scores were significantly statistically marginally higher in IM patients than DM patients (3.80 +/− 0.833 in IM patients vs 3.44 +/− 0.741 in DM patients, p < 0.05).

However, GOS scores at 3 months, 6 months and 1 year did not differ significantly between IM patients and DM patients (4.29 +/− 0.806 in IM patients vs 4.48 +/− 0.769 in DM patients at 1 year, p = 0.3722). Survival rates at 28 days, 3 months, 6 months and 1 year did not differ significantly between IM and DM patients (71.7% in IM patients vs 82.8% in DM patients at 1 year, p = 0.1629). pSAH patients demonstrate a possible survival benefit from early definitive treatment. IM and DM after pSAH are associated with similar morbidity and mortality. Delaying intervention on poor-grade patients may be prudent to allow for planning of the procedure and stabilisation of the patient, ensuring optimal patient care

Revising the RCA; Streamlining Root-Cause Analyses for bacteraemias on ITU

Rachel Saunders, Sam Clark, Yvonne Stubbington and Ascanio Tridente

Whiston Hospital, Liverpool, UK

Abstract

Sepsis is a leading cause of mortality within the Intensive care environment.¹ Over the last 3 decades, the recognition of sepsis has greatly improved and blood cultures form an integral part of both diagnosis and management. At the authors’ institution, any detected bacteraemia is examined through a process of Root Cause Analysis (RCA) to ensure that, where necessary, lessons are learnt and any improvement to patient care is implemented.

The current RCA process is extremely cumbersome, with potential for streamlining for specific pathogens, particularly Gram negatives. A large proportion of bacteraemias within our institution are due to Gram negative organisms and often do not represent Healthcare Associated Infections. We aimed to create a separate process tailored to succinctly evaluate the cause of Gram negative bacteraemias, without compromising on the quality of the process.

In collaboration with the infection control and medical microbiology teams, while satisfying the mandatory requirements of the Department of Health, we redesigned the RCA process for ITU bacteraemias. We identified two different pathways; one dedicated to those bacteraemias requiring full investigation (usually MRSA, MSSA, CDI and VRE), and a second, more succinct, process for the initial screening of the Gram negative and other low-risk bacteraemias.

Out-of-hours discharge from intensive care is associated with increased mortality. What next?

Sarah Vollam and Peter Watkinson

Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, UK

Abstract

Discharge from an intensive care unit (ICU) out-of-hours has been explored in many ICU database studies^1–3. As background work to the REFLECT study (ISRCTN14658054) we undertook a systematic review and meta-analysis to explore the association between time of discharge and both mortality and ICU readmission.⁴

We searched Medline, Embase, Web of Knowledge, CINAHL, the Cochrane library and OpenGrey to June 2017. Inclusion criteria were: studies reporting in-hospital mortality and/or ICU readmission rates by “out-of-hours” and “in-hours” ICU discharge time; patients aged ≥16 years discharged alive from a non-specialist ICU to a lower level of hospital care. We excluded studies restricted to specific diseases. We extracted published data, summarising using a random-effects meta-analysis. From 1,961 studies identified in searches we used our eligibility criteria to exclude studies not relevant. 18 cohort studies were included, with data from 1,191,178 patients from 1994–2014. There was variability in the definition of “Out-of-hours”, commencing between 16:00 and 22:00 and finishing between 05:59 and 09:00. Patients discharged out-of-hours had higher in-hospital mortality, relative risk (95% CI) 1·39 (1·24, 1·57) p < 0.0001 and readmission rates, relative risk (95% CI) 1·30 (1·19, 1·42), p < 0.001 than patients discharged in-hours. Heterogeneity was high (I² 90.1% for mortality and 90·2% for readmission). Our meta-analysis shows out-of-hours discharge from an ICU is associated with substantial increases in subsequent hospital mortality and ICU readmission. The effect remained when studies were stratified by geographical location and by definition of out-of-hours. As all studies included were retrospective it is impossible to attribute causation. It is also unclear why patients are discharged from ICU at night. This may occur under pressure to make capacity for an emergency admission, or may be due to delays in identifying a suitable ward bed.

As part of the REFLECT study, we reviewed the medical notes of 300 patients who died following discharge from ICU, across three hospital trusts of varying sizes in the UK. Patients in this cohort were often discharged in the evening or overnight. Only 32.3% of non-palliative discharges occurred before 16:00. We also interviewed 30 staff and 20 patients about their experiences of transfer from ICU to the ward. Out-of-hours discharge was a common theme discussed by both patients and staff. Receiving patients in the evening, at a time of reduced staffing and high workload, was a concern frequently raised by ward nurses and doctors. We are currently developing this work to investigate the discharge processes and reasons for out-of-hours discharge from ICU.

This abstract presents independent research funded by the NIHR under its Research for Patient Benefit (RfPB) Programme (Grant Reference Number PB-PG-0215-36149). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health & Social Care.

Haemophagocytic Lymphohistiocytosis-like syndrome reduces survival in patients with ARDS

Patrick McAleavey¹, Andrew Boyle^1,2, John Conlon¹, Cecilia O'Kane¹ and Danny McAuley^1,2

¹Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

²Regional Intensive Care Unit, Royal Victoria Hospital, Belfast, UK

Abstract

Background: Haemophagocytic Lymphohistiocytosis (HLH) is a rare hyperinflammatory condition. Ferritin and IL-18 are elevated in cases of HLH, however their role in the pathogenesis of HLH is not understood. Plasma ferritin and IL-18 are recognised biomarkers for HLH. In patients with sepsis, an HLH-like syndrome (HLH-LS) has been reported [1]. However, it is unknown if a similar HLH-like syndrome co-exists in patients with ARDS, and if it modifies outcomes in patients with ARDS.

Hypothesis: We hypothesised that an HLH-LS occurs in patients with ARDS, and that its presence is associated with increased mortality.

Methods: A post-hoc analysis of the HARP-2 clinical trial was undertaken. HARP-2 was a randomised, controlled clinical trial evaluating simvastatin in 540 patients with ARDS. Baseline plasma samples obtained were analysed for ferritin and total IL-18 using an enzyme-linked immunosorbent assay. Ferritin >4000 ng/mL or a total IL-18 >2500 pg/mL have been reported to have a high sensitivity and specificity for the diagnosis of HLH-LS [1,2]. Patients with high ferritin or high IL-18 values were compared to a control group with normal ferritin and IL-18 values.

Results: 511 patients where baseline samples were available were included. High baseline ferritin was present in 58 (11.4%) patients. These patients had a higher baseline SOFA score, fewer ventilator free days (VFDs) and higher 28-day and 12-month mortality. High baseline IL-18 was identified in 54 (10.6%) patients. These patients had a higher baseline SOFA score and higher 28-day mortality. Although these patients also had fewer VFDs and higher 12-month mortality, these findings were not statistically significant (Table 1).

Conclusion: This post-hoc analysis demonstrates that an HLH-LS, identified by high ferritin or high IL-18, occurs in patients with ARDS. Furthermore, the presence of an HLH-LS with ARDS is associated with worse clinical outcomes. Identification of this co-existing HLH-LS might offer an opportunity to improve outcomes for this group of patients with ARDS.

Table 1.

Group	Control (n = 416 [81.4%])	High ferritin (n = 58 [11.4%])	High IL-18 (n = 54 [10.6%])
SOFA score	8.0 [6.0–10.0]	9.0 [8.0–12.0]*	11.0 [8.5–13.0]*
Ventilator-free days	15.0 [0.0–22.0]	0.0 [0.0–14.5]*	7.5 [0.0–21.0]
Mortality at 28 days	21%	38%*	35%*
Mortality at 12 months	32%	52%*	44%

^* p < 0.05 compared to control.

Bronchioalveolar Lavage Microscopy Culture and Sensitivity vs Point Of Care Testing in the Diagnosis and Correct Antibiotic Therapy at 48 hours

Barnaby Lewin, Helen Surgenor, Clare MacEwen, Stuart McKechnie, Emma-Louise Jones and Lucinda Barrett

Oxford University Hospitals NHS Foundation Trust, Oxford, UK

Abstract

Background: We compared a point of care (POC) microbiological testing with microbiology, Culture & Sensitivity (MC&S) on antibiotic treatment strategies as a part of a larger QIP.

Current best practice for the diagnosis of VAPs and HAPs in the ICU is by MC&S. Diagnosis is difficult and based on clinical probabilities eg CPIS.

With clinical indication, broad spectrum antibiotics are started and narrowed at 48 hours when definitive microbiological results are available.

This QIP was to assess therapeutic benefits of pneumonia organism isolation at 4 hours vs 48 hours.

Method: All patients were intubated and ventilated on a general ICU.

Suspected VAP or intubated suspected HAPs would receive septic screens on clinical deterioration. Endotracheal aspirate (non directed Bronchoalveolar lavages) were carried out using a 20 ml syringe with 0.9% NaCl.

This sample was sent for MC&S and point of care testing.

Microbiology and Intensive Care treating teams were blinded to the results of the gene probes until the study was completed.

Consultant microbiology advised ward rounds occurred on a daily basis.

Decision to start antibiotics resides with the treating Intensivist as per trust standard care. CPIS score was not routinely used to guide starting antibiotic therapy.

MC&S results were available as standard to guide antimicrobial therapy at 48 hours or before with heavy growth.

After the trial period, clinical expertise were applied retrospectively to determine if there would have been actual therapeutic benefit of using the POC results at 4 hours vs 48 hours of standard therapy.

Results: N = 37 (VAPs and suspected HAPs)

Antibiotic therapy correct with MC&S, i.e no change to antibiotic therapy 67.6%

Lab support for not starting or stopping antibiotic therapy earlier 24.3%

Change of therapy to Correct Antibiotic (sensitivity) 5.4%

Narrowed the Antibiotic therapy 2.7%

Discussion: This small study showed agreement of Point of Care diagnostics against MC and S.

The majority of results in this study showed no change in therapeutic strategy (blind treatment was adequate whilst MC&S results were awaited).

The point of care testing showed advantage in accurate rapid assessment of VAPs to allow de-escalation to narrower antibiotic strategy and to stop antibiotics with negative test results.

Recognition of a specific organism still needs clinicians to buy in to narrowing the spectrum of antibiotic treatment, CPIS score calculation and risk assessment of surveillance BALs could be carried out to improve HAP and VAP treatment.

Larger pooled data is required to assess the accuracy of POC Endotracheal aspirates and whether it replaces the need for MC&S.

The two cases where the antibiotic was changed to one more effectively treating the pathogen may represent real therapeutic benefit.

Original CPIS

Pugin J, Auchentaler R, Mili N, Jannsens JP, Lew PD, Suter M. Diagnosis of ventilator-associated pneumonia by bacterilogic analysis of bronchoscopic and nonbronchoscopic “blind” broncoalveolar lavage fluid. Am Rev Respir Dis 1991; 143: 1121-9.

J Clin Microbiol. 2014 Jul; 52(7): 2487–2492.

Evaluation of Curetis Unyvero, a Multiplex PCR-Based Testing System, for Rapid Detection of Bacteria and Antibiotic Resistance and Impact of the Assay on Management of Severe Nosocomial Pneumonia

Wafaa Jamal,a,b Ebtehal Al Roomi,b Lubna R. AbdulAziz,b and Vincent O. Rotimia,b

Using GPICS Standards to Improve Services in a District General Hospital Intensive Care Unit

Philip Coakley, Rebecca Watson, Katie Wardle, Bijan Doostdar and Sarah Gillis

Whittington Hospital, London, UK

Abstract

Introduction: Following guidance is essential in order to provide a safe and clinically effective service. The Whittington Hospital is a District General Hospital in Inner London with a general Intensive Care Unit. The Intensive Care Society, in collaboration with the Faculty of Intensive Care Medicine, produced a structured national guideline for the provision of intensive care services. The jointly authored Guidelines for the Provision of Intensive Care Services (GPICS) were published in 2015. The decision was made to audit the Whittington ITU's own practice against national guidance.

Methods: The particular focus of the audit was the patient pathway through ITU. The GPICS chapter used for audit was ‘Admission, Discharge and Handover’. The new intervention for this audit was a new day-long audit proforma into which all documentation occurred. Admissions for a 5 week period in September-October 2017 were reviewed against the standards and compared to previous audits, which had shown poor documentation of these key standards.

Results: 31 patients were admitted and discharged during the study period. Average length of stay was 3.8 days (range 1–15). 100% of admissions were discussed with the duty consultant, and 87% had time of acceptance to ITU documented and in all cases were admitted within 4 hours of this time. 100% of patients had a documented escalation pathway on discharge, as documented on the ITU discharge summary. 74.2% of patients had a documented consultant review within 12 hours of admission (improved from 50% in previous audit). There were no non-clinical transfers to other units. A morning consultant-led ward round was documented in the notes 85% of the time, when not documented this was usually due to admission time after the ward round. An evening consultant-led ward round was documented 47% of the time. Use of the evening ward round box was variable with 49% of documented evening ward rounds recorded in the space available, but 51% recorded elsewhere. 100% of patients had daily input from nursing staff, physiotherapists, pharmacists and microbiology, regular input from dietetics, SALT, OT and clinical psychology. The nurse in charge was present on all ward rounds. All patients had a formal discharge summary and a standardised handover to the admitting team. Of the 31 patients, just 29% were discharged within 4 hours of the decision being made, 65% were not and 6% were not applicable due to death on ITU. In all cases the reason for not discharged within 4 hours was awaiting a ward bed, but in all cases discharges occurred between 07:00 and 23:59. 16% of patients were readmitted to ITU following discharge to the ward during the study period. There were no repatriations.

Discussion: This audit has showed improvement in the percentage achieval of most of the standards set out in GPICS. This was in part due to education and in part due to the documentation proforma. The main area that failed to demonstrate improvement was discharge within 4 hours of decision. This result was consistent with previous results and is likely to be due to a wider issue with patient flow and bed availability.

Conclusion: Frequent audit against national standards is a valid and useful tool for ensuring the ongoing provision of high quality care.

GPICS – available online at: https://www.ics.ac.uk/ICS/guidelines-and-standards.aspx (accessed 24/09/18).

Reporting of outcomes and measures in studies of interventions to prevent and/or treat delirium in the critically ill (Del-COrS): a systematic review

Louise Rose¹, Meera Agar², Lisa Burry³, Noll Campbell⁴, Mike Clarke⁵, Jacques Lee⁶, John Marshall⁷, Najma Siddiqi⁸ and Valerie Page⁹

¹King's College London, London, UK

²University of Technology, Sydney, Australia

³Mt Sinai Hospital, Toronto, Canada

⁴Indiana University-Purdue University, Indianapolis, The United States of America

⁵Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

⁶Sunnybrook Health Sciences Centre, Toronto, Canada

⁷St Michael's Hospital, Toronto, Canada

⁸York University, York, UK

⁹Watford General Hospital, Watford, UK

Abstract

Background: Lack of consistency in outcome selection, measurement, and reporting has been established for many conditions and leads to uncertainty with regard to the effects of interventions. Delirium in the critically ill is a common yet potentially preventable condition that poses immense burden on patients, their family, and the healthcare system. Currently there is no expert-informed guidance for researchers on the selection of a core outcome set (COS) comprising study outcomes and their corresponding measures for trials of interventions to prevent and/or treat delirium.

Objective: To inform COS development, our objective was to evaluate the scope and variability of outcomes, definitions, measures, and time-points of measurement in registered and published clinical trials of interventions, including quality improvement strategies such as the ABCDE bundle, to prevent and/or treat delirium in the critically ill.

Methods: We searched electronic databases, systematic review repositories and the international trial registry (1980 to December 2016) for studies of interventions for delirium prevention and/or treatment. Researchers working in pairs independently conducted citation screening, data extraction, and quality assessment. We extracted data on study characteristics, verbatim descriptions of study outcomes, and measurement characteristics. We assessed quality of outcome reporting using the MOMENT study scoring system, risk of bias using the Cochrane tool. and study quality using Scottish Intercollegiate network Guideline checklists. We categorized reported outcomes using Core Outcome Measures in Effectiveness Trials (COMET) taxonomy.

Results: From 158 studies recruiting 67,519 participants and reporting a mean (SD) 10 (6.2) outcome domains per study, we identified 12 delirium-specific outcome domains, with delirium incidence (116, 73%), delirium duration (53, 34%), and antipsychotic use (32, 20%) most commonly reported. We identified 95 non-delirium specific outcome domains within 19 COMET taxonomy categories. Within delirium-specific and non-specific outcome domains, we found multiple outcomes within domains due to differing descriptions and time-points. CAM-ICU plus RASS was the most common measure used (44, 36%). Measurement generally commenced on randomization or ICU admission, measurement discontinuation ranged from 24 hours to 30 days or discharge. Measurement frequency was highly variable with daily measurement reported for 32% of studies.

Conclusions: The substantial heterogeneity and multiplicity of outcome selection and measurement provides empirical evidence to support the need for COS to inform delirium research in the critically ill.

Knowledge of the Glasgow Coma Scale on a general ICU

Zoe King

John V Farman Intensive Care Unit, Addenbrooke’s Hospital, Cambridge, UK

Abstract

Background: The Glasgow Coma Scale (GCS) is a simple and objective tool used to methodically assess neurology (Bhaskar, 2017). As a widely used assessment tool on the ICU, there may be variations in technique between nurses and varying levels of understanding. Stern (2011) highlights that GCS usually receives criticism due to practitioners being required to use a numerical scale, the limitations of the categories and perceived inaccuracy of its scoring, due to those using it lacking in experience. Bledsoe et al (2014) state that GCS is inaccurate and a simpler scoring system should be established. Reith et al (2016) identified the requirement for training, education and standardisation of assessment in GCS. Al-Quraan and AbuRuz (2016) discovered that knowledge about GCS is a problem globally, and they recommend education of all nurses irrespective to the area of practice.

Method: A two- stage audit of documented GCS in medical flowsheets of current patients on a 20-bedded general intensive care unit was performed. This was followed by an online anonymised survey conducted which consisted of ten questions relating to assessing GCS and related knowledge. 125 survey requests were sent and 32 were completed, giving a response rate of 26%. This produced quantitative and qualitative data regarding nurses’ knowledge of the GCS and its application.

Results: The initial two-stage audit of observations showed that nurses of varying levels of seniority were assessing GCS incorrectly and in inappropriate patients where deep sedation was a confounding factor. In the subsequent survey, although 78% of nurses indicated that they were confident in assessing GCS, a gap in knowledge was identified, correlating with the results shown in the previous audit. 22% of nurses in one question and 69% in another question provided answers indicative of lack of understanding of the significance of the different motor scores and causes of increased intracranial pressure respectively. The main limitation of the study is that teaching of GCS was being implemented during the time that the surveys were being disseminated, resulting in potential bias of the sample. The sample size was small with a fairly poor response rate. Additionally, the questions asked in the survey were not validated.

Conclusions: Considering the limitations of the study, the results do support existing research that knowledge needs to be improved, and highlights the need for further implementation in this area. There is also a repeated theme in current evidence that the method of assessing GCS must be standardised to improve accuracy of results. This research may be beneficial to other similar ICUs in demonstrating that knowledge can be lacking in nurses of all grades and practice may be incorrect, despite their confidence and opinions of their pre-existing knowledge.

Critical Care Outreach Teams: The Use of an Interactive Card Game to Explore Practitioners Views on the Ideal Composition of an Outreach Team

Dhupal Patel¹, Tom Bashford¹, Monica Trivedi¹, Jessica Bowen¹, Sam Clayton¹, Stuart Tuckwood¹, Lisa Wood¹, John Clarkson² and Charlotte Summers¹

¹Addenbrooke's Hospital, Cambridge, UK

²Department of Engineering, Cambridge, UK

Abstract

Introduction: Critical care outreach teams (CCOT) were developed in response to evidence that adverse events are preceded by deterioration in a patient’s vital signs. A consensus model for outreach is lacking and services across the UK have been implemented variably, and differ widely with regard to service provision, team composition, and hospital coverage. Identifying the optimal model of CCOT and understanding the associated financial implications may improve our ability to deliver this service and maximise its impact.

Aim: To explore practitioner’s views on the ideal composition of a CCOT, using resource constraints as a tool to drive discussion.

Methods: An interactive card game was used to explore the ideal CCOT composition, and subsequent changes when a cost restriction was imposed. Six different medical identities (medical registrar (STR), Consultant Physician, ICU STR, ICU Consultant, outreach nurse, and advanced critical care practitioner (ACCP)) were developed, each with skills matching the minimum expected from that role. A relative financial value was assigned to each character. We approached several different medical and non-medical practitioners and initially asked them to choose five cards representing their ideal CCOT. A financial constraint was then imposed, and participants were asked to rationalise their choices and explain their underlying reasoning. Initial and final choices were collected for each participant using voice recording, with consent, and subsequently transcribed. Analysis was performed using Microsoft Excel (Redmond, USA) and diagrams created using SankeyMATIC.

Results: Clinical personnel involved in acute care including CCO nurses, doctors, and resuscitation officers, and non-clinical personnel working in patient safety and management participated. Most participants chose an outreach nurse to assist with practical tasks, followed by an even distribution of ICU Consultant, ICU STR, and ACCP, based on their advanced skills. The Consultant Physician and medical STR were chosen the least frequently (Fig1). The imposition of a cost restraint highlighted a pyramidal team structure with a sole decision maker, a second tier of advanced skills embodied by either the ICU SpR and/or the ACCP, and a base consisting of outreach nurses. Insights from the qualitative analysis included the need for retaining senior end-of-life decision making and communication skills within the CCOT.

Discussion: The landscape and demands of CCO are changing, reflecting the rise in a complex and ageing population. Communication skills and decision making around escalation of care and appropriate limitations are increasingly valued, with less emphasis on acute resuscitation skills. Participants who initially selected physician representation did so for their perceived knowledge of the patient and their background, yet considered this dispensable if financially restricted. There are no nationally agreed recommendations for the minimum skill-set a CCOT must possess, nor the activities they should undertake. CCO providers appear aware of the skills they feel are valuable in a CCOT, but there is divergence regarding the personnel best embodying these. We suggest that CCOTs may be optimised by focussing on particular skills required to manage commonly encountered scenarios, rather than a ‘one-size-fits-all’ model based on job titles.

*Authors contributed equally.

Figure 1.

Sankey diagrams demonstrating ideal and rationalised card choices

Factors associated with admission to the Intensive Care Unit following referral: prospective examination of referrals to a Critical Care unit in a tertiary centre

Nikki Faulkner and Ram Matsa

University Hospital North Midlands, Stoke-on-Trent, UK

Abstract

Background and aims: In a time of increasing austerity intensive care beds are a scarce resource. There exist no national guidelines determining which patients should be accepted for treatment in Critical Care, although some regions have local policies. It is widely accepted that patients who may benefit from admission to the intensive care unit would have reversible pathology and could be reasonably expected to survive an ICU stay with an acceptable quality of life.

There are several scoring systems which are used to grade the severity of acute illness however none of which have been validated for use before admission. The mainstay of the decision regarding whether or not to admit a patient is clinical judgement of individual clinicians.

We have evaluated referrals for admission to critical care with the aim to compare characteristics between the groups of patients who were accepted to critical care, those deemed not to be requiring critical care input and those deemed to be too unwell to benefit from ICU. The aim was to assess the factors that influence ICU admission and the patient outcomes.

Methods: A prospective evaluation of ICU admissions during specific shifts of the author of the study over the course of 3 months. Data was collected regarding patients age, WHO performance status, co-morbidities, diagnoses and organs which required support. Follow up data collected included mortality and length of ICU stay.

Results: A total of 94 patients were referred to critical care in the timeframe of the study. Of these 70% (66) patients were declined admission, with 41% (39) of patients not requiring ICU care at the time of referral and 29% (27) of patients deemed to be too unwell or physiologically frail to benefit.

Median age (interquartile range) was 67 (54–79) years, 52 (60.8%) were male. Age, gender, ethnic origin, organ to be supported and markers of acute physiological derangement did not impact on likelihood of admission to critical care.

WHO performance status significantly altered likelihood of admission to ICU with odds ratio (95% CI) for admission of 3.1 (1.72 to 5.56) for patients with performance status 0 (p < 0.001) to odds ratio of 0.28 (0.1–0.76) for patients with performance status 4.

Presence of pre-existing cardiovascular (P = 0.003), renal (P = 0.004) and neurological (P = 0.033) co-morbidities or disseminated malignancy (P = 0.021) also significantly decreased the likelihood of admission.

Conclusion: Age was not a factor in determining admission to ICU in our unit. Poor baseline function as determined by WHO performance status, as well as pre-existing co-morbidities significantly reduced the likelihood of admission.

To a Better Recovery: Developing a patient orientated multidisciplinary Critical Care Rehabilitation service

Joanne Steele, Lucy Powell and Ram Matsa

University Hospitals of North Midlands, Dept. Critical Care, Stoke-on-Trent, UK

Abstract

Background: Critical illness has significant impacts on physical, psychological and cognitive dysfunction in the short and long-term. Evidence suggests that morbidity could be significantly reduced through a structured multidisciplinary approach to care [NICE, 2009]. Recommendations from NICE [2009] and GPICS [Faculty of Intensive Care Medicine, 2015] act as drivers to promote high impact early rehabilitation for best patient outcomes. However, resource restrictions, culture and financial implications remain challenges in developing and delivering such services. The purpose of this project is to demonstrate how we as a team developed an integrated patient-centred, multi modal, multidisciplinary rehabilitation team despite such challenges.

Methods: A team with a rehabilitation focus was formed that included medical, nursing and physiotherapy staff. Strategic and stakeholder meetings were held as well as a comprehensive scoping exercise to agree the rehabilitation pathway and best practice. There was a recognition of the need to shift from a predominantly physiotherapy led rehabilitation service to a fully integrated multidisciplinary approach. A three-month pilot weekly multidisciplinary rehabilitation ward round demonstrated a reduction in mechanical ventilation days. The ‘perfect rehabilitation team’ was agreed including Physiotherapists, Occupational Therapists, Medical staff, Speech and Language Therapists, Dietitians, Clinical Psychologist and a Rehabilitation Coordinator to oversee the delivery of the service.

Result: A successful business case was granted to provide a fit for purpose, fully funded multidisciplinary Critical Care rehabilitation workforce, enabling improved patient outcomes with a focus on the development of an enhanced rehabilitation pathway. Seven-day therapies services, patient-centred goal setting, timely specialist referrals and numerous other strategies have resulted in reductions in patient length of stay and released capacity to meet upcoming departmental demands. Provisional outcome data is as documented [Table 1]. Further data is awaited which evaluates the service in full.

Conclusion: Whilst challenging, we have developed a new integrated multidisciplinary rehabilitation service, focussed on improving patients’ functionality, reducing the physical and psychological impact of critical illness, reducing overall length of stay and ensuring compliance with NICE [NICE, 2009] and GPICS [Faculty of Intensive Care Medicine, 2015] rehabilitation recommendations.

Table 1.

Outcome Description	Pre- Rehab Service	Post Rehab Service	Achievement
Manchester Mobility Score at Critical Care discharge	6	7.35	22.5% improvement. 53% improvement since Dec 2010
8 hourly assessment of ICU Delirium using CAM-ICU screening tool	<50% of applicable patients assessed	>85% of patients identified at risk of delirium measured & treated	Enabling early interventions to manage delirium
Nutritional Adequacy	25% of patients achieved 80% of their estimated nutritional requirements	52% of patients achieved 80% of their estimated nutritional requirements	27% improvement
CG 83 and Quality Standards 158	Non-compliant	Partial compliance	Risk Assessment & goal setting within 4 days of admission, ‘At risk’ patients receive formal handover of care on discharge to ward & additional information

Partial Anomalous Pulmonary venous Return (PAPVR) Case Report

Raghda Abed, Guy Rousseau and Mark Meller

Northern Devon Health Care Trust, Barnstaple, UK

Abstract

Background: Partial anomalous pulmonary venous return (PAPVR) is a vascular anomaly when some of the pulmonary veins connect to the right atrium or one of its venous tributaries rather than the left atrium. Approximately 10% are left sided. An isolated PAPVR is usually small, without haemodynamic compromise and rarelyrequires surgical correction. It’s rarely seen in adults and more commonly an incidental finding in asymptomatic patients undergoing pulmonary vascular studies for other indications.

Clinical case: A 61 year old male, presented with a one month history of severe shortness of breath and chest pain. He had a background of rheumatoid arthritis.

Clinically the patient had crepitations bilaterally and was hypoxic. A CTPA ruled out pulmonary embolism, but showed ground glass changes on both lungs.

Pneumocystis Carinii was grown from sputum culture. He continued to deteriorate and was transferred to ICU.

Central venous cannulation of the right internal jugular vein(IJV) under ultrasound guidance was complicated by an accidental arterial puncture. Subsequent Ultrasound-guided cannulation of the left IJV was uneventful, and the lumen of the left IJV was noticeably more distended and easier to locate than the right.

Post procedure CXR showed the CVC tracing a path along the left heart border.

With the patient on 2litres of oxygen via a nasal cannulae, a blood sample taken from the CVC showed a pO2 of 37.2 kPa and pCO2 of 4.5 kPa, the peripheral arterial blood sample (taken simultaneously), showed a pO2 of 9.07 kPa and pCO2 of 4.6 kPa. The transduced waveform from the central line had the appearance of pulmonary artery trace.

The CT pulmonary angiogram (CTPA) which was performed to exclude pulmonary embolism a few days prior to CVC was reviewed and on closer inspection revealed an anomalous left upper lobe pulmonary vein draining into the left subclavian vein.

Conclusion: In our case the left IJV was more filled and easier to cannulate compared to the right. The oxygen saturation from CVC blood sample was higher than that of the arterial line and the wave form similar to that of pulmonary circulation. Such findings together with the radiological evidence of abnormal CVC placementcaused a certain amount of confusion as to the actual location of the CVC. We hope that this report will highlight PAPVR as a possible yet rare cause of CVC “displacement” and anomalous venous blood gas analysis results.

Prediction of potential organ donation after circulatory death in Scottish patients

Aashray Lal¹, Radha Sundaram¹, Scott Cormack², Colin Faichnie², Susan Hannah² and Lesley Logan²

¹NHS Greater Glasgow and Clyde, Glasgow, UK

²NHS BT, Glasgow, UK

Abstract

Background: The Scottish government has set an ambitious target to increase authorisation rate for donation after circulatory death (DCD) to 80% from 57% in 2016. The process is complex and involves referral, screening, estimation of time to asystole after withdrawal of life saving treatment (WLST) and family approach. There is evidence that if we streamline this process by using a screening tool initially and then accurately estimate time of death after WLST, we can make family discussions more meaningful and secure their confidence in the process, increase referrals and reduce DCD stand downs.

Aim: To improve the DCD process by increasing the accuracy of estimation of imminent death, a crucial factor that determines whether the referral proceeds or not.

Methods: We performed a retrospective observational study looking at all DCD referrals for a period of a year in Scotland. The data is collected routinely by the specialist nurses in organ donation at referral and thereafter. A univariate analysis was performed to confirm any statistically significant correlation between diagnoses and time to asystole (within 240 min) and six physiological parameters namely vasopressor use, cough, gag, mode of ventilation, sedation and percentage of inspired oxygen. Fishers exact t test was used because of the small numbers. There was a total of 76 referrals with 64 becoming asystolic after WLST and 55 proceeding to successful donation.

Results: A total of 76 referrals, the median age amongst the adult patients was 59 and there were 47 males and 29 females. When accounting for multiple testing and so testing against a significance level of 0.008, there was no statistical evidence to suggest any of the variables significantly influence whether a DCD patient dies within 4 hours of WLST in this dataset.

Variable		Asystole within 240 minutes	Not asystolic within 240 minutes	Fishers exact t test
Ventilation	spontaneous	23	9	0.03
	mandatory	42	2
Inspired oxygen concentration	Less than 50%	57	12	0.5885
	More than 50%	7	0
Vasopressor	no	33	7	0.759
	yes	31	5
Presence of cough	no	21	3	1
	yes	41	8
Presence of gag	no	22	5	`
	yes	32	6
Sedation	no	29	4	0.5325
	yes	34	8

Discussion: We were unable to demonstrate any statistically significant association between physiological variables at referral and asystole within 240 minutes, a factor that increases the likelihood of a successful donation. The limitations of this study include small numbers and possibility of bias (i.e. patients not being referred because they are deemed unlikely to die within the time frame). A larger study looking all deaths in intensive care, not just referrals and using machine learning would help in predicting time to asystole, increasing not just successful DCD number but reduce DCD stand downs and improve family approach for authorisation.

Woven EndoBridge (WEB) device in the management in ruptured and unruptured intra cranial aneurysms – Is it safe?

Marcus Belasco, Ravi Chauhan, Randeep Mullhi, Catherine Snelson, Nandan Gautam, Tony Whitehouse and Tonny Veenith

Department of Critical Care medicine, University Hospital of Birmingham NHS trust, Birmingham, UK

Abstract

Background: The Woven EndoBridge (WEB) device is an intrasaccular flow disruptor used on intracranial wide-necked bifurcation aneurysms. Currently, there is a lack of data on the short and long-term outcomes of WEB embolization of ruptured and unruptured aneurysms

Methods and Materials: A retrospective descriptive cohort study of 121 patients with 124 aneurysms at a specialist centre treated with the WEB device from May 2013 to August 2017. Both ruptured (n = 63) and unruptured (n = 61) aneurysms were included. Angiographic and functional outcomes were assessed using the WEB Occlusion Scale and mRS, respectively. Complications were also recorded.

Results: Angiographic data was available for 95 and 44 aneurysms at short and long-term follow-up points, respectively. Functional data was available for 90 and 60 aneurysms at short and long-term follow-up points, respectively. The short-term adequate aneurysm occlusion rate was 82% and the long-term adequate occlusion rate was 87.2%. In the ruptured subgroup, 65.1% of patients were discharged in good clinical condition. 70.7% and 85% of patients in the ruptured subgroup had good clinical condition at short and long-term follow-up. No unruptured aneurysm patients had significant morbidity at discharge.95.9% and 85% of patients in the unruptured cohort had good clinical condition at short and long-term follow-up.

Conclusions: The WEB is a safe and effective treatment for wide-necked bifurcation aneurysms. Angiographic and functional outcomes were encouraging.

Targeted temperature management and neurological outcome after out-of-hospital cardiac arrest. A retrospective 4 year review of survivors from a large district general hospital in the UK

Christopher Pritchett¹, Joshua Thorns¹, Jez Hunter², William Yau¹, Jessica Brayley¹ and Michael Spivey¹

¹Royal Cornwall Hospital, Truro, UK

²Derriford Hospital, Plymouth, UK

Abstract

Introduction: In 2017 it was shown that target temperature management (TTM) for 48 hours vs 24 hours did not significantly improve neurological outcome.¹ Following this publication we reviewed all 186 ICU admissions for out of hospital cardiac arrest (OHCA) from 2013–2017. We aimed to assess local use of TTM before and following our local guideline after the 2013 TTM publication² and our neurological outcome of survivors alongside it.

Methods: Using the Philips ICCA system we identified all OHCA admissions to the ICU between 1/1/13 and 31/3/17. Patients aged less than 18 were excluded. Patient electronic records were read. Baseline data was collected for: age, gender, time of cardiac arrest, presenting rhythm and time of return of spontaneous circulation (ROSC). From the clinical notes it was ascertained whether TTM was explicitly considered in the management plan. Method of TTM was recorded as either active cooling (via Cool Guard) or passive. Average temperature at 12, 24, 48 and 72 hours was recorded. Neurological outcome of survivors was scored retrospectively through use of the cerebral performance category (CPC) tool against out-patient clinic letters.

Results: Of the 186 patients, 76% were male. Average age was 58.4 years. Of presenting rhythms, 52% were VF/VT, 19% asystole, 18% PEA and 11% unknown. Shortest average time to ROSC was in the VF/VT group at 27 minutes.

94% of all patients had TTM explicitly prescribed in the patient management plan. Of These, 51% were actively cooled. Use of active cooling in patients aged between 18–75 was highest in 2013 with 100% use and lowest in 2014 with 10% use. Corresponding average temperatures at 24 hours for these years were 34°C and 36.4°C respectively.

66 patients (35%) of patients survived to discharge. 46 patients underwent CPC scoring from 6 month out-patient follow up letters. This showed CPC 1–3 of 78%, 15% and 7% respectively.

Conclusions: There is a marked change in approach to TTM following 2013. This reflects the literature with the publication of TTM & the change on which local guidelines were written. Use of active cooling was surprisingly low in 2014 following guideline changes and has increased since. Despite an increasing use, our average temperatures at 24 hours were still above 36°C.

CPC data was encouraging but should be interpreted with caution. The large number of patients who did not receive an out-patient clinic follow up and were therefore unable to be scored may represent a follow-up bias in favour of those with a good neurological outcome.

We intend to review use of the Cool Guard. We aim in particular to see its use increase in those who suffer VF/VT OHCA aged between 18–75.

An educational intervention targeting ICU nurses to improve adherence to lung protective ventilation in patients after heart surgery

Tania Soriano Navarro, Ioanins Karpouzis, Kate Diomede, Julie Sanders and Mark Griffiths

St Bartholomew's Hospital, Barts NHS Trust, London, UK

Abstract

Patients after cardiac surgery are at a high risk of developing respiratory complications, including Acute Respiratory Distress Syndrome (ARDS).1 Amongst several other factors, mechanical ventilation (MV) may make a significant contribution to ARDS risk.2 This Ventilator Induced Lung Injury (VILI) may be mitigated by the use of lower tidal volumes (Vt): 6–8 ml/kg of ideal body weight (IBW) is recommended. 3 However, adherence to low Vt MV is lacking.4 We tested the effect of an education programme targeting Intensive Care nurses on MV parameters administered to patients after heart surgery (HS).

The author trained a group of ARDS link nurses that were responsible of disseminating the training to the whole ICU nursing team. MV parameters were recorded for 7 days for consecutive HS patients before and after training. Questionnaires (before and after the educational intervention) assessed nurses’ knowledge and confidence managing MV. When data was analysed Fisher’s exact test was employed for categorical variables. Data sets were subjected to the D'Agostino & Pearson normality test (alpha = 0.05). A Mann-Whitney U test was used when comparing populations that were not normally distributed; a significance level of p < 0.05 was set.

There was no difference in Vt before and after the intervention (median (IQR) day 0: 8.0 (7.1–8.8) n = 29 vs 7.4 (6.8–7.9) ml/kg IBW n = 41, p = 0.07; day 1: 8.3 (7.5–9.5) n = 10 vs 7.4 (6.8–8) ml/kg IBW n = 6, p = 0.09). However, after the intervention patients were more likely receive Vt <8 ml/kg IBW (OR = 0.3; 95%CI 0.11–0.85; p = 0.02) on day 0. Females were more prone than males to benefit from the intervention: females pre-intervention audit (n = 7) 8.2 (7.6–10), post- intervention audit (n = 8)7.4 (6.8–7.8) (p = 0.04). An increase in nursing documentation of IBW and Vt by IBW was observed (OR (95%CI) at completing IBW 0.13 (0.04–0.47) and completing Vt 0.11 (0.02–0.51); p < 0.01 for both). The proportion of nurses’ who was aware of the Vt upper limit did not improve after the intervention (before 30% and after 32 % OR (95% CI) 0.99 (0.47−2.14); p > 0.99). However, a higher proportion of nurses knew the Peak inspiratory pressure (Ppeak) upper limit after the educational update (before 51% and 76% after intervention OR (95%) 0.34 (0.16–0.72); p = 0.04). Confidence with MV settings and challenging the doctors were assessed. The reason for those levels of confidence was explored in the questionnaires.

Despite gaps in nursing knowledge, fewer patients were ventilated with presumed injurious Vt (>8 ml/kg IBW) after targeted training. Continuous nursing education may decrease the incidence of ARDS by increasing adherence to evidence-based care in high risk patients.

The author received funding from Barts Charity in order to complete MSc studies. This study received ethics committee approval or equivalent.

Code Black: An evaluation of a standard operating procedure to expedite neurosurgical intervention in traumatic brain injury

Philippa Gleed, Parjam Zolfaghari and Akshay Patel

Royal London Hospital, Barts Health NHS Trust, London, UK

Abstract

Introduction: Traumatic brain injury (TBI) constitutes a significant health and socioeconomic problem, being a leading cause of death and disability. TBI is a time-critical injury, demanding clinicians act decisively to prevent secondary progression, and mitigate further disability. Consequently, the Code Black Trauma Team Response Standard Operating Procedure was implemented at the Royal London Hospital (a Major Trauma Centre- London) in July 2015. This protocol aims to expedite rapid imaging, assessment and transfer of critically brain-injured patients requiring emergent neurosurgical decompression from the emergency department to an operating theatre for definitive evacuation of surgically amenable mass lesion(s).

Aims: The primary aim was assessment of the Code Black Trauma Team Response Standard Operating Procedure in reducing time to surgical decompression of amenable mass lesion(s). Secondary outcomes include effect on survival at 30 days and thirdly, to detect any changes in demographics of the severe TBI population.

Methods: A retrospective case-control study was carried out between July 2013 and July 2017 on severe TBI patients (defined by Head AIS >2 and GCS ≤8) at the Royal London Hospital. Patient data was obtained from a combination of the Trauma Audit Research Network (TARN) and Collector databases as part of a local audit and quality improvement project, and was deemed not to require ethics approval.

Results: 770 patients were included for analysis and categorised into three cohorts: 308 patients presenting 1^st July 2013-15^th July 2015, 300 patients presenting 16^th July 2015-31^st July 2017 and 162 patients managed under Code Black between July 2015-17. Median time to craniotomy was 1.8 hours (0.4–22.7 h) in 2013-15, 1.1 (0.1–19.2 h) in 2015-17 and 0.8 (0.1–17.6 h) for Code Black patients, p < 0.001. For those patients who were operated on, the survival rate at 30 days was 67% in 2013-15 and 72% for Code Black patients (not statistically significant). No significant differences in demographics were observed (Table 1).

Conclusion: This study demonstrated significant reduction in time to surgery for patients with severe TBI following implementation of the Code Black protocol. However, this did not confer improvement in short-term mortality. Whilst mortality is an important outcome, in this population further studies should focus on the impact of expedited surgery on neurological outcomes

Table 1.

Severe TBI patient demographics at the Royal London Hospital 1^st July 2013- 31^st July 2017.

		1st July 2013-15th July 2015	16th July 2015- 31st July 2017
Age		47 years (18–95y/o)	46 years (18–95y/o)
Gender		79% M, 21% F	78% M, 22% F
Mechanism of Injury	Fall <2 m	22%	26%
	Fall >2 m	22%	21%
	Vehicle Incident	45%	40%
	Crush	1%	0%
	Blow(s)	6%	9%
	Stabbing	1%	1%
	Shooting	0%	1%
	Other	3%	1%
Mode of Arrival	Helicopter	28%	21%
	Ambulance & Helicopter	1%	0%
	Ambulance	70%	77%
	Other	1%	1%
Head AIS	3	8%	7%
	4	24%	14%
	5	68%	79%
ISS		30 (9–66)	33 (9–75)
ICU Length of Stay		9 days (0-67d)	10 days (0-200d)
Hospital Length of Stay		26 days (1-204d)	27 days (1-371d)

M – Male, F- Female, AIS- Abbreviated Injury Score, ISS- Injury Severity Score..

The need for Continuous Veno-Venous Haemofiltration in patients in intensive care after cardiac surgery at Papworth Hospital

Caitlin Pley¹, Katie Purohit¹ and Stephen Webb²

¹School of Clinical Medicine, University of Cambridge, Cambridge, UK

²Royal Papworth Hospital, Papworth Everard, UK

Abstract

Acute kidney injury (AKI) is common in intensive care patients, with renal replacement therapy (RRT), associated with the highest mortality risk(1), required in 4–5% of ICU admissions. Continuous Veno-Venous Haemofiltration (CVVH) is a type of continuous RRT, indicated for severe AKI, that has been shown to improve patient outcomes(2). It can be maintained for up to several weeks and relies on ultrafiltration and convection for effective clearance. Continuous renal replacement therapy is preferred to intermittent options because it allows for greater stability and predictability in volume control, blood chemistry and intracranial pressure, and can be used in patients in whom conventional RRT is contraindicated due to hypotension(3). Limitations of CVVH include the requirements for central venous access, continuous anticoagulation and immobilisation. In intensive care at Papworth Hospital, a tertiary cardiothoracic hospital, CVVH is used for a diverse population of patients who have developed severe AKI post-cardiothoracic surgery. For cardiac surgery patients with AKI, CVVH has been shown to increase both in-hospital and long-term survival, and 96.3% of survivors recovered renal function in one study(4). However, mortality amongst cardiac surgery patients requiring CVVH for AKI remains high at 38.8%(5). Between April 2014 to April 2015, Papworth introduced a new fluid policy, which aims to improve the recognition and management of AKI, in order to prevent progression. The policy recommends AKI rounds together with pharmacists, and standardised fluid regimes for common post-surgical clinical situations, including an alert/oriented patient, a drowsy/nauseous/vomiting patient, a frail/elderly/cardiac failure patient, and a diabetic patient.

This study aims to examine the proportion of patients requiring CVVH following cardiac surgery. Furthermore, this study investigates whether the implementation of new fluid guidelines between April 2014 and April 2015 has effected a change or trend in the requirement of CVVH in cardiac surgery patients at Papworth. This study includes data from 7850 patients from January 2015 to August 2018 inclusive.

In the time period of January 2015 – August 2018, 533 out of 7850 cardiac surgery patients at Papworth required CVVH to treat their AKI in intensive care. The data also showed that following the implementation of the standardised, situationally-guided fluid protocol, there appeared to be an increase in the variation amongst the monthly values. By grouping the data together into 6-month periods, a clearer trend emerges (Figure 1). Following the implementation of the fluid protocol, there appears to be an initial rise in the number of cardiac surgery patients requiring CVVH therapy, with a subsequent decrease to lower levels than before the change in fluid guidelines. This may indicate that the change in fluid guidelines engendered initial problems, possibly due to human factors in the process of changing clinical practice. However, in the long-term, the new guidelines may have had a positive effect on the incidence of severe kidney injury requiring CVVH in these patients.

The incidence of CVVH use in cardiac surgery patients at Papworth is 6.8% from data collected from 7850 patients (January 2015-August 2018). A significant change in fluid guidelines has produced interesting trends in this data.

A Case of Peripartum Cardiomyopathy Supported by VA-ECMO

Alexandra Baker, Francisca Caetano, Susanna Price, Shahana Uddin, Richard Trimlett and Deepa Arachchillage

Royal Brompton and Harefield NHS Foundation Trust, UK, London, UK

Abstract

Introduction: Peripartum cardiomyopathy (PPCM) is a rare, severe and idiopathic type of heart failure in previously healthy young women, during late pregnancy or immediately after delivery.

In severe cases refractory to full medical therapy, mechanical cardiovascular support may be needed as a bridge to recovery or transplantation.

Case Report: A 26-year-old, 38-week pregnant woman (G7P2) presented to obstetric triage after 4 weeks of chest pain (treated as reflux) and worsening oedema. She was found to be hypertensive, oedematous to her waist and on lying flat became acutely hypoxic. She immediately underwent emergency intubation, ventilation and caesarean section resulting in delivery of a live infant. A transthoracic echocardiogram was performed in theatre, which revealed a moderately dilated left ventricle (LV) with secondary mitral regurgitation and severe left ventricular systolic dysfunction, ejection fraction <30%.

Post-operatively the patient remained in a low cardiac output state, despite escalating inotropic support, culminating in a cardiac arrest with successful resuscitation. She was retrieved to an ECMO centre on Veno-Arterial Extracorporeal Membrane Oxygenation (VA-ECMO). An Intra-Aortic Balloon Pump (IABP) was inserted for LV off-loading.

Transoesophageal echocardiogram (TOE) showed a large RA thrombus (in association with the drainage cannula) and a possible patent foramen ovale (PFO) increasing simultaneously the risk of pulmonary embolism, but also of stroke due to paradoxical embolism, particularly during ECMO decannulation. She thus required systemic anticoagulation, but due to bleeding complications (from the caesarean wound) only a low heparin was maintained.

To treat the PPCM she was loaded with levosimendan and commenced on bromocriptine. Serial echocardiography demonstrated improvement in biventricular function. On Day 7 post-ECMO cannulation a bubble study excluded any intra-cardiac shunt, and the patient was decannulated from ECMO. Subsequently, the IABP was removed and she was extubated. By Day 11, her LV was of normal dimensions, and the ejection fraction >40%. She was repatriated to her local hospital on chronic heart failure therapy and under surveillance of the heart failure team.

Discussion: This case demonstrates several challenges in the assessment and management of PPCM. First, the non-specific nature of its presenting symptoms can make it difficult to differentiate from physiological and symptomatic changes in pregnancy. Second, PPCM can be difficult to differentiate from other causes of cardiomyopathy diagnosed during/after pregnancy. Third, a pregnant patient presenting with cardiogenic shock may have additional causes for circulatory collapse, including pregnancy-induced hypertension and eclamplsia, or amniotic fluid embolism. Finally, although increasingly used and recommended in current guidelines, the effects of bromocriptine remain difficult to predict in individual patients.¹

VA-ECMO has been described as a management option in patients refractory to conventional medical treatment, INTERMACS I, as a bridge to cardiac recovery or transplantation. Its use must be weighed against potential complications, including stroke, infection, circuit failure, bleeding and thrombosis. Treatment with bromocriptine and the peripartum state put the patient at high risk of thrombus formation. Therefore, collaboration with haematologists and imaging experts (to diagnose/exclude thrombus and intra-cardiac communication) is essential to mitigate the risk of paradoxical embolism resulting in stroke in this young cohort of patients.²

Echocardiography in the evaluation of sepsis-induced cardiomyopathy managed with Venoarterial Extracorporeal Mechanical Oxygenation (VA-ECMO)

Dr Hatem Soliman¹, Marco Morosin¹, Stephane Ledot¹, Francisca Caetano¹, Richard Fisher², Shahana Uddin² and Susanna Price¹

¹Royal Brompton and Harefield NHS Foundation Trust, UK, London, UK

²Kings College Hospital, London, UK

Abstract

Abstract: Septic shock is reported to have a mortality of 25−45%.¹ Although up to 64% of patients with sepsis are found to have some degree of concomitant myocardial dysfunction, overt sepsis-induced cardiomyopathy (SiC) is unusual and is associated with significantly increased mortality. For those patients who survive, SiC usually resolves within 7 to 10 days^2,3. Therefore, it may be reasonable to consider aggressive therapies, including veno-arterial extracorporeal membrane oxygenation (VA ECMO), to support organ perfusion in this subset of patients who have the potential for recovery.

Patient Profile: A 48-year-old man presented to his local hospital with dyspnoea and fever where a diagnosis of influenza B was made. He was intubated and ventilated for severe acute respiratory failure. Additional investigations: electrocardiogram demonstrated ST-T segment changes, Troponin was elevated. Coronary angiogram showed angiographically unobstructed coronary arteries. Initial FoCUS showed very poor biventricular function (LV ejection fraction 15% with global hypokinesia and LV mural thrombus). He was started on dopamine and adrenaline, but remained haemodynamically unstable and therefore an intra-aortic balloon pump was inserted. As he remained INTERMACS I, he was referred for VA-ECMO.

Clinical Course: Trans-oesophageal Echocardiogram (TOE) after commencing VA ECMO showed poor biventricular function with spontaneous echo contrast in the LV (Figure 1), mild mitral regurgitation and no aortic valve opening. Further laboratory investigations showed RSV A/B RNA +ve, Legionella urinary antigen +ve and Streptococcus pneumoniae +ve. He was started on antimicrobials and Intravenous Immunoglobulin. He remained on VA ECMO for two weeks during which he developed large pericardial effusion (2.2 cm) (Figure 2) with no compromise in ECMO flows. The pericardial effusion was closely monitored until it resolved completely by the time the heart started to recover around day 14, at which stage he was decannulated. He was later extubated and discharged home after full recovery and with a normal echocardiogram.

Discussion: SiC is one of the major predictors of morbidity and mortality in septic patients increasing mortality rates from sepsis from 38.6% to 70%, and can be so severe that it can mimic cardiogenic shock. VA-ECMO is increasingly used in adults with SiC, with survival and complete cardiac recovery reported in up to 70% of patients.

Conclusion: Expert echocardiography is a valuable tool in the diagnosis, cannulation and monitoring of patients who are managed with VA ECMO, as the interpretation of relatively ‘simple’ findings (ventricular function, pericardial collection, tamponade) may be particularly challenging in this patient population.

Figure 1. Mid-oesophageal four-chamber view showing dilated LV with spontaneous echo contract in the LV cavity

RA; right atrium, LA; left atrium, RV; right ventricle, LV; left ventricle

Figure 2. Deep trans-gastric short axis view showing the large pericardial effusion collection

LV; left ventricle, RV; right ventricle

Stabilisation of unwell children by Anaesthetists and Intensivists in a DGH setting

Fui Xuan Pong and Saibal Ganguly

New Cross Hospital, Birmingham, UK

Abstract

Introduction: New Cross Hospital is a busy District General Hospital (DGH) where there is no Paediatric Intensive care unit (PICU). A multidisciplinary team (MDT) comprising of ED physicians, Paediatricians, Anaesthetist and Intensivist manage unwell children in Emergency Department (ED) and wait for Paediatric Intensive Care Retrieval (PICR) team to transfer to a hospital with PICU facility.

Aims & Objectives: The aim of the study was to identify and compare the numbers of unwell children who were intubated and ventilated in ED between 2016 and 2017. We also reviewed the average time spent by the MDT in the ED.

Methods: It was a retrospective study. Data selection was done by Information Technology (IT) team from ED. Children that required airway support and those who were transferred out of the hospital were selected. Mann-Whitney and Chi-Square tests were used to calculate statistical significance difference.

Results: Paediatric ED attendances were 23,456 and 23,273 in 2016 and 2017 respectively. Among these, there were eight and eleven children in 2016 and 2017 respectively needed intubation and ventilation and were retrieved by the PICR team. The age group in 2016 ranged from two months to fourteen years and one month to fourteen years in 2017. Children who got intubated and ventilated presented with various aetiologies such as status epilepticus, cardiac arrest, meningitis, respiratory failure secondary to acute lower respiratory infection, head injury and multiple trauma. The time spent in 2016 was between two hours and thirty-nine minutes to ten hours and thirty-eight minutes. In 2017, it was between two hours and fifty-two minutes to nine hours and two minutes. There was no statistical significance difference in the incidence of paediatric intubation (p: 0.481) between the two years. Mann-Whitney test showed there was no statistical significance difference either for the time spent by MDT in ED between the two years (p: 0.600).

Conclusion: From the cases identified, the admission rates and the incidence of airway management between the two years did not significantly increase, but the number of hours spent by MDT in ED was significant. We agreed that managing an unwell child who would need intubation and ventilation does not occur frequently. However, when it happens, the hospital’s resources are stretched for a few hours before retrieval team arrives.

Recommendation: With the specialist centre’s bed occupancy remaining high, district general ICU’s might consider admitting critically ill children for stabilisation with adequate support from the tertiary centre’s rather than transferring from ED. We are mindful that this would require an extensive training, teaching and reconfiguration of an existing workforce.

Improving the initial recognition and management of Acute Kidney Injury

Kaenat Mulla, M. Asim T Khan, Hannah Cullumbine, Kate Fairman, Jamil Chowdhury, Sheeraz Iqbal, Jack Herlihy, H Karunanratne, Sara Davison and K Aziz

Barking, Havering Redbridge NHS Trust, Romford, UK

Abstract

Acute Kidney Injury (AKI) can present in up to 18% of hospital admissions and is associated with increased all-cause morbidity and mortality rates. It is still under-recognised and inadequately managed. A 2009 NCEPOD study revealed that only 50% of patients with Acute Kidney Injury (AKI) receive “good” care. Currently, it costs the NHS £434 million per year to treat AKI which represents 1% of the NHS budget, and it is thought that improving the current standards of care could save the NHS £150 million per year.¹

For the pilot study we reviewed all general medicine admissions at a district general hospital over the course of 10 consecutive days in September 2017 (n = 205). We identified which patients had an AKI based on the gold-standard definition set by the Acute Kidney Injury Network (AKIN). Cases were evaluated to determine whether investigations and management were in line with recent NCEPOD recommendations and NICE guidelines.² Once we identified areas for improvement, a driver diagram was developed to enable serial analysis of further PDSA cycles. We conducted 7 PDSA cycles and introduced various interventions; we educated doctors and nurses about AKI, displayed posters on wards, implemented proformas, emailed reminders, and presented at grand-rounds. We also implemented questionnaires to help understand observed trends.

Our results show a gradual but significant improvement in all defined areas for AKI recognition, assessment and management. Continuous encouragement is needed to meet targets, as demonstrated by minor fluctuations in standards. 10% of patients admitted had an AKI and this was in keeping with the national prevalence. The average age of patients with AKI was 70 years, with a 3:2 male-to-female ratio. 31% had a background of chronic kidney disease. 68.5% had stage 1 AKI and this was primarily due to pre-renal causes such as hypovolaemia. 14% had stage 3 AKI.

The AKI proforma sticker was underutilised, perhaps because only 44% of doctors surveyed were aware of its existence. However, when the AKI proforma was used, 100% of patients were correctly identified to have an AKI and received a 48-hour review. There was also an 8% improvement in number of patients having a fluid balance assessment. Recognition and cessation of nephrotoxic agents is poorly done, and a survey revealed that only 33% of doctors recognised nephrotoxic agents from a list of medications. We noted a deviation in standards at one point in the cycle, which was attributed to winter pressures and junior doctor changeover. We suggest that further education and use of the AKI proforma sticker will reduce AKI associated morbidity, mortality and economic burden. The sticker is cheap to print and will act as a checklist for doctors and nurses.

A quality improvement project to improve the documentation of resuscitation and ceiling of care decisions in elderly patients at a large district general hospital

M. Asim T Khan, M Hamza V, Huiyi Esther Law and Raidah Haider

Barking, Havering Redbridge NHS Trust, Romford, UK

Abstract

Aims: All elderly patients seen on post-take ward rounds should be assessed by the Gold Standards Framework (GSF), with resuscitation status and ceiling of care discussed and documented accordingly in the medical notes and discharge letters.

Methods: We opportunistically reviewed the medical notes and discharge letters of all patients aged over 75 years immediately prior to discharge. We conducted three cycles of data collection, each lasting one week in duration, from November 2017 to January 2018. Data was recorded into an anonymised excel template. We determined a pre-intervention baseline in cycle one, with a view to compare the trend after implementing two interventions separately in the next cycles. In cycle two, we delivered a presentation at the quarterly geriatrics meeting. In cycle three, we emailed recommendations to all doctors based at the trust. We determined the outcome of each intervention.

Results: In the pre-intervention sample (n = 35), only 51% patients had GSF documented. In cycle two (n = 45) this was 49%, and in cycle three (n = 22) this fell to 32%. In cycle one, 65% (13/20) of patients deemed ‘not for resuscitation’ (NFR) had no documented escalation status in the notes. In cycle two this was 46% (12/26) of NFR patients, and in cycle three this fell further to 14% (1/7). NFR status in cycle one was not communicated in 70% (14/20) patients’ discharge letters. In cycle two this fell to 35% (9/26) and rose to 100% (7/7) in cycle three. Seven patients met the GSF criteria in cycle one, 13 patients in cycle two, and four patients in cycle three. Regarding GSF patients in cycle one, 28% had no documented resuscitation discussion and 57% had no documented ceiling of care. In cycle two all GSF patients had a documented resuscitation discussion and only 25% had no documented ceiling of care. In cycle three, 75% had no resuscitation discussion documented and 25% had no documented ceiling of care.

Conclusion: The study demonstrates an overall improvement in documentation, particularly with the second cycle intervention. However, there is a relapse in standards in cycle three. This implies that the direct delivery of teaching sessions is far more effective than blanket emails. It also suggests that interventions need repeated administration to achieve the desired adequacy in documentation. The subpar documentation in cycle three may reflect the increased winter pressures at this point in the study and may also be confounded by the low discharge rates reducing sample size.

Successful treatment of Capnocytophaga sepsis in a patient with polycythaemia rubra vera

Helen Church, Anand Padmakumar, Sawsan Awad, David Cambray-Deakin, Michael Collins, Gemma Wheldon and Faiha Eltayeb

Chesterfield Royal Hospital, Chesterfield, UK

Abstract

Background: A 72-year-old Caucasian Mr E was admitted to Emergency Department (ED) with upper gastrointestinal (UGI) bleeding. He had been found by his neighbours; confused, sweating and had haematemesis. He had no travel history, but had a past history of polycythaemia rubra vera (PRV) (on hydroxycarbamide and aspirin), and hypertension (on amlodipine). On examination he was tachypnoeic (RR 36/min), hypoxic and wheezy. He was hypertensive and tachycardic (HR 118 bpm, BP 172/144), apyrexial and had marked bruising on both lower legs. Initial work-up revealed acute kidney injury (AKI) (urea 31.0 mmol/L, creatinine 540 µmol/L), pancytopaenia (Hb 86 g/L, WCC 3.4x10⁹/L, Platelets 9x10⁹/L), elevated CRP (668 mg/L) and right-sided consolidation on chest-xray.

Initial diagnosis was neutropaenic sepsis with AKI from pneumonia, with a possibility of Panton-Valentine leucocidin infection (given very high CRP). He was admitted to High Dependency Unit (HDU) and initially treated with co-amoxiclav, clarithromycin and linezolid, but then escalated to ceftriaxone and gentamicin following the growth of a gram-negative bacteria on blood culture. After 24 hours his AKI worsened and was transferred to ICU for renal and vasopressor therapy. Mr E received multiple blood products to correct bone marrow suppression (WCC 3.4, Platelets 11 x 10⁹/L).

On day 2 of ICU stay, vasopressor requirements increased and skin bruising spread more proximally, resulting in extensive skin exfoliation and peripheral dry gangrene. Blood cultures failed to cultivate an organism, but 16S rRNA PCR molecular technology identified Capnocytophaga canimorsus. Collateral history revealed Mr E had been bitten by a dog, which had left two excoriations but had healed without medical intervention. By day 6, he made good recovery. He was transferred to a tertiary hospital for haemodialysis for resolving AKI.

Discussion: Capnocytophaga canimorsus is a gram-negative commensal bacillus found in saliva of healthy dogs (16–25%) and cats (17%), transmitted to humans through bites, scratches or licking of the skin. Clinical presentation includes septic shock, gangrene of extremities, skin failure, meningitis, endocarditis, and eye infections (1). AKI, respiratory distress, disseminated intravascular coagulation, GI symptoms and pancytopenia have also been reported (2–4).

Although general immunocompromise is an acknowledged risk factor for C. canimorsus infection, the literature focuses mainly on patients who have had splenectomy and alcoholism (1). Other case reports have revealed a broader range of immunocompromised states in which C. canimorsus sepsis was confirmed. However to the best of our knowledge, this is the first reported case of C. canimorsus sepsis in a patient with PRV.

Conclusion: Where C. canimorsus causes sepsis it has a short incubation (24–72 hours) from exposure (e.g. dog bite) and a high mortality (19–36%) (1,2,5). However, laboratory confirmation can take several days. A high index of clinical suspicion, especially in immunocompromised splenectomised patients, and specifically looking for animal bites might direct appropriate early institution of antimicrobial therapy, hence reducing morbidity and mortality. Prophylactic antibiotics should be considered in high-risk immunocompromised patients following exposure to animal bites.

Validation of the Hospital Frailty Risk Score in the ICU

Dominic Marshall¹, Justin Salciccioli², Robert Hatch¹ and Matthew Rowland¹

¹Critical Care Research Group, Nuffield Department of Clinical Neurosciences, Oxford, UK

²Department of Medicine, Mount Auburn Hospital, Harvard Medical School, Cambridge, The United States of America

Abstract

Background: Frailty is a syndrome of chronic illness, reduced physiological reserve and decreasing ability to function independently. It is associated with aging and strongly associated with prognosis. However it is often difficult to measure frailty as it is often dependent on subjective assessment or collateral history. Gilbert et al recently generated a novel Hospital Frailty Risk Score (HFRS) using EHR which achieved moderate agreement with accepted clinical measures of frailty and identified groups of patient over 75 at greater risk of adverse events.¹ Our objective was to validate this score in the ICU population.

Methods: Data for this study were collected from the MIMIC-III database – an open source clinical database.² Gilbert et al. used ICD codes from the 10^th revision to develop the HFRS.¹ For this study, we utilized the 9^th version of ICD as this is the classification scheme available in the MIMIC-III. We used a stepwise approach to validate the use of the HFRS in general adult ICU patients >75 years of age. The primary outcome of interest was 28-day mortality. Following application of the available SAS macro code to compute the frailty risk scores, we derived univariate and multivariate logistic regression to assess the relationship between frailty score and outcomes of interest. In these first models, we treated frailty score as a single continuous variable without transformation. We then used a two-step approach: first, we computed the 33^rd and 66^th centile for frailty risk scores and categorized patients into groups based on this as low, medium and high; second, we produced graphical representations of mortality trends with increasing frailty risk scores and generated categories based on apparent changes in mortality trends. each outcome, after multivariable adjustment.

Results: 12,854 patients made up the final cohort for analysis. The median age of the cohort was 81 years (IQR 77–86) and 48% of patients were male. The median SAPS-II score was 40 (33–48) and the median SOFA score was 4 (IQR 2–6). The median Hospital Frailty Score for the entire cohort was 5.8 (IQR 3.0–9.6). In univariate analysis OR for death at 28 days was 1.86 (IQR, 1.65–2.11) for the medium risk group and 2.01 (IQR, 1.78–2.27) for high risk group compared with low risk. This association remained significant after multivariate adjustment.

Conclusions: The Hospital Frailty Risk score developed for use with EHR can define a sub-group of ICU patients >75 who are at increased risk of 28-day mortality. Results from our study do not demonstrate a significant difference between medium and high-risk frailty score suggesting that in the ICU, frailty may be a binary risk factor for adverse events in patients.

Length of stay and mortality following repatriation from intensive care at a tertiary specialist cardiothoracic centre

Katie Purohit¹, Caitlin Pley¹ and Stephen Webb²

¹School of Clinical Medicine, University of Cambridge, Cambridge, UK

²Royal Papworth Hospital, Papworth Everard, UK

Abstract

After specialist cardiothoracic care at Papworth Hospital, Cambridgeshire, patients may be repatriated; meaning they are transferred to their local hospital for further care. Due to Papworth being a tertiary specialist cardiothoracic hospital, patients often come from far in the UK, and even Ireland. The purpose of repatriation is to enable patients to be closer to friends and family, and is considered when their stay is expected to be prolonged. This study aims to investigate the outcomes, as well as the length of stay in intensive care, of patients after they have been repatriated from intensive care at Papworth. When transferring critically ill patients, extreme care must be taken as a number of serious consequences can arise(1,2). Risks of transport can be mitigated with network organisation, dedicated teams, detailed planning, training for staff and high-quality equipment(3).

This is a retrospective analysis of 568 patients repatriated from intensive care at Papworth between April 2012 and March 2018 inclusive, looking at pre-discharge mortality and average length of stay in the repatriated intensive care unit. Patients were stratified into different groups by specialty.

Of all patients, 28% died before discharge from the hospital that they were repatriated to. The difference in survival to discharge was stark between the groups. Of patients classified as cardiology non-transplant and heart failure, the proportions who died before discharge were 45% and 46% respectively. For surgical patients, 23% of pre-operative and 34% of post-operative patients died before discharge. However, for critical care medicine and respiratory ECMO patients, outcomes were much better; with only 9% and 10% dying before discharge, respectively. The average length of stay in the intensive care unit after repatriation was compared between patients who died before discharge and those who survived to discharge. For all patients, the average length of stay of those who survived to discharge was 9.8 days, and for those who died before discharge the average length of stay was 9.3 days. This difference is not statistically significant. Within the different groups, the mean length of stay between those who survived and died was more divergent, but not statistically significant due to large variation. When comparing the different groups to the total mean length of stay, it is interesting to note that respiratory patients on ECMO stayed in intensive care much longer, for 11.7 and 16.9 days respectively for those who survived and died. Cardiology patients (not including transplants) had much shorter stays, with 5.3 and 4.0 days respectively.

Patients who were repatriated from a tertiary specialist cardiothoracic centre had a high mortality rate before discharge. Mortality was especially high for cardiology and heart failure patients. There was no difference in the mean length of stay between patients who survived to discharge and those who did not. However, the reason for which patients required intensive care appeared to affect average length of stay. This study highlights the importance and value of stratifying repatriated patients by specialty, in order to better understand differences in patient outcome.

Mechanical circulatory support in pneumonia-related cardiomyopathy: a challenge of cardiopulmonary pathophysiology during weaning

S Fizza Haider, Marco Morosin, Richard Fisher, James Doyle, Richard Trimlett, Francisca Caetano and Susanna Price

Royal Brompton and Harefield NHS Foundation Trust, UK, London, UK

Abstract

Introduction: Weaning from extracorporeal support can be challenging, in particular where the heart and lungs recover at different rates. Knowledge of the range of extracorporeal support configurations is mandatory for success where options include VV-, VAV-, VVA- and VA-ECMO. The choice of each depends upon not only intrinsic cardiac and/or lung function, but also tolerance to the different configurations of support.

Clinical case: A 47-year-old woman presented to her local hospital with a ten-day history of malaise, dyspnoea and cough. Despite initial intensive care treatment she deteriorated, meeting the criteria for VV-ECMO, and was retrieved to national centre on VV-ECMO for severe acute respiratory failure due to Group A Streptococcal pneumonia.

On arrival she was haemodynamically unstable. ECG demonstrated widespread ST elevation, and trans-thoracic echocardiography (TTE) revealed severely impaired biventricular function. Coronary angiography demonstrated unobstructed coronary arteries. As she was INTERMACS I she was converted to veno-arterial (VA)-ECMO together with an intra-aortic balloon pump (IABP). Trans-oesophageal echocardiography showed marked echo-contrast in the left ventricle with no aortic valve opening (Figure 1).

Subsequently cardiac function improved more rapidly than pulmonary, and during attempted weaning she developed Harlequin syndrome. In order to assess adequacy of RV recovery, she was temporarily converted to VAV-ECMO and the weaning study repeated. This demonstrated adequate RV function and she was ‘downgraded’ to VV-ECMO with dobutamine support. She remained on VV-ECMO for a further two weeks while her respiratory function and sepsis resolved. At this stage she was successfully decannulated and discharged back to her referring hospital.

Discussion: Indications to initiate ECMO include potentially reversible severe acute respiratory, cardiac or cardiopulmonary failure. The decision to initiate cardiac and/or respiratory ECMO is dependent on the patient’s history and nature of the disease process. In general, VV-ECMO is reserved for severe hypoxaemia refractory to traditional ventilatory support and VA-ECMO for primary cardiac failure, cardiogenic shock or cardiac arrest. Understanding the disease course and quantifying the degree of cardiac and/or respiratory failure also help predict ECMO duration, and influence the decision of ECMO configuration [1]. Where particularly challenging, additional ECMO configurations may be used (VVA- or VAV-ECMO).

There is little consensus on weaning strategies and timing of weaning from VA ECMO[3]. Although sometimes used, VAV-ECMO (returning oxygenated blood to the right atrium in order to minimise Harlequin syndrome) may result in acute right heart dilatation. It does, however, provide a mechanism to trial tolerance to weaning from full VA-ECMO support in the short-term. Where not successful, VVA-ECMO (with increased drainage from the right heart, reducing/stopping ejection of de-oxygenated blood) may be considered for (relatively) longer-term support whilst the heart and/or lungs recover. Expert echocardiography is mandatory to accompany haemodynamic and laboratory investigations during weaning trials in such patients[4].

Figure 1. Trans-oesophageal echocardiogram on VA-ECMO

Bi-plane transoesophageal echocardiogram (in systole) in a patient receiving VA-ECMO. Note both AV and MV are closed simultaneously, and the LV is opacified with spontaneous echo contrast. LA; left atrium, LV; left ventricle, RV; right ventricle, AV; aortic valve, MV; mitral valve

Heparin Induce thrombocytopenia post cardiac surgery compared to patients exposed to Extra Corporeal Membrane Oxygenation (ECMO)

Miguel Garcia¹, Tim Hayes², Giles Page² and Laura Head²

¹Withenshaw Hospital MFT foundation trust, Manchester, UK

²Withenshawe Hospital MFT foundation trust, Manchester, UK

Abstract

Introduction: Heparin induce thrombocytopenia (HIT) is a syndrome characterised by reduced platelets and clinical thrombosis after exposure to heparin. It is understood that thrombosis is mediated by antibodies against PF4. Antibodies against PF4 are produced in 20% to 50% of patients exposed to heparin, with an increased incidence associated to unfractionated heparin. Clinically it is characterised by an abrupt reduction of the platelet count after day 5 of the heparin exposure, not explained by other causes and with or without clinical venous or arterial thrombosis. Despite the prevalence of antibodies in the population exposed, HIT is only developed in 0.5 % to 1 % of patients. It is more common post heart surgery than in medical patients. Its treatment requires stopping heparin exposure and anticoagulating with alternative anticoagulants such as argatroban or danaparoid. The prevalence of HIT on ECMO and mechanical circulatory support is uncertain. The literature has reported a prevalence between 0.3 % to 10 %.

Methods: We performed a retrospective case note analysis of in all patients diagnosed with HIT from January 2017 to present. Our practice involves respiratory ECMO, acute heart failure including mechanical circulatory support and transplantation.

Results: In the period from January 2017 to present we have been able to provide data of 16 patients diagnosed with HIT: 3 patients with optical density below 1. ECMO CPR 1 patient, BIVAD (biventricular assist device) 1 patient, VV ECMO 3 patients, VA ECMO 2 patients, Lung transplant on VA (one converted to VV) =2 patients, post cardiac surgery =3 patients, Post thoracic surgery =1 patient, Heart failure (no mechanical support) =3 patients. We found a high incidence of DVT in 10 cases out of 16 positives. We also found an increase frequency of filter thrombosis. In this period of time we performed 49 VV ECMO and 58 VA ECMO a total of 107 ECMO patients. 8.2 % of patients on ECMO developed HIT in this period of time.

Discussion: We found the introduction of a bespoke pathway to make the diagnosis for HIT has probably improved our detection rate (figure 1). In our cohort of patients, the incidence for HIT in patients post general cardiac surgery is similar to the one reported in the literature, we found that the rate of diagnosis on ECMO is around 8 %. We think it is necessary to improve the laboratory base for diagnosis. We propose the introduction of well validated screening test available 24/7 and confirmatory tests in a national referral centre, bearing in mind how rare this diagnosis is.

Introduction of a pre-operative screening tool to identify and guide management of patients at high-risk of developing acute kidney injury (AKI) requiring dialysis post cardiac surgery

Peter Bamford¹, Omar Al-rawi², David Walker² and Mathew Shaw²

¹ST9 ICM/Anaesthesia, Royal Liverpool and Broadgreen University Hospital Trust., Liverpool, UK

²Liverpool heart and chest hospital, Liverpool, UK

Abstract

Acute kidney injury (AKI) is common after cardiac surgery with a reported incidence of up to 30%. Around 1–2% of all patients will subsequently require renal replacement therapy (RRT), with most remaining dialysis dependent after the perioperative period. It is widely accepted that AKI confers a significantly increased risk of morbidity and mortality. This is supported by the literature, which reports a mortality amongst patients developing AKI of 15–30% and 60–70% in those requiring RRT. (1)

At Liverpool Heart and Chest Hospital (LHCH) there was no mechanism to identify cardiac surgical patients at high risk of developing AKI requiring RRT. We therefore adopted the Cleland Scoring System (CSS) as a tool for risk stratifying patients into low, moderate or high-risk categories. (2)

The CSS was developed between 1993–2002 and originally incorporated 33,217 patients, which were divided between a test group and validation group. The study identified a number of risk factors that appear to infer an increased risk of developing AKI requiring RRT post-operatively. (2)

In light of this we retrospectively analysed 4 years of peri-operative data at LHCH to firstly evaluate the incidence of AKI requiring RRT, and secondly determine each patients Cleveland Score so we could compare our data against a model which has been widely tested, and shown to have the best discrimination amongst tested populations. (3)

We reviewed the data for 3,029 consecutive patients undergoing either: coronary artery bypass grafting (CABG), CABG plus valve or valve only procedures, and categorized the patients as either:

• Low Risk– Cleveland score 0–2

• Moderate Risk – Clevelend Score 3–5

• High Risk – Cleveland Score >5

Our results revealed that we clearly had a higher proportion of patients requiring dialysis in all three-risk groups:

1. Low – LHCH 2% vs. Cleveland 0.4% (n = 2305)

2. Moderate – LHCH 10.7% vs. Cleveland 1.8% (n = 685)

3. High – LHCH 51.3% vs. Cleveland 9.5–21.3% (n = 39)

We also looked at the mortality figures for LHCH patients and demonstrated that there was a huge difference between those dialysed vs. those not dialysed for each risk group.

1. Low – 28.9% in RRT patients vs. 0.3% in those not receiving RRT

2. Moderate – 11% vs. 1.6%

3. High – 25% vs. 10.5%

Our results indicate that although we appear to dialyse a larger proportion of patients at LHCH our outcomes appear to be better than the reported literature for patients requiring RRT post-operatively. However RRT remains associated with a higher morbidity and mortality. Our data supporting that a higher Cleveland score is associated with; a higher in-hospital mortality; an increased risk of dialysis and an increased length of critical care and perioperative stay (all P < 0.001).

There is no doubt that the identification of patients at high risk of developing AKI could lead to the introduction of interventions designed to minimize the risk of developing AKI and reduce the need for RRT. We have therefore introduced the CSS into the pre-operative assessment of all patients undergoing cardiac surgery at LHCH, and intend to roll out a renal bundle of care in the coming months.

Mucoactive agent use in acute respiratory distress syndrome – audit of practice in a Regional Intensive Care Unit

Ho Wong¹, Gordon Anderson² and Murali Shyamsundar¹

¹Centre for Experimental Medicine, Queen's University Belfast, Belfast, UK

²Regional Intensive Care Unit, Royal Victoria Hospital, Belfast, UK

Abstract

Introduction: Pharmacotherapies aimed at improving airway clearance have shown benefit in various chronic respiratory conditions such as cystic fibrosis. Mechanically ventilated patients with acute respiratory distress syndrome (ARDS) could potentially benefit from mucoactive therapy. Critical care trials on the use of mucoactive therapy are sparse and have not demonstrated benefit while their use in clinical practice is common. The potential downsides include harm from such therapy due to obstruction of heat and moisture exchangers by crystallization of salts, increased cost to care delivery and workload. The use of a dependable drug delivery system and drug dosage are also integral to maximising the potential benefit as the airway concentration of the mucoactive will determine the efficacy of the drug.

Objective: The objective of this audit is to ascertain mucoactive use and its effects on outcomes in patients with ARDS.

Study Design: Retrospective observational cohort study of mucoactive use in mechanically ventilated (MV) patients with ARDS.

Method: Electronic chart review of MV ARDS patients admitted to the Regional Intensive Care Unit, Royal Victoria Hospital over 12 months between March 2017 to Feb 2018 was conducted. Data on demographics, aetiology, mucoactive use, type of mucoactive and duration of MV were collected. Continuous data were tested for normality and compared using unpaired t test or Mann-Whitney U test as appropriate. Data are presented as mean (SD) or median (IQR).

Results: 129 patients were MV for ARDS during this period of which 88 (68%) of the patients were male and 76 (59%) had a non-pulmonary aetiology for ARDS. Mucoactive agents were administered to 56 (43%) of the patients. The mucoactive agents used were carbocysteine (70%), 3% hypertonic saline (4%) and DNAse (9%) and a combination in the rest. Jet nebulisers were used for the inhaled route. Different doses of carbocysteine was noted of which 45% of patients were administered 375 mg thrice daily. There was no difference in ARDS severity in patients prescribed ARDS and those who weren’t as determined by their p/f ratio at the time of starting mucoactive agents (16.7 (11.1–24.4) vs 18.2 (13–28), p = 0.22). Patients who were prescribed mucoactive agents had a significantly longer duration of MV when compared to patients who did not receive mucoactive agents (13 (7–20) vs 7 (3–14), p < 0.004).

Discussion: In this single centre retrospective study in ARDS patients, use of mucoactive agents is associated with a significantly increased duration of MV. The potential reasons for the increased could be secondary to alveolar dispersion of bacteria colonising the airways or a pulmonary toxic effect through an unexplored mechanism. The low doses used in systemically administered mucoactives agents might have contributed to the lack of efficacy. Finally, the type of nebuliser through its impact on drug delivery, would have influenced the efficacy of inhaled mucoactive agents. High quality trials of mucoactive agents in ARDS is essential with a specific focus on optimising drug delivery and precision medicine based on underlying therapy responsive trait in ARDS patients.

TUR(S) be or not TUR(S) be: Iatrogenic fluid overload following Holmium Laser Excision of Prostate (HoLEP)

Christopher Holt, Charlotte Smith and Larry Mulleague

Epsom and St Helier University Hospitals NHS Trust, London, UK

Abstract

Holmium Laser Excision of Prostate (HoLEP) is increasingly the method of choice for surgical management of benign prostatic hyperplasia (BPH) due to equivalent or superior surgical outcomes compared to Transurethral Resection of Prostate (TURP) and fewer surgical complications. Transurethral resection syndrome (TURS) is a well-recognised complication of TURP although relatively rare with a significant associated morbidity and mortality. HoLEP has a reduced rate of irrigation fluid absorption compared to TURP and has only two previously reported cases of massive absorption causing fluid overload, electrolyte disturbance and dilutional anaemia with a similar pathophysiology to TUR syndrome.

We present a further case report of a 62-year-old patient undergoing elective HoLEP who developed sudden signs of fluid overload with respiratory and cardiovascular compromise, requiring an emergency surgical tracheostomy perioperatively and admission to ICU for ongoing cardiovascular and respiratory support. Subsequent investigations revealed a normal anion gap hyperchloraemic acidosis and haemodilutional anaemia.

Massive fluid absorption and associated complications similar to TUR syndrome following HoLEP is seemingly rare with only two other reported cases. As the number of HoLEP procedures rises the incidence of this complication may be increasingly reported. We recommend severe cases are management in critical care with supportive therapy, diuresis and correction of electrolyte abnormalities.

Survey and monitoring of usage of chest radiography and chest ultrasonography in the cardiothoracic perioperative setting

Zlatka Belamarić, Giampaolo Martinelli, Gerry Van Rensburg, Greta Giuliano, Atif Husain and Suresh Sanapala

St Bartholomew's Hospital, Barts NHS Trust, London, UK

Abstract

Introduction: Respiratory complications remain a leading cause of postcardiac surgical morbidity and can prolong hospital stay and increase costs.

Conventional lung imaging in critically ill patients are bedside CXR and lung CT.

Lung US is also a tool for assessing lung status in critically ill patients and can be easily used and repeated at the bedside, allowing the effects of therapy to be monitored, as highlighted in the International evidence-based recommendations for point-of-care lung ultrasound published in 2012.

Objectives:

1) To survey if the LUS is utilised in the contexts where it is supported by the literature to be superior and more effective than CXR.

2) To compare if the different modes of investigation are performed in a timely and efficient manner.

Methods: This was a local service evaluation held at Barts Heart Centre limited to 4 adult Intensive Care Units.

Data collection was obtained from the computer system and patient health records which were held in the Intensive Care Unit.

Results: 299 cases of supine AP CXR or LUS were identified in the period from the October 2016 to the February 2017.

Overall CXR was used on 277 occasions (93%) whilst LUS was utilised on 22 (7%) occasions. Clinical questions triggering the request for the CXR/LUS were the presence of pneumothorax, pleural effusion, consolidation, pulmonary oedema and positioning of indwelling devices.

Of 108 cases where pneumothorax was suspected, CXR was utilised 106 times (98%) and LUS was utilised on 2 occasions (2%).

Of 43 cases where pleural effusion was suspected by the clinician, CXR was utilised 30 times (70%) and LUS was utilised on 13 occasions (30%)

Of 29 cases where consolidation was suspected CXR was utilised 27 times (93%) and LUS was utilised on 2 occasions (7%).

Of 27 cases where pulmonary oedema was suspected, CXR was utilised 25 times (93%) and LUS was utilised on 2 occasions (7%).

On 47 occasions CXR was utilised to ensure positioning of indwelling devices: NG tubes, chest drains, CVC catheters and ET tubes.

Only twice was LUS utilised to assist positioning, on both occasions regarding chest drain insertion.

The average time from the request of the CXR to the CXR having been performed was 2 hours 18 minutes. Meanwhile, LUS took 57 minutes on average from its request to its performance.

Conclusion: Our service evaluation showed that the most commonly performed investigation was CXR despite the evidence base strongly supporting the use of the LUS which could ameliorate effectiveness and rapidity of the intervention in perioperative care, resulting in the better patient care.

FONA isn't always the answer: a case report on difficult airway management in the Emergency Department from a massive thyroid mass

Shah Mizanur Rahman

Emergency Department, Wexham Park Hospital, Slough, UK

Abstract

This case report relates to the multi-disciplinary approach needed for the management of the critically unwell patient with a difficult airway, and how actions from the roadside through to the emergency department and intensive care unit contributed to a successful outcome in a patient with spontaneous airway compromise at home.

Prehospital events: An emergency call was placed by family for a 53 year old patient with sudden collapse at home. On arrival, the ambulance crew found a patient with a large neck swelling, and an extremely hypoxic patient with agonal gasping. Laryngoscopy was completely unsuccessful, and efforts to ventilate with a bag valve mask unfruitful. The paramedic team placed an igel, noting sats of 74% with a good pleth trace, and proceeded to ventilate with high flow oxygen and ‘scoop and run’ to the nearest ED.

Emergency Department: Upon receipt of the priority call, the ICU and Anaesthetic SpR were fast bleeped to resus, and the theatre ODP called for directly, being asked to bring additional difficult airway and fibreoptic equipment to the ED.

Preparations were made for surgical front of neck access using current accepted practice (scalpel, bougie, ET tube), as well as video laryngoscopy (glidescope with hyperangulated blade) and FOI set up. Upon arrival, the patient was in a agonal state, being unresponsive, with gasping breathing efforts, sats still in the high 70 s and otherwise tachycardic with a bounding radial pulse and normotensive.

Eventually, after much airway assessment and consideration of different options, an MLT was placed, with a reinforced COETT railroaded successfully, without respiratory or cardiac arrest. The extremely hypercapneoic patient’s physiology normalised with ventilation and was subsequently transferred to ICU. She had a <72 hour stay, was GCS 15 and transferred to tertiary care for her newly diagnosed thyroid mass.

Lessons learnt:

1. The value of maintaining oxygenation in the presence of failed ventilation through the use of supraglottic devices

2. Difficult airway management is a team effort

3. Difficult airway trolleys in remote site anaesthetic practice may require additional equipment such as MLT and reinforced endotracheal tubes.

Retrospective case analysis of extracorporeal membranous oxygenation resuscitation (E-CPR) for in-hospital & out-of-hospital refractory cardiac arrest

Giles Page, Laura Head, Alan Ashworth, Miguel Garcia and Julian Barker

Manchester University NHS Foundation Trust, Manchester, UK

Abstract

There is a substantial sub-group of patients who suffer cardiac arrest and do not respond to standard cardiopulmonary resuscitation (CPR), as such there has been a movement towards using veno-arterial extra-corporeal membranous oxygenation (VA-ECMO) as a salvage procedure to aid resuscitation in refractory cardiac arrest. Indeed the idea of ECMO being used in cardiac arrest was first mooted in 1966.¹Extra-corporeal membranous oxygenation resuscitation (E-CPR) has been demonstrated to provide survival rates of 50% when initiated within 30 minutes of cardiac arrest, 30% at 30–60 minutes and 18% after 60 minutes, with particularly favourable outcomes for in-hospital cardiac arrest.²Prior to several studies being published from 2013 onwards, out-of-hospital cardiac arrest (OOHCA) was thought to have variable outcome despite ECPR. The subsequent studies have since demonstrated that E-CPR when used in an experienced ECMO centre and in conjunction with mechanical chest compression is able to provide favourable neurological outcomes for patients in refractory out-of-hospital cardiac arrest.^3,4As such the Wythenshawe Hospital is an experienced ECMO centre and has been offering E-CPR since April 2016 on an ad-hoc basis and from May 2017 in a protocoled manner, as evidenced in diagram 1.

We performed a retrospective case note analysis of patients in refractory cardiac arrest in-hospital and out-of-hospital requiring E-CPR at Wythenshawe Hospital from when the service was first offered in April 2016 to the present time. The patients were started on E-CPR and then transferred to the coronary catheterisation laboratory for angiography and percutaneous intervention; from there the patients were moved to the cardiothoracic critical care unit. Statistical significance was determined using the student’s t-test.

In the period from April 2016 to present we have been able to provide E-CPR to 16 patients, 12 of whom had out-of-hospital cardiac arrest. As is evident in table 1 the overwhelming majority were male and the mean age was 47. The mean time from arrest to ECMO was 75.6 minutes across both OOHCA and in-hospital cardiac arrests, the time taken from the ECMO team reaching the patient to successful cannulation and starting E-CPR was 20 minutes. Four patients survived to hospital discharge, all with cerebral performance category (CPC) scores of 1.

Our overall survival rate of 25% to hospital discharge with CPC 1 is comparable to the published literature, particularly as survival is quoted at 18% for arrest to E-CPR time of greater than 60 minutes.² The median time to cannulation from ECMO team arrival is 22 minutes, this is similar to that achieved by the CHEER group.³ The main complications were injuries to the femoral vessels. The low number of patients recruited thus far has meant an inability to prove statistical differences between the surviving patients and the non-survivors, although 75% of the survivors are female.

We have been able to demonstrate that E-CPR is feasible and achieves outcomes similar to those published in the literature, we also anticipate that the time to E-CPR commencement and time from arrest will reduce with more patients recruited and wider awareness of the service in the pre-hospital community.

Therapeutic plasma exchange in VV ECMO patients with severe sepsis: Experience of a single centre

Giles Page, Cara Godfrey, Miguel Garcia and Tim Hayes

Manchester University NHS Foundation Trust, Manchester, UK

Abstract

The use of therapeutic plasma exchange (TPE) is well documented in the paediatric literature as an adjuvant therapy for either sepsis or thrombocytopaenia associated multi-organ failure, where it has been shown to improve multi-organ failure and outcomes.^1,2 It has been explored in the adult population where similar benefits have been seen in severe sepsis.³ It has also been demonstrated to be safe and effective when used in conjunction with extra-corporeal membranous oxygenation (ECMO) and renal-replacement therapy (RRT).^4,5 However, there is a paucity of evidence for the use of TPE in adults who require VV ECMO with similar presentations.

We performed a retrospective observational study of the use of TPE as an additional salvage therapy in conjunction with VV ECMO and renal replacement therapy in 9 adult patients admitted to the Wythenshawe Cardiothoracic Critical Care Unit between July 2017 and March 2018 with severe sepsis related multi-organ dysfunction. Statistical analysis was performed using a paired students t-test.

At baseline the patients had a median SOFA score of 15 (7–16) and a median MODS of 14 (12–16). The mean age of these patients was 42 (16–58); five of the patients were female. The patients underwent 3 (3–10) consecutive sessions of plasmapharesis, and this was commenced within 12 hours (median, 3-238 hrs) of admission to the critical care unit. The median total volume of plasma exchanged was 10800 ml and Octaplas was used for all of the plasma exchanges.

As highlighted in table 1, we were able to demonstrate a significant reduction in the level of organ dysfunction (MODS p0.045), although there was no significant difference in the SOFA score (p0.266). There were significant reductions in the quantity of vasoactive and inotropic medications (p0.03) used in the patients as well as lactate clearance (p0.02) following plasma exchange. Plasma exchange did cause a significant reduction in the haemoglobin (p0.004) and rise in the bilirubin (p0.005); these could be attributed to the volumes and products used in the plasmapharesis, as well as on-going resuscitative efforts, improved tissue perfusion and haemolysis with VV EMCO.

The all cause mortality at 30 days was 33% (3/9), and at 90 days four patients (45%) survived to hospital discharge (three of whom had SOFA scores of 16).

Our experience suggests that therapeutic plasma exchange allows for haemodynamic stability and modification of multi-organ failure in some patients with severe sepsis requiring VV ECMO. However, further investigation is warranted into the use of TPE in severely septic adults.

Audit to assess the impact of daily multidisciplinary antimicrobial review on antimicrobial stewardship with respect to antibiotic usage in critical care patients

Sarraa Almahdi and R Gopal

London North West Healthcare University NHS Trust, London, UK

Abstract

Antimicrobial resistance is a global problem; the world health organisation (WHO) has set a global action plan to address this issue. The department of health in the UK has introduced the Five Year Antimicrobial Resistance Strategy 2013–2018 to combat the increased resistance. The Commissioning of Quality and Innovation have set targets for reducing the impact of serious infection.

This audit aims to assess the impact of daily antimicrobial consultant and multidisciplinary team (MDT) review of antimicrobials on antibiotic usage in ICU with respect to defined daily doses (DDD). The objective of the audit is to collect data on the antibiotic DDD before and after the implementation of the MDT antibiotic review. The audit also aims to ascertain the standardised mortality ratio and the antibiotic expenditure before and after implementation of the respective intervention. The audit standards are to have a 20% reduction in antibiotic usage.

The intervention was implemented in an 11 bed general intensive care unit at a 722 bed hospital in the UK. ICU patients were reviewed daily on weekdays by a multidisciplinary team at a meeting room consisting of a microbiology consultant, ICU consultant, trainee doctor, pharmacist, nurse, dietician, physiotherapist and a speech and language therapist. The ICU consultant presented the patient’s clinical update to the MDT, the antimicrobials were reviewed by the microbiology consultant and ICU pharmacist with the MDT. Agreed recommendations were documented and implemented on the patient’s drug chart and medical notes. Data was collected on antibiotic usage 1 year before implementation and 1 year after implementation of the intervention.

The results showed that the total DDD was 8624.9 and 6580 in the control and intervention group respectively. The Wilcoxon signed The Z-value is −2.1734, if the Z-value is less than −1.96 we can reject the null hypothesis. The average DDD for the control group is 128.73 95% CI [77.6656, 179.7944] the average DDD for the intervention group is 98.12 95% CI [56.3051, 139.9349], p = 0.03. The DDD per 1000 admissions for the control group is 14209 and for the intervention group it is 7891. The intervention group has 44% less antibiotic usage per admission. The DDD per 100 patient bed days for the control group is 294 and for the intervention group is 119. This is a 60% reduction in antibiotic usage per patient bed day.

The standardized mortality ratio for the control group was 1.01 and for the intervention group was 1.05. The expenditure on antibiotics was reduced from £1889 to £1110 per 100 patient bed day and subsequently from £91,290 to £67310 per 1000 patient admission.

Table to show antibiotic usage in ICU as defined daily doses (DDD), antibiotic expenditure and the standardised mortality ratio.

	Control (2016)	Intervention (2017)
Total defined daily doses (DDD)	8624.9	6580.9
Average DDD	128.73 95% CI [77.67, 179,79]	98.12 95% CI [56.30, 139.93]
Number of patient admissions	607	834
Number of patient bed days	2934	5551
DDD per 1000 admissions	14209	7891
DDD per 100 patient bed days	294	119
Standardised mortality ratio	1.01	1.05
Average APPACHE score	19.6	18.6
Actual annual expenditure on antibiotics reviewed	£55,413	£56137
Expenditure per 100 patient bed day	£1889	£1011
Expenditure per 1000 patient admission	£91,290	£67,310

In conclusion the daily antimicrobial consultant and multidisciplinary ICU team review of antimicrobials reduced antibiotic usage in critical care patients and reduced antibiotic expenditure with no change in mortality.

Risk Assessing and Managing Aggression and Violence in Mental Health Patients in a Tertiary Intensive Care Unit

A Wood, Cheryl Achary, J Pennington, L Barrass, A Zurecki and M Healy

Adult Critical Care Unit, Royal London Hospital, London, UK

Abstract

Violent and aggressive episodes are common in the acute medical setting, particularly in areas caring for elderly patients¹ and those with cognitive or mental health disorders.² Incidents of physical assault reported to security staff at our institution have risen by 60% between 2016 and 2017.³

In response to a critical incident involving a patient suffering from an acute psychotic episode, our tertiary critical care unit has initiated processes to equip staff in the management of violence and aggression associated with similar patients admitted to the intensive care. These include a series of practical training sessions with security staff on safe restrictive interventions, new guidelines to both risk assess and manage the spectrum of aggression in mental health patients and plans to integrate these new guidelines into the staff induction program.

Our guidelines focus on five core steps, to be exercised in parallel with the clinical critical care needed by the mentally unwell patient:

Step One. When a patient with known or suspected acute mental health illness is admitted, staff are advised to consider their potential to be violent or aggressive. Risk factors cited as predictive of violent behaviour were gathered according to current evidence base.²

Step two involves advising on a collateral history.

Step three encourages early involvement of the psychiatry team to conduct a mental health assessment and provide advice on managing both the acute mental health pathology and potential for aggression.

Step four utilises the Broset^4,5 checklist and advises on preventative strategies in patients identified as moderate or high risk. This guideline details non pharmacological measures including de-escalation strategies and indications for restrictive interventions, with a checklist to highlight patient safety measures when using manual restraint manoeuvres. This step also includes a rapid tranquilisation pathway detailing a structured escalation plan when pharmacological measures are urgently needed.

Step five addresses crisis situations where despite measures violent behaviour escalates and it becomes necessary to conduct an emergency rapid sequence induction, possibly in a remote area. The standard operating procedure details two main phases; initial dosing to gain control of the situation with/without i.v. access in preparation for an RSI, then advice on RSI and post RSI elements to address. We highlight the need for cooperation between psychiatric and medical teams to coordinate an individualised care plan and avoid prolonged intubation periods.

Following such an event we highlight the importance of attentive psychosocial support for those involved.

These guidelines and training sessions were organised to address the emerging needs of our patient population, but also to empower our staff in moments of crises. By focusing on risk assessment and early preventative measures we hope to avoid violent episodes in our intensive care. Where they do still occur, we hope these guidelines provide a structured, safe and evidence based method of keeping both our patients and staff safe.

Acknowledgements: Medical, nursing, security and pharmacy staff at the RLH ACCU. Thank you also to Bartshealth psychiatrists who advised and ratified drug selection and dosing.

An immersive augmented-reality environment of the intensive care bed space as a tool to orientate patients and families

Henry Munby

Gloucestershire Hospitals NHS Foundation Trust, Gloucestershire, UK

Abstract

Despite recent moves to improve user experience, the intensive care environment remains frightening and alien for a large proportion of patients, families and carers, and admissions can have adverse psychological impacts.

For patients this is known to contribute to delirium and the post-intensive care syndrome (PICS) with its numerous psychological sequelae, including post-traumatic stress disorder.1 Patients with extended stays or receiving mechanical ventilation are particularly at risk. Significant others often do not know what to expect from an ICU admission and may be overwhelmed by the unfamiliar devices, staff and noises. This is often a distressing life event that can have significant psychological morbidity for family members. 2

The job of orientating patients and families to the environment often falls to the nurse at the bedside, with intermittent input from rounding clinicians. Technical tasks inevitably compete with time to offer thorough explanation and reassurance. There is a lack of high-quality, interactive sources of information to help orientate families to the workings of a critical care unit, and we are proposing an educational role for novel immersive virtual reality technology

We are developing an immersive augmented reality (AR) video and audio tour of our local ICU, taken with a 360-degree camera. Embedded interactive educational content explains the devices, people and processes common to critical care. The video will be accessible via the Virti application which is intuitive to navigate and is supported by Android and iOS smartphone and tablet operating systems. This can be viewed in immersive ‘VR mode’ through a low-cost headset adaptable to a user’s own device or navigated by scrolling a touchscreen device in a two-dimensional mode. Virti is established as a platform for medical education and is currently being used for immersive learning of resuscitation, procedural and communication skills.3

The application is free for users to download, and access codes to the ICU tour modules can be distributed to patients and families to access in a number of settings. The tour could be included in pre-operative assessment for patients undergoing high-risk surgery and therefore planned to recover in ICU and potentially experience a prolonged stay and higher rates of delirium and PICS. Families and carers of admitted patients could be given access to aid their understanding and orientation to the unit on visiting. And finally, patients themselves could use a device in VR-mode to take the tour while lying in bed, to better comprehend their environment. Other groups have piloted the use of immersive AR technology in supine ICU patients and proven its safety for a cognitive rehabilitation programme. 4

We will be studying the impact of the intervention with a qualitative survey of patient, family and carer satisfaction, as part of a local quality improvement project in the department of critical care. We envision this novel educational tool becoming a means to inform patients and families and mitigate some of the stress inherent in a critical care visit.

The Burden Of Maternal Critical Care In 365 Days At The University Of Portharcourt Teaching Hospital Nigeria

Job Otokwala

Department of anaesthesiology & intensive care university of portharcourt teaching hospital, port Harcourt, Nigeria

Abstract

Background: Developing countries such as Nigeria constitute about 50% of maternal deaths; Women with life threatening illnesses that require organ support either during pregnancy or in the puerperium constitute a significant risk group for maternal morbidity and mortality. The scourge of maternal deaths in Nigeria is about 600/100,000 pregnant women and such an alarming figure provokes responses to curb the increasing trend.

The epidemiology of maternal deaths revolve around the commonly known causes such as sepsis, haemorrhages from ruptured uterus, abnormal placentation, placental abruption, uterine atony, eclampsia, severe pre-eclampsia, pulmonary embolism and trauma. Preventive measures have been adopted at various levels to reduce the increasing burden of maternal deaths. Most of these measures are in the areas of education, training of traditional birth attendants and antenatal care.

Although all these efforts are laudable, a small percentage of these patients still follow the common pathway of critical illness defined by sudden deterioration of organ-system failures and required efficient management to reduce maternal deaths.

Aim: This study looked at the percentage of women that were admitted into the intensive care unit of the university teaching hospital in Southern part of Nigeria and assessed the epidemiology of the risk factors for maternal deaths within a limited period of time, 365days.

Methodology: Ethical clearance was obtained from the university teaching hospital’s ethical committee and all data as documented in the record books of parturients that were admitted either from outside the hospital, referred from the delivery wards or postnatal wards and from the obstetric emergency clinic to the intensive care unit and that died between july 2017-july 2018 were collated and analysed.

Results: Forty-six (46) parturients were admitted between 01 July 2017 and 31 July 2018 in the mixed ICU. The mean age of the parturients was 33 ± 1.3 years. The mean BMI was 32 kg/m2. The mean time lag at presentation was 72 ± 10.4hours. seventeen(17) parturients died in 356days from the following causes:

Total number of parturients admitted 46.

Maternal deaths within one year(july 2017-july 2018): 17(36.9%)

CAUSES OF MATERNAL DEATHS	NUMBER(%)
Postpartum haemorrhage	8(47%)
Pulmonary Embolism	4(23.5%)
Eclampsia	2(11.8%)
Sepsis	2(11.8%)
Aspiration pneumonitis	1(5.9)

ICU PROFILE

Duration of stay in icu(DAYS)	INVASIVE VENTILATION	NIV
3 ± 0.1	5(29%)	2(11.8%)

Conclusion: Haemorrhage still remain a major risk factor for maternal deaths in developing countries. The contributions of delays at presentation to prompt definitive interventions such as early hysterectomies when indicated to the absence of blood products are challenges. Pulmonary embolism could easily be missed as a result of complex pathologies at presentation. The prevalence of anaemia, liberal use of preloading fluids during caesarean deliveries, malaria infestations and diagnostic dilemma make pulmonary embolism an easily missed entity.

The complimentary utilization of preventative measures and strengthening of health systems especially the intensive care strategies will help circumvent the growing burden of maternal critical care in developing countries.

Pecha Kucha Abstract

Lee Angeline

ACCS Anaesthetics CT3 Base Hospital: Wexham Park Hospital

Abstract

He came to me on a paramedic stretcher, on one of those days that the emergency department was boiling over.

The emergency department doctors had called to say they were sorry to trouble us, but he was too far gone, and that they’d tried to speak to his family to tell them that there was nothing that could be done, but they still wanted us to see him.

COPD is a terrible disease. I thought this as I laid eyes on him. He was in tripod position, what was left of his intercostal muscles being sucked in to outline bone as he tried to fill his lungs with the stale air of the resus bay. He could barely keep his eyes open. His blood gas results were in my hand and they told the tale of months of worsening airway disease and an exacerbation that had brought him to his knees in his front room – making a proud old man who had survived two world wars finally concede to let his family call the ambulance.

His wife and son had seen me coming and stood between us as I approached. I felt their helplessness. I felt their pain. I felt their fierce desire to protect him. I felt their mistrust as I introduced myself as an intensive care doctor. They had met four doctors since he had been brought in fifteen minutes ago and every single one had his future written on their faces like an open book. His son said as much when I mentioned resuscitation. ‘It’s like you think he’s already dead! You haven’t even tried and you’re letting him go!’

I hardened my heart.

In my 10 years of medical training, I have seen thousands of PowerPoint slides on communication skills, participated in hundreds of conversations, real or simulated, to help me have these conversations with patients and their relatives. By the time I met him and his family, I had already had five similar conversations with other people that day.

I had faith in my communication skills, but as we talked, I realised that I had also been lucky. In almost all the times I’d had these conversations previously, the patients and their families were already on board with my agenda before we’d even started. This time, his family and I were in different boats going in different directions. The room was cold and reeked of fear, worry, mistrust and separation.

But he was still alone on the stretcher.

I stopped and went back to the last point of common ground. ‘You are right. I haven’t even tried to make him better yet. I’m sorry. Let’s go to him. Please stay with him while we give him more treatment. We’ll talk again once he’s had some time to respond.’

His wife had been mostly quiet during the conversation, but now she stood and took my hand, and together we went back to his bedside.

Pecha Kucha Presentation – A day that changed my practice

CR MacEwen, SpR Intensive Care Medicine, John Radcliffe Hospital

The author uses a selection of images and poetic forms to share thoughts and insights on grief, triggered by watching an elderly man bid farewell to his wife of sixty years in the middle of a busy intensive care unit, and extending to the universal experience of loss.

“Hey, that’s no way to say goodbye”

Excerpt:

Sleep my baby, close your eyes,

Sandman’s calling from the skies

So many nights, pursuit disbanded

Gashed with glimpses, empty-handed,

Sightings growing time-worn tarnished

Mere cadaver, mere imposter,

How many restless, limbless nights has

Searched the wraith whose soul is lost her?

And there you are, the hunt starts over!

Sprinting, gleeful, far-doomed creature!

The bugle sounds, the lens slips down

As viscous torsion chokes endeavour,

Wading, wailing, predetermined

The song of sorrow loops forever

But this full moon, your eyes meet mine,

Smiling, shining, soft and lined,

I clutch at you, my grief confounded

Coarse warmth real between my fingers

I don’t believe, until I breathe,

Deep in skin folds, her soul lingers

Sense that can’t be coaxed or conjured

Long I’d mourned our realms as sundered,

But yet in limbic twists persist

Potential changed not by goodbyes,

One breath, I knew, one breath before,

The running sand filled up our eyes.

Name: Emma Houston

Job Title: ST5 Anaesthetics

Base Hospital: Liverpool Heart and Chest Hospital

In late summer 2013, I was half way through foundation training and about to land in the High Dependency Unit (HDU) of a large teaching hospital. I had survived my first year as a doctor relatively unscathed, and was riding the crest of that wave. Little did I know that I was hurtling head first into a rotation, and one patient in particular, that would change not only the course of my career, but my outlook on medicine.

Starting on the HDU was a challenge. The first handover alone felt like it was being conducted in a foreign language. What exactly is a Swedish nose? And who is Passy-Muir? But I quickly started to feel at home amidst the beeping and infusion pumps and enjoyed the challenge of managing some of the most acutely unwell patients in the hospital.

Critical care seemed to highlight the dichotomy of highs and heartbreaking lows that are inherent in medical practice. This was particularly true for J, a young man, who was admitted with a large pharyngeal tumour obstructing his airway. J had an emergency tracheostomy and was admitted to the HDU while awaiting a MDT treatment plan.

J had a wicked sense of humour. He was always ready to joke and share a smile. We laughed together when a delirious patient told me how pleased she was that I’d brought my boyfriend to work, as she gestured at the nearby drip stand, and when an elderly patient barricaded himself in a linen cupboard, declaring that it was his office. I always looked forward to seeing him.

Unfortunately, one evening J suffered a torrential bleed from his tracheostomy site. I was the first to attend. J looked terrified. I requested senior support and witnessed a coordinated, efficient and effective response from the critical care consultant. I was struck by how calmly, swiftly and expertly she handled the situation. I reflected that this was definitely practice to aspire to. J was rapidly transferred to theatre. Shortly after the MDT outcome was decided. J was informed that curative treatment was not an option. I’ll never forget the conversation we had when he asked, “what should I do now?”.

Critical care is a place of extremes, of brilliant highs and brutal lows. Through this a team of caring, compassionate and extremely competent healthcare professionals work together to try and achieve sometimes, seemingly impossible, feats for their patients. They advocate for patients, support juniors in clinically challenging circumstances and are honest with families about the reality of the circumstances they are facing. This rotation opened my eyes to the great possibilities of critical care. Restoring the health of seriously unwell patients, and providing dignity and comfort when that is not possible. This experience changed my outlook and paved the way for my entry into that field of practice, while inspiring me to follow the examples of the expert clinicians that I encountered when working there as a junior.

Abstract

Name: Kate Tantam

Job Title: Specialist Research and Clinical Education Sister

Base Hospital: University Hospital Plymouth NHS Trust

Abstract (500 words maximum)

One day I walked into an office and read a letter from a 35 year old female patient, 6 months after her admission to critical care.

“When I came home I was barely able to walk…physically I had a long rehabilitation period ahead of me (with no support given) but I was also deeply traumatised by what had happened to me in ICU. I had flashbacks and couldn't make sense of what happened (I had very confused memories from ICU, some of which were from delirium and were unreal, but I didn't know which were true and which were false). I was also 12 weeks pregnant before my critical illness, so I was profoundly sad at the loss of that pregnancy. It felt like my life had been torn apart. Three months after my illness, my husband had a break down over what he had gone through while I was ill”.

This story broke me, I cried and it opened my eyes. We didn’t have a formal follow up service, and we had no one to whom I could get this lady any help. I vowed on that day that it had to change.

We have set up a follow up team, a rehabilitation team and we have a business case in for an ICU outreach rehabilitation service.

We realised that patient stories are key to unlocking engagement in the clinical setting. We have used patient stories as part of “Rehab Revolution” in the South West. We have embedded a patient narrative in every piece of teaching, every business case, every small innovation, and in everyday language when we talk about care.

Cultural change in a busy MDT team can be hard to start, maintain and sustain and one of the key components of that cultural change has been our #rehablegend campaign. Every day we nominate a #rehablegend and share their genius online.

Nominations are based on excellence in clinical settings. Examples include going the extra mile, inspiring others, caring for colleagues, artistic skills in bedside care, innovating care, supporting loved ones and anything that is deemed by anyone to have improved the experience of being in the NHS. This campaign recognises brilliance and the impact of the little things. We work on the very simple concept that everyone needs to take charge of supporting recovery after critical illness, irrespective of role or designation.

We have seen a profound change in the culture of our MDT team by initiating this very small and simple intervention. Public acknowledgement, sharing of patient stories, sharing of local achievements and the simple act of saying thank you has been huge.

We want to present this work because when we tweet we see that other departments are starting to join our team of #rehablegends. We would like to see if we can extend our rehab revolution further and share the impact of sharing stories. @plymouthicurehab

Name: Simon Gates

Job title: Senior Paediatric Physiotherapist

Base Hospital: Nottingham Children’s Hospital (Nottingham University Hospitals NHS Trust)

A day that changed my practice – Abstract

Andrew* is a two year old boy. Following a sudden collapse at play group he was admitted to hospital and diagnosed with pulmonary langerhans cell histiocytosis (PLCH), a condition that results in severe cystic degradation of lung tissue and significant respiratory failure. The incidence in children is estimated to be less than 1 in 10 million.

Despite initially making good progress, Andrew was admitted to the paediatric intensive care unit following a seizure and required intubation and ventilation. The consequences of this were significant. The level of ventilation required resulted in multiple pneumothoraces and subsequent chest drains, a CT scan showed areas of pneumonia in addition to the cystic changes. Over the following days Andrew became persistently hypoxic despite high FiO2 and maximal ventilation, he was discussed with the local ECMO centre but was deemed to be not suitable due to his underlying condition.

At the time, our unit was trialing an electrical impedance tomography (EIT) monitor. This technology had not been widely used in children but we were able to use ECG electrodes to attach the monitor and begin to use it to understand how our interventions were impacting on Andrew’s ventilation. The breath to breath feedback allowed us to optimise his ventilation and within 12 hours his oxygenation began to slowly improve. Over the following days Andrew continued to improve and was eventually extubated. As a team we were excited how EIT had influenced our decision making and were delighted that we had beaten the odds in getting him through his deterioration.

But this wasn’t the day that changed my practice. That was five weeks later when one of the PICU consultants mentioned that Andrew’s mother had been writing a blog throughout her son’s illness. I went home that evening and decided to read.

The sheer emotion in the writing jumped out at me. The confusion at what was happening and the uncertainty of the situation. Having this insight into how a parent was feeling while their child was critically unwell left me questioning how we interact with parents when our minds are focused on the clinical situation we are dealing with. Since then I have always made time to ensure families understand what is happening at the time of the event, not just when they receive updates. It has also convinced me of the value of families making a record or diary of events that happen while their child is in PICU. As clinicians it can be very easy to lose sight of the emotional cost that can arise from our clinical success.

*Name changed to maintain anonymity

Consent has been gained from the patient’s family to share the information contained in this abstract.

Sonic? Boom!

S. A. Hayward

Blackpool Teaching Hospitals NHS Foundation Trust

I must admit, I may have bitten off more than I can chew but I am a completely different professional to the one I was a few years ago. In 2013 I was perfectly comfortable with my assessments skills as an ICU physiotherapist. I believed I was doing the best I could for my patients using observation, auscultation and chest radiographs to deliver the care they needed. That was until a “happy” Scottish intensivist introduced me to a technique that changed my practice and the direction of my professional career.

“To be honest, I don’t bother with auscultation and I can’t rely on chest, x-rays!” he said. What! Two of my main assessment skills had just been annihilated. “What do you use?” I asked tentatively. “Lung ultrasound, do you want me to show you?” That was it, the point at which my outlook on my profession shifted. Why had I never heard of this before and why aren’t all critical care physiotherapists using it? I discussed learning lung ultrasound with my senior physiotherapist and received a categorical; “No!” “It’s too specialist!” That just wasn’t good enough, this was too much of an important discovery to let it pass my profession by.

Since that day I have made it my mission to learn lung ultrasound and spread the word about this technique to as many respiratory and critical care physiotherapists as I possibly can. It took a while but with the help of my intensivist colleague I eventually learnt how to independently perform lung ultrasound and became the first physiotherapist within the UK to gain CUSIC accreditation. Now I felt confident I needed to spread the word further, so I sent my first Tweet about lung ultrasound in February 2016 and within an hour and a half I had the Association of Chartered Physiotherapists in Respiratory Care wanting to know more.

To date I have taught an introductory lung ultrasound course to over 100 physiotherapists across the country. I recently completed another essential milestone by becoming the first physiotherapy CUSIC mentor, which allows me to train and accredit from within my own profession. As a result of my incessant pestering of the CUSIC committee I was invited to join as the AHP lead which I accepted without a second thought. Along this journey I discovered the POCUS community, and they were as passionate as I am. Now I need to spread the word even further, showcase the technique, teach the way and prove anyone can learn lung ultrasound.

I want to be the best physiotherapist I can be for my patients, and I believe lung ultrasound can help me, and others, do that. Lung ultrasound has given me the autonomy to assess and treat my patients more effectively and I now plan to prove that by being the lead co-ordinator on research projects nationally. In the end I didn’t need a day. I can remember the exact moment when my practice changed, stood next to bed 2, on our general ICU.

The evaluation of a biomarker-based exclusion of ventilator-associated pneumonia to improve antibiotic stewardship. A multi-centre validation study and randomised controlled trial

TP Hellyer

Institute of Cellular Medicine, Newcastle University, Newcastle-Upon-Tyne

Abstract

Ventilator-associated pneumonia (VAP) is a common intensive care unit (ICU) nosocomial infection. The diagnosis of VAP is challenging and there is considerable variation in approaches in clinical practice.¹ VAP is confirmed in 30–50% of clinically suspected VAP, which implies that it is over-treated. Antimicrobial resistance is a global problem and antibiotic stewardship could be improved through biomarker-based diagnostics. In a single centre study, bronchoalveolar lavage (BAL) fluid biomarkers demonstrated diagnostic potential to exclude VAP.²

This body of work aimed to validate BAL fluid biomarkers to exclude VAP in a multi-centre setting and then determine whether a rapid, biomarker-based exclusion of VAP could reduce antibiotic use in patients with suspected VAP in a randomised controlled trial (RCT).

A multi-centre observational study was conducted in 12 UK ICUs.³ Patients with suspected VAP underwent a standardised bronchoscopy and BAL. VAP was confirmed by growth of a potential pathogen in BAL fluid at >10⁴ colony forming units per ml (cfu/ml). Interleukin-1 beta (IL-1β), IL-8, matrix metalloproteinase-8 (MMP-8), MMP-9 and human neutrophil elastase (HNE) were measured in BAL fluid by cytometric bead array. Patients were dichotomised into ‘VAP’ and ‘non-VAP’ groups and receiver operating characteristic (ROC) curves were constructed for individual and combinations of biomarkers. Cut-points were derived to achieve a minimum negative predictive value (NPV) of 0.95 to rule-out VAP, with the optimal combination of biomarkers going forward to the RCT phase of the project.

The optimal biomarker test was assessed in a multi-centre RCT, conducted in 24 UK ICUs.⁴ Patients with suspected VAP were randomised 1:1 to a ‘biomarker-guided recommendation on antibiotics’ (intervention arm) or ‘routine use of antibiotics’. Following randomisation and bronchoscopy, BAL fluid from patients in the intervention arm underwent testing for biomarkers and results aimed to be returned to clinical teams within 6 hours. Semi-quantitative culture was carried out on all BAL samples. When biomarkers excluded VAP, clinicians were advised to consider early discontinuation of antibiotics. The primary outcome measure was change in antibiotic duration, measured by the frequency distribution of antibiotic-free days (AFD), in the 7 days following BAL. Secondary outcome measures included AFD at 14 and 28 days; antibiotic days at day 7, 14 and 28; ventilator-free days at 28 days; 28-day and ICU mortality; sequential organ failure assessment (SOFA) score at day 3, 7 and 14; and duration of ICU and hospital-stay. Planned subgroup analyses included a per-protocol analysis, clinician assessment of likelihood of VAP (low, medium or high) and admission category (medical, surgical, trauma/head injury). The effect of deaths within the primary outcome period was determine by two sensitivity analyses, using discrete-time Cox proportional hazard model and repeating primary analysis resetting AFD to zero for patients that died in the 7 day period.

The validation study included 150 patients with paired BAL culture and biomarker results. VAP was confirmed in 53 (35%) patients. All biomarkers were significantly higher in the VAP group and the area under the ROC curve was 0.81 for IL-1β, 0.74 for IL-8, 0.76 for MMP-8, 0.79 for MMP-9 and 0.78 for HNE. The optimum cut-point for IL-1β (minimum NPV of 0.95) was 17 pg/ml and had a sensitivity and specificity of 96.2% and 43.3% respectively. A combination of IL-1β and IL-8 had a sensitivity and specificity of 100% and 44.3% respectively and was determined the optimum combination of biomarkers to use in the RCT.

Two hundred and ten patients with suspected VAP were randomised in the RCT. The VAP rate was 34% and clinician pre-test suspicion of VAP was predominantly ‘high’ (57.8% and 45.3% in the intervention and control arms respectively). At baseline, 81% of patients were receiving antibiotics. The NPV of the biomarker test was 0.929 (95% CI 0.643–0.990). The mean time from sample collection to reporting to clinical teams was 8.2 hours (SD 2.5). There was no difference in the frequency distribution of AFD at 7 days between patients having a biomarker-guided approach to antibiotics and those in the control arm (Chi-squared = 5.70, p = 0.575). The median antibiotic days at day 7 was 6 (IQR 4–7) in both trial arms (HR 0.838 [95% CI 0.626–1.122], p = 0.2360). There remained no difference in AFD between the two groups at day 14 (HR 1.126 [95% CI 0.826–1.536]) and day 28 (HR 1.008 [95% CI 0.725–1.402]). Sensitivity analyses did not reveal any effects of mortality in the 7 day period or any centre effects. There was no mortality difference at 28 days (OR 1.519 [95%CI 0.782–2.982]). There were no differences between trial arms in any secondary outcome measure. In the subgroup analysis of clinician pre-test likelihood of VAP, there was more antibiotic use in the ‘high’ likelihood group (‘high’ vs ‘low’: HR 2.358 [95% CI 1.205–4.906]; p = 0.011).

This work has validated the diagnostic value of BAL fluid biomarkers to exclude VAP. Furthermore the diagnostic performance of the test remained largely consistent in the RCT. Despite the strength of this diagnostic test to rule-out VAP, when applied to an antibiotic discontinuation intervention in a RCT, antibiotic use was not reduced. ICU is a clinical environment of high antibiotic use, with often diagnostic uncertainty and high actual or perceived clinical risk. This biomarker-based intervention was unable to alter prescribing practices. The evaluation of other molecular-based diagnostics for VAP offers additional opportunities for antibiotic stewardship.⁵ Future applications of these tools to reduce antibiotic use should consider embedding in a wider antibiotic stewardship programme.

These studies received ethics committee approval. Validation study: 11/NE/0242, 11/SS/0089; RCT: 13/LO/0651, 13/SS/0074.

This work was supported by the Health Innovation Challenge Fund (HICF-0510-078; 094949/Z/10/X), a parallel funding partnership between the Department of Health and Wellcome Trust. The views expressed are those of the author and not necessarily those of the Department of Health or Wellcome Trust.

Becton Dickinson Biosciences provided reagents, equipment, technical support and training for the flow cytometric assay.

The effect of circulating humoral factors on mitochondria in sepsis-induced acute kidney injury

Sean J Pollen

Bloomsbury Institute of Intensive Care Medicine, UCL, Gower Street, London, WC1E 6BT

Abstract

Sepsis is a major clinical problem with a persistently high mortality and morbidity. Concomitant acute kidney injury (AKI) occurs in half of all such patients, in whom the mortality risk doubles. Whether AKI is cause or epiphenomenon for this increase in mortality is unclear. Nevertheless, novel therapies currently available for treatment of both sepsis and AKI are limited, with most treatment focussed on resuscitation, stabilisation and organ support therapy until native organ function recovers.

The pathophysiology underlying sepsis remains to be elucidated. The observation that failed organs appear largely normal histologically suggests a functional rather than structural abnormality. Mitochondrial modulation of cellular bioenergetics has been posited as a mechanism to explain the apparent paradox between the clinical and biochemical presentation of septic AKI, despite minimal cell death, maintenance of tissue oxygenation and eventual recovery of organ function.¹ The postulate is that bioenergetic failure leads to a metabolic shutdown that may be protective under conditions of prolonged and extreme stress, akin to a hibernation response. A temporary down-regulation of metabolic activity in sepsis presents clinically as organ failure but may in fact aid organ recovery and thus the chances of long-term survival. This may act through preventing an overwhelming mismatch between oxygen supply and demand that would otherwise lead to irreversible cell death and an inability to regenerate. Additionally, in vitro studies show that exposing naïve tissue and cells to septic serum causes mitochondrial dysfunction. These studies exclude any impact from the abnormal circulation also observed in sepsis and suggest circulating factors (‘evil humours’) may play a fundamental signalling role in triggering cellular bioenergetic and metabolic shutdown. As mitochondria are also the major source of reactive oxygen species (ROS) production, the down-regulation of electron transport chain activity, seen as a decrease in mitochondrial membrane potential (MMP), may offer protection through decreasing potentially overwhelming ROS.²

To probe the phenomenon of septic serum induced mitochondrial dysfunction, I utilised a well-established long-term, fluid-resuscitated in vivo model of sepsis in my host lab³ and then refined an ex vivo technique whereby thin slices of kidney tissue, taken from healthy animals, were exposed to serum harvested from septic or control rats after 24 hours.⁴ In this ex vivo model real-time changes in mitochondrial function could be visualized using fluorescent dyes or endogenous fluorescence under confocal microscopy. MMP, global and mitochondria specific ROS, mitochondrial redox state, and cell viability were probed using tetramethylrhodamine methyl ester (TMRM), dihydroethidium (HEt), mitochondrial superoxide indicator (mitoSOX), constitutive NADH, and calcein acetoxymethyl ester (calcein AM), respectively. Thus, over a 90 minute time course, the impact of humoral mediators present in septic serum on mitochondrial function could be tracked. Pooled serum from sham-operated animals acted as controls.

Following characterisation, proposed modulators of the changes observed were included in experiments to potentially prevent or mitigate the dysfunction caused by exposure to septic serum. Preliminary studies were also performed to determine which components within the septic serum were predominantly responsible for the changes in mitochondrial function observed. This was achieved by separation of the sera into heavy and light fractions, split at 50 kDa which were then recombined with their respective counterpart from the alternative experimental condition (control or septic) to create a chimeric serum.

Septic serum increased mitochondrial ROS generation, decreased MMP, and shifted NADH redox state towards oxidation (Table 1). Taken together these results suggest increased mitochondrial uncoupling (i.e. away from ATP production and directed towards heat generation). A fall in MMP could be due to either increased uncoupling or decreased ETC activity, however a decrease in ETC activity would be expected to increase the NADH signal whereas uncoupling would decrease it, as seen here. To investigate this further, I quantified protein levels of the main uncoupling protein (UCP) present in renal tissue, UCP2, in kidney slices following exposure to sham or septic serum for 90 minutes. While I saw no change at this early timepoint following serum exposure, I did find marked UCP2 upregulation in whole kidney tissue taken from septic animals at 24 h. UCP2 protein was increased 1.65 ± 0.52 fold (mean ± SD) compared to levels in sham tissue (p < 0.05). To investigate the activity levels of UCP2, the healthy kidney slices were co-incubated with septic serum and genipin, a known UCP2 inhibitor. Genipin abrogated the changes in MMP and redox state seen with septic serum, although exacerbated ROS generation (Table 1). This indicates that UCP2 plays an important role in decreasing MMP via uncoupling but that this may be protective by preventing greater increases in ROS.

Scavenging ROS with the antioxidant mitoTEMPO prevented the changes observed with septic serum if given at Time 0 (Table 1), and largely reversed them when administered part-way through exposure (Figure 1). This provides further evidence for the central importance of ROS as both a consequence of, and a signalling mechanism within the process of mitochondrial dysfunction in sepsis. The reversibility of changes through scavenging ROS also indicates the potential for therapies to be effective following a septic insult.

In attempting to identify the mediators within serum responsible for the changes observed, fractionation suggested that components ≥50 kDa were primarily responsible (Table 1). The main constituent of serum above 50 kDa is albumin, which has previously been suggested as an important mediator in critical illness, due to its ability to transport several substances, many of which may be deleterious or useful in signalling pathways.⁵ Inclusion of exogenous fresh albumin (equivalent to 60 g/l) partially prevented the dysfunction caused by septic serum (Table 1), perhaps by providing clean binding sites for humoral mediators free within the serum to attach to, thus rendering them ineffective.

I have thus shown how humoral factors circulating in septic serum can cause mitochondrial dysfunction in naïve kidney slices. ROS appears to be a key mediator of these changes, scavenging of which may prove protective. Uncoupling may however offer a protective negative feedback, reducing ROS and thereby shielding the cell from oxidative damage. Exploring which specific components of the serum cause mitochondrial dysfunction, and the mechanisms through which exogenous albumin attenuates mitochondrial dysfunction are warranted.

I was awarded a UCL Medical School MBPhD Programme bursary, an Astor Foundation Research scholarship, and an Intensive Care Foundation New Investigator Award which funded my work described here.

Table 1.

Relative fluorescence intensity (%) at 90 minutes compared to the 0 minute baseline recording (mean ± SD).

	Control	Septic	Septic +  Genipin	Septic +  MitoTEMPO	Sham conc. +  septic filt.	Septic conc. +  sham filt.	Septic +  Albumin
HEt	86 ± 4	145 ± 4*	164 ± 10*†	81 ± 8†	106 ± 8*†	135 ± 8*¶	127 ± 8*†
TMRM	98 ± 5	64 ± 5*	80 ± 6*†	86 ± 6*†	95 ± 4†	66 ± 4*¶	81 ± 7*†
NADH	101 ± 7	78 ± 5*	95 ± 6†	90 ± 7†	96 ± 4†	85 ± 3*¶	83 ± 6*
MitoSOX	97 ± 8	140 ± 7*	–	98 ± 10†	–	–	–
Calcein AM	86 ± 4	84 ± 5	80 ± 7	86 ± 7	–	–	86 ± 6

HEt = dihydroethidium, TMRM = tetramethyl rhodamine methyl ester, NADH = reduced nicotinamide adenine dinucleotide, MitoSOX = mitochondrial superoxide indicator, calcein AM = calcein acetoxymethyl ester.

^*p < 0.05 vs sham serum, ^†p < 0.05 vs septic serum, ^¶p < 0.05 vs sham concentrate + septic filtrate.

Microvesicles are key mediators of inflammation in acute lung injury

S Soni, M.R Wilson, K O’Dea and M Takata

Section of Anaesthetics, Pain Medicine and Intensive Care, Department of Surgery and Cancer, Imperial College London

Abstract

Acute lung injury (ALI) and its clinical presentation Acute Respiratory Distress Syndrome (ARDS) have an unacceptably high mortality.¹ Treatment remains principally supportive² despite a vast amount of clinical/pre-clinical research. Therefore, a re-direction in ALI research is required, moving away from highlighting individual inflammatory mediators and instead investigating how inflammatory cargo are transmitted between cells. Microvesicles (MVs) are membrane-circumscribed extracellular particles (size: 100–1000 nm), providing an alternative yet crucial role in intercellular communication by carrying a variety of bioactive cellular cargo.³ However, their role in ALI/ARDS remains unknown. We hypothesised that MVs are key to the pathogenesis of ALI as mediators of intercellular communication. Our aims were to examine: the role of MVs in the pathophysiology of ALI; the inflammatory mediators carried by alveolar MVs; how MVs communicate with target alveolar cells; and MV populations in patients with severe inflammatory lung disease.

We used various in vitro/in vivo techniques and developed a number of intricate experimental protocols (using flow cytometry, ELISA, western blotting, confocal microscopy, immune-electron microscopy) to fully characterise/phenotype MVs. We identified MVs populations in bronchoalveolar lavage fluid (BALF) taken from a lipopolysaccharide-induced murine model of ALI. Primary alveolar macrophages/bone marrow-derived macrophages were harvested from untreated mice and stimulated with lipopolysaccharide (1 µg/ml) and ATP (3 mM) to produce MVs. To assess their inflammatory activity, these MVs were incubated with epithelial cells in vitro or instilled intratracheally into mice. To ascertain MV-mediated communication within the alveolus, labelled MVs were instilled intratracheally into the lungs of mice or incubated within a co-culture of primary alveolar macrophages and epithelial cells (under inflammatory/anti-inflammatory conditions). In addition, pharmacological/molecular inhibitors were used to identify mechanisms of MV-mediated communication. Finally, 61 patients with severe chronic obstructive pulmonary disease (COPD) underwent rigorous clinical evaluation using symptom/lung function/severity scores. BALF/blood samples were obtained, analysed for MVs and correlated to clinical markers/indices of disease severity.

Initially we profiled MV populations produced during ALI. We demonstrated for the first time, that MVs from different intra-alveolar precursor cells are sequentially produced within the alveolar space early in the course of ALI. We focussed upon alveolar macrophage-derived MVs since they were the dominant population and found they packaged significant amounts of TNF rather than other pro-inflammatory material (e.g. IL-1β/IL-6). These alveolar macrophage-derived MVs increased epithelial cell ICAM-1 expression, via a TNF dependent mechanism (p < 0.001). When instilled intratracheally into mice, these MVs induced changes consistent with ALI e.g. increased BALF neutrophils, protein and epithelial cell ICAM-1 expression (p < 0.05). Therefore these alveolar macrophage-derived MVs, are potent initiators of ALI, mediated by their molecular cargo, particularly TNF and these results highlight that MVs are key components in the pathophysiology of ALI/ARDS.

Following this, we examined how cells packaged TNF into MVs. In severe sepsis/ARDS, substantive cell injuries occur with subsequent release of danger signals (e.g. ATP).⁴ We discovered that ATP totally redirects TNF trafficking, inhibiting soluble TNF release from cells, instead preferentially packaging the membrane TNF isoform into potent, pro-inflammatory MVs. Within MVs, TNF is protected from extracellular degradation, dispersion or neutralisation. Using pharmacological/genetic inhibitors, we discovered that TNF secretion within MVs bypasses the conventional endoplasmic reticulum/Golgi-dependent pathway (despite being a classically-released cytokine), and is mediated by acid sphingomyelinase. Unlike TNF + ve MVs, TNF-ve MVs (harvested from TNF knockout mice) did not produce inflammation in vivo (assessed using BALF neutrophils/monocytes, protein, CXCL1 and epithelial cell ICAM-1 expression (p < 0.05)) highlighting the physiological importance TNF enclosed within MVs. These data, offer important, previously unappreciated clinical implications. Clinical trials of anti-TNF therapy have shown little beneficial impact in sepsis/ARDS, despite ample preclinical evidence of TNF involvement in their pathophysiology.⁵ In severe sepsis/ARDS, substantial ATP release takes place, which would lead to secretion of TNF in MVs (rather than soluble form). Antibody-based blocking treatments may not be effective to inhibit the effects of TNF enclosed within MVs. These data suggests that targeting MV-mediated TNF signalling (by reducing MV internalisation/production) may be essential to block TNF biological effects in sepsis/ARDS.

We next investigated how MVs communicate with target cells within the alveolus. Although both epithelial cells and alveolar macrophages internalise MVs, we discovered that alveolar macrophages take up the overwhelming majority, regardless of the environmental condition (p < 0.001). Using pharmacological/molecular inhibitors, we demonstrated that epithelial cell uptake is dependent on integrin/phosphatidylserine binding whilst alveolar macrophage uptake is more dependent on scavenger receptors indicating clear mechanistic differences between these cell types. Our results strongly suggest that alveolar macrophages play a crucial role in modulating MV processing within the alveoli, regulating MV-mediated intra-alveolar inflammation. These data highlight future areas of study, where MV signalling can be interrupted within the alveolar space.

Finally, we consolidated and translated our in vitro/in vivo data by characterising different MV populations in 61 patients with severe COPD. We confirmed the presence of multiple MV populations within BALF including leukocyte, neutrophil, monocyte, alveolar macrophage and epithelial MVs. We assessed whether these BALF MV populations could be potential biomarkers for COPD and found that BALF neutrophil MVs strongly correlated with a broad range of functional and clinically relevant disease severity indices including symptoms (MRC dyspnoea score) (p < 0.01); exercise tolerance (6 minute walk test) (p < 0.01); disease severity (FEV₁/RV/TLC) (p < 0.05); exacerbation frequency (p < 0.01) and mortality (BODE index) (p < 0.01). These data strongly corroborate our in vitro/in vivo data and suggest that BALF neutrophil MVs are a novel, clinically relevant biomarker for COPD disease severity, and may have a pathogenic role in COPD progression.

In conclusion, we have demonstrated a number of breakthrough findings which have considerable translatable and clinical significance. Our data provide substantial evidence that MVs are key components in the pathophysiology of lung inflammation/ARDS. Furthermore in these clinical scenarios, where there is a distinct lack of biomarkers, MVs are readily measureable indicators of disease severity that could be used in clinical practice. FInally, MVs are disease-modifiable targets and our studies have pioneered novel diagnostic and therapeutic perspectives in patients with ARDS, where no curative treatment currently exists.

We gratefully acknowledge the Medical Research Council and British Journal of Anaesthesia for funding this work.

All studies received ethics committee approval.

Improving sepsis resuscitation with reinforcement learning

M Komorowski

Department of Surgery and Cancer, Imperial College London, London, UK

Department of Bioengineering, Imperial College London, London, UK

Laboratory of Computational Physiology, Harvard-MIT Division of Health Sciences & Technology, Cambridge, Massachusetts, USA

Abstract

Sepsis is the third leading cause of death worldwide and the main cause of mortality in hospitals¹, but the best treatment strategy remains uncertain. In particular, evidence suggests that current practices in the administration of intravenous fluids and vasopressors are suboptimal and likely induce harm in a proportion of patients^1,2.

To tackle this sequential decision-making problem, I used reinforcement learning, a class of machine learning algorithms whose goal is to estimate an optimal set of rules that maximises some form of reward.³ Similarly, a clinician’s goal is to make therapeutic decisions in order to maximize a patient’s probability of a good outcome.

An artificial intelligence (AI) agent was built and validated on two large non-overlapping ICU databases, containing data routinely collected from adult patients in the U.S.A. The Medical Information Mart for Intensive Care version III (MIMIC-III) was used for model development and the eICU Research Institute Database (eRI) for model testing. All patients included fulfilled the sepsis-3 criteria. The final datasets contain 17,083 admissions from MIMIC-III from 5 separate ICUs in one hospital and 79,073 admissions from 128 different hospitals from eRI. I extracted up to 72 hours of data per patient, in the form of time series of 48 variables, including demographics, Elixhauser premorbid status, vital signs, laboratory values, fluids and vasopressors received.

A Markov decision process (MDP) was used to model the patient environment and trajectories.³ The various elements of the model were defined using patient data time series from the training set (Figure 1). I deployed the AI agent to solve the MDP and predict outcomes of treatment strategies. First, I evaluated the actual treatments of clinicians, by analysing all the prescriptions and computing the average return of each treatment option, which can take values from −100 to +100 in my model. Then, the MDP was solved using an algorithm called policy iteration, which identified the treatments that maximised return, that is, the expected 90-day survival of patients.³ This process generated a new treatment strategy (termed “a policy” in reinforcement learning) for sepsis optimizing 90-day mortality, which could be applied in real-time at the patient level.

Figure 1: Data flow of the AI agent. Eighty percent of the MIMIC-III dataset was used to define the elements of the Markov decision process. Time series of patient data were clustered into finite states. The dose of intravenous fluids and vasopressors were discretized into 25 possible actions. Patients’ survival at 90 days after ICU admission defined rewards. Reinforcement learning was used to estimate optimal treatment strategies. The remaining 20% MIMIC-III data was used to identify the best model among 500 candidates, which was then tested on an independent dataset from the eRI database.

It was crucial to obtain reliable estimates of the performance of this new policy without deploying it, since executing a bad policy would be dangerous for patients^3,4. In reinforcement learning, evaluating the performance of a new policy using trajectories generated by another policy (the clinicians’ policy) is termed off-policy evaluation^3,4. I implemented a type of off-policy evaluation called weighted importance sampling, and used bootstrapping to estimate the true distribution of the value of the AI policy.⁴ 500 candidate models were built, and I selected the one maximising the theoretical statistical safety of the new AI policy. Then, the chosen AI policy was tested on the independent eRI dataset.

Figure 2a shows the distribution of the estimated value of the clinicians’ policy and the AI policy in the selected final model tested on the eRI cohort. The median value of the AI policy, at 84.5 (interquartile range 84.3–87.7), was reliably higher than that of clinicians’ (median 56.9, interquartile range 54.7–58.8). Figures 2b and 2c show the distribution of treatment doses according to clinicians’ and AI policies. On average, the AI agent recommended lower doses of intravenous fluids and higher doses of vasopressors, compared to clinicians’ actual treatments. While the proportion of time the eRI patients received vasopressors was only 17%, that would have been 30% if following the AI agent.

I analysed patient mortality when the dose actually administered corresponded to or differed from the dose suggested by the AI agent. The patients who received doses similar to the doses recommended by the AI agent had the lowest mortality. Giving more or less than the AI policy of either treatment was associated with increasing mortality rates, in a dose-dependent fashion (Figures 2d and 2e). When the actual dose given was different from the suggested dose, clinicians gave more or less fluids in similar proportions, and less vasopressors 75% of the time. The median dose deficit in patients who received too little vasopressors was 0.13 µg/kg/min (interquartile range 0.04–0.27 µg/kg/min).

Figure 2. Comparison of clinicians and AI policies in eRI (a, b, c); Average dose excess received per patient of both drugs in eRI and corresponding mortality (d, e). a. distribution of the estimated value of the clinicians’ and the AI policy in the selected model. b and c, Visualization of the clinicians’ and AI policies. All actions were aggregated over all time steps for the 5 dose bins of both medications. On average, patients were administered more intravenous fluid (b) and less vasopressor (c) medications than recommended by the AI policy. Vasopressor dose is in µg/kg/min of norepinephrine equivalent and intravenous fluids dose is in mL/4 hours. d and e, The dose excess refers to the difference between the given and suggested dose averaged over all time points per patient. The shaded area represents the standard error of the mean. In both plots, the smallest dose difference was associated with the best survival rates (vertical dotted line). The further away the dose received was from the suggested dose, the worse the outcome.

In this research, I demonstrated how reinforcement learning could be applied to solve a complex medical problem, and suggest individualized treatment strategies for sepsis. When clinicians’ actual treatments varied from the AI agent’s suggested policy, this was most commonly to administer too little vasopressor. Early use of low-dose vasopressor has been suggested in sepsis^2,5. This may avoid administering excessive amounts of fluids, which has been linked with poorer outcome^2,5. My results support this strategy but importantly allow the treatment to be individualized for each patient.

My vision is that this system will be used in real-time, using data streams in electronic health records to suggest a course of action. Physicians will always need to make subjective clinical judgments about treatment strategies, but computational models can provide additional insight about optimal decisions, avoiding targeting short-term resuscitation goals and instead following trajectories towards longer-term survival.

There are limitations to this study. Although the datasets are large and used routinely collected clinical data, some sites and patients had to be excluded due to poor quality data recording or missing data. Due to differences between the two datasets, slightly different implementations of the sepsis-3 criteria were used, and hospital mortality was used in eRI instead of 90-day mortality. Finally, some laboratory values would not have been immediately available to clinicians to inform decision making but were available to the AI agent.

This work will clearly require prospective evaluation using real-time data and in clinical trials, but if only a few percent reduction in mortality from sepsis could be achieved, this would represent several tens of thousands of lives saved annually, worldwide. In the last 10 to 15 years, attempts to develop novel treatments to reduce sepsis mortality have uniformly been unsuccessful^2,5. The use of computer decision support systems to better guide treatments and improve outcomes is therefore a much-needed approach.

Acknowledgements: I am grateful to my PhD supervisors, Professor Anthony C. Gordon and Doctor A. Aldo Faisal, as well as to Leo Celi from the Massachusetts Institute of Technology and Harvard University and to Omar Badawi from Philips Healthcare for their support in this project.

Direct measurement of intracellular mitochondrial respiration in skeletal muscle taken from critically ill patients demonstrates functional alterations with potential adaptive implications

HT McKenna^1,2, AJ Murray³ and DS Martin^1,2

¹Royal Free Hospital Intensive Care Unit

²Division of Surgery and Interventional Science, University College London

³Department of Physiology, Development and Neuroscience, University of Cambridge

Abstract

The syndrome of critical illness, which can result from many different triggering insults, is poorly understood at a cellular level. The prevailing theory, which underpins most clinical practice on intensive care units, is that this syndrome is a consequence of inadequate oxygen delivery to the mitochondria, such that they cannot produce enough ATP to power the processes essential for the maintenance of cell integrity and function. Many intensive care interventions target the ‘normalisation’ of factors that determine global oxygen dispatch (such as arterial oxygen content and cardiac output). Achievement of these targets has not been reliably associated with improvements in clinical outcome (1) and the interventions themselves may be associated with harm (2). Evidence from animal models and indirect studies of mitochondrial complex activities or metabolites in critically ill patients have suggested that mitochondrial function itself is modified during the stress of critical illness, and that the degree of dysfunction is related to clinical outcomes (3). The aim of this study was to characterise skeletal muscle mitochondrial respiratory function in critically ill patients, and to determine if any measured functional indices were associated with survival.

Subjects were recruited from a mixed intensive care unit (ICU) at the Royal Free Hospital. Vastus lateralis biopsies were taken from critically ill patients within 48 hours of their admission to the ICU (n = 21; 8 females, mean age 61.4 years, 95% CI 54.4–68.4), and from age- and gender matched control patients undergoing elective hip replacement under general anaesthesia (n = 9, 4 female, mean age 64.2 years, 95% CI 56.1–72.3). The critically patients had a mean APACHE II score of 28.2 (95% CI 24.8–31.6) and mean SOFA score of 12.7 (95% CI 11.5–13.9), and 9 (43%) survived to hospital discharge. All were receiving at least two forms of organ support, including vasopressors (>0.1 mcg/kg/min) and mechanical ventilation.

Muscle fibre bundles were dissected and the sarcolemmal membrane permeabilised with saponin. Mitochondrial respiration was measured using an Oxygraph-2K (Oroboros Instruments, Innsbruck, Austria), which monitors real-time oxygen consumption in tissue preparations, in conjunction with a substrate-uncoupler-inhibitor protocol designed to quantify the following components of mitochondrial respiration: LEAK (oxygen consumption not coupled to ATP production); fatty acid oxidation (FAO) capacity, complex I capacity (CI), complex II capacity (CII), maximal oxidative phosphorylation (OXPHOS), and maximal capacity of the electron transport system (ETS).

Maximum respiratory capacity per mass of tissue did not differ in early critical illness compared with controls, but specific components of mitochondrial respiration, expressed relative to the maximum oxygen consumption rate, were lower in the critically ill patients, namely LEAK (p = 0.025), FAO (p = <0.0001) and complex I (p = 0.014) capacities. Critically ill patients who survived to hospital discharge (n = 9) did not differ from non-survivors (n = 12) in their maximum respiratory capacity per mass of tissue, but did demonstrate a lower relative complex I capacity (p = 0.012) than non-survivors.

Figure 1. Mitochondrial respiratory function as measured using high resolution respirometry in conjunction with a substrate-uncoupler-inhibitor protocol. A illustrates the maximum oxygen consumption rates (OCR) per mg muscle tissue, for oxidative phosphorylation (OXPHOS) and the electron transport system unlimited by the phosphorylation system (ETS) in critically ill patients vs controls. B shows the capacity of individual components of respiration: LEAK, fatty acid oxidation (FAO), complex I (CI) and II (CII) relative to the maximum capacity (ETS). C shows the maximum OCR per mg in survivors and non-survivors of critical illness and D demonstrates the relative capacity for individual components between these two groups. *p < 0.05, **p < 0.01, *** p < 0.001, **** p < 0.0001

This is the first study in which direct measurements of intact mitochondrial function in fresh skeletal muscle have been measured in patients during the acute phase of critical illness. It shows that critically ill patients undergo functional changes in specific intra-mitochondrial components within the first 48 hours, rather than a global change in respiratory capacity. It supports the findings of a recent study that reported decreased fatty acid oxidation in the early phase of critical illness (4), but goes further to show that such changes do actually reflect a reduction in the capacity of the mitochondrial machinery to respire using fatty acid substrates. The use of fatty acids is known to produce fewer ATP molecules per oxygen molecule consumed, compared with carbohydrate, and thus this switch in early critical illness could represent an adaptive response towards increased efficiency. Similarly, the lower capacity for leak respiration (oxygen consumption not coupled to ATP synthesis) in critically ill patients may reflect another component of adaptation to promote mitochondrial efficiency during a period of cellular stress.

Complex I is a known producer of reactive oxygen species (ROS) in response to cellular stress (5), and the finding of lower complex I capacity in early critical illness could thus represent a specific downregulation to reduce exposure to damaging ROS. The fact that those patients who survived to hospital discharge had lower complex I capacities during the first 48 hours, compared with non-survivors, corroborates the proposal that early downregulation of complex I capacity in response to stress may be a protective response.

This study revealed previously unknown mechanisms for how the end-point of all our clinical interventions on ICU, the mitochondria, alter during critical illness in real human patients. The results help explain why clinical interventions targeting systemic oxygen dispatch may not improve patient outcomes, and instead highlight specific subtle mechanisms within the cell, which represent potential therapeutic targets to improve our treatment of this high mortality syndrome.

This study was funded by the Intensive Care Foundation and the Royal Free Charity.

All protocols received ethical approval from the Camden and Islington NHS Research Ethics Committee and local site approval from the Royal Free Hospital Research and Development Department.