Mobile extracorporeal membrane oxygenation service for severe acute respiratory failure – A review of five years of experience

Abstract

Provision of extracorporeal membrane oxygenation as part of support escalation in severe refractory acute respiratory failure in England is provided by five specialist centres that operate within a well-defined quality and safety framework. We conducted a qualitative study of the extracorporeal membrane oxygenation retrieval service provided by one of the five centres. We analysed 176 consecutive debrief reports written between October 2013 and April 2018 by the consultant. Main identified issues were short delays in retrieval predominantly due to insufficient communication or equipment failure. All issues were addressed in subsequent practice. Our results suggest a need for improved communication between the referring intensive care unit and retrieving team. Our findings highlight the value of regular reflection-based evaluation to ensure continued provision of safe and efficient service.

Introduction

Supportive management including least damaging mechanical ventilation and prone positioning remains the approach of choice in management of acute respiratory distress syndrome (ARDS).¹ Added benefit of extracorporeal membrane oxygenation support (ECMO) remains controversial despite several attempts to indisputably justify its use through randomized control trials.^2–4 Notably, the improved outcomes associated with transfer of patients to a specialist ECMO centre identified in the CESAR trial³ form the foundation of the current national ECMO service provision in England. This service is formed of five specialist ECMO centres commissioned by the National Specialised Commissioning Group. Each centre supports a network of regional hospitals by the means of a specialist ECMO assessment and retrieval team available 24 h a day, seven days a week.^5,6 Centres support each other to ensure each referred patient is provided for.

A specialist team will travel to the patient bedside when both referring and ECMO team agree that the patient is perceived to have a potentially reversible condition. The ECMO team will usually rely on information collected on a standard referral form⁷ and radiology imaging sent across. The ECMO team will review the patient on site with three potential outcomes: continued management at the referring centre with advice from the retrieval team if the patient is found to have potential for improvement on conventional therapy or to lack potential for recovery; retrieval on mechanical ventilation with view to optimize treatment at the specialist centre; retrieval on ECMO support established at the referring hospital.^8,9

This study evaluates provision of ECMO assessment and retrieval service by a specialist ECMO centre with the intention to recognize factors impacting on service provision and identify potential areas for improvement.

Methods

We conducted an institutionally approved single-centre retrospective qualitative study of 176 consecutive ECMO retrieval debrief reports written in the period between October 2013 and April 2018 by the consultant leading the retrieval team. The non-structured debrief 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.

Data collection

All reports were screened and extracted data included about patient demographics, grounds for referral, duration of stages of retrieval, on-site assessment including patient’s condition on arrival and conclusion of the assessment, discussion with family, details of the cannulation procedure and over-all reflections and lessons learned. Free text entries were reviewed and assigned to different categories.

Definitions

Completed on-site assessment was defined as a fully executed decision, resulting from the on-site assessment, to initiate ECMO, retrieve patient without ECMO or continue conservative care at the referring hospital with advice from the ECMO retrieval team.

Decision to conservatively manage the patient at the referring hospital was made either due to perceived futility owing to irreversibility of the condition or lack of physiological reserve defined by Gillon et al.⁸ as “the individual’s capacity to survive the stresses of critical illness and is determined largely by comorbidities”⁸or due to improvement in patient’s condition as a result of advice given by the specialist team about trial of proning, ventilatory support settings and medication or improvement on conventional therapy since the initial assessment.

Three key phases of a retrieval were identified and defined as: dispatch phase, from acceptance to retrieving team departure; assessment phase, from arrival at the referring hospital to ECMO initiation, which includes consultant handover at arrival, assessment of the patient, discussion with patient’s next of kin, transfer to theatre and cannulation procedure; stabilization phase, from ECMO initiation to retrieving team departure.

Statistical analysis

Descriptive statistical analysis was performed using R¹⁰ to calculate mean and standard deviation for continuous variables. Counts and percentages were calculated for categorical variables. Reported causes of respiratory failure are based on referral information rather than definitive discharge diagnosis. Every cannula insertion was counted as a separate procedure. Chi-squared test was performed for categorical variables.

Retrieval index

To assess potential impact of retrieval distance on the number of retrievals from a given hospital, distance between referring hospitals within our region and the specialist ECMO centre was determined using Google maps (https://www.google.com/maps), an online navigation programme, set to calculate the shortest route¹¹ in miles with one decimal place accuracy. To account for the size of the referring hospital, 2016/17 ITU admission statistics for each trust¹² were used to calculate the average number of retrievals from the hospital for every 1000 ITU admissions (retrieval index). Impact of the distance of the referring hospital from the ECMO centre as a predictor of the retrieval index was determined using a linear regression by adding a line of best fit. Suitability of the model was assessed in R¹⁰ using diagnostic plots. Statistical significance of the regression was determined using ANOVA under the assumption that distance was a categorical variable because of the limitations of retrospective analysis and the online navigation programme.

Results

The team handled 734 referrals in the period between October 2013 and April 2018 of which 176 from 48 hospitals were accepted for on-site assessment by a specialist retrieval team.

Three patients were referred for mechanical cardiac support and excluded from the analysis. Six (3.5%) patients died at the referring hospital at various stages of the retrieval (Table 1). Cause of death was not specified in the retrieval reports.

Table 1.

Number of deaths at various stages of the retrieval process.

Stage of retrieval	Number of deaths (%)
Before team arrival at the referring hospital	1 (0.6)
During assessment	1 (0.6)
During ECMO cannulation	1 (0.6)
During stabilization phase	3 (1.7)

ECMO: extracorporeal membrane oxygenation.

Of the total of 167 completed retrievals, 31 (17.9%) were declined ECMO following on-site review either due to perceived futility following on-site review¹³ or improvement in patient’s condition¹⁴ as a result of advice given by the retrieving team¹⁵ or improvement on conventional therapy since initial assessment²; 124 (71.7%) were established on ECMO at the referring site; 12 (6.8%) were retrieved without ECMO, 6 (3.4%) of whom subsequently required ECMO after transfer to the specialist ECMO centre (Figure 1).

Figure 1.

Study flowchart.

The mean age of the patients was 42.1 (±15.9) years, the mean BMI was 31.0 (±9.1) and 98 (57%) were male. Cause of the severe acute respiratory failure suggested at referral was specified in 155 (89.6%) cases. Bacterial infections (n = 73; 42.2%) were the leading cause, followed by other respiratory causes such as viral infections (n = 21, 12.1%), aspiration pneumonias (n = 14; 8.1%), asthma exacerbations (n = 13; 7.5%), trauma (n = 7; 4%) and non-respiratory causes such as pancreatitis (n = 4; 2.3%). The cause was not specified in 19 (11%) reports (Table 2).

Table 2.

Patient characteristics.

Patient characteristics	Mean (±SD) or n (%)
Age	42.1 (±15.9)
BMI	31.0 (±9.1)
Male	98 (57%)
Murray score	3.1 (±0.6)
RESP score	3.2 (±3.6)
Aetiology of respiratory failure	n (%)
Bacterial	73 (42.2%)
Viral	21 (12.1%)
Aspiration	14 (8.1%)
Asthma	13 (7.5%)
Trauma	7 (4%)
Pancreatitis	4 (2.3%)
Fungal	3 (1.7%)
Pulmonary embolism	2 (1.3%)
Other	17 (9.8%)
Not specified	19 (11%)

Retrieval process

The 176 retrievals during the period between October 2013 and April 2018 were handled by 25 consultant intensivists, 14 ECMO nurse specialists and 24 perfusionists. Median number of retrievals performed by each individual member of the team is detailed in Table 3. Training staff were present on 40 (23.1%) retrievals.

Table 3.

Number of retrievals per each individual member of the team.

Team member	Median (interquartile range)
Consultant Intensivist	6 (3–10)
ECMO nurse specialist	6 (2–19)
Perfusionist	5 (3–9)

ECMO: extracorporeal membrane oxygenation.

Duration of key phases of ECMO retrieval is highlighted in Table 4. A total of 49 (29.3%) of the retrievals took place on a weekend (defined as 00:01 a.m. Saturday to 00:01 a.m. Monday).

Table 4.

Duration of key phases of ECMO retrieval.

Phase of a retrieval	Mean (±SD)
Dispatch phase	1 h 18 min (±37 min)
Assessment phase	2 h 51 min (±1 h 4 min)
Stabilization phase	2 h 01 min (± 55 min)

ECMO: extracorporeal membrane oxygenation.

In 44 (25%) reports, the retrieving consultant intensivist subjectively noted excellent communication with the referring team and teamwork among the retrieving team. Main issues reported during retrievals are reported in Table 5.

Table 5.

Issues reported during key phases of ECMO retrieval.

	n (%)
Dispatch phase
Delays in ambulance availability	4 (2.3%)
Ambulance malfunction	1 (0.6%)
Navigation problems	1 (0.6%)
Assessment phase
Delays on arrival	6 (3.5%)
Waiting for family	2 (1.2%)
Operating theatre unavailability	6 (3.5%)
Blood product unavailability	6 (3.5%)
Missing equipment	6 (3.5%)
Lack of spare equipment	5 (2.9%)
Stabilization phase
Equipment failure	5 (2.9%)
Difficulties securing bariatric  patients on the trolley	10 (5.6%)
Delays on arrival to ECMO centre	6 (3.5%)

ECMO: extracorporeal membrane oxygenation.

Communication issues with the referring team were reported in eight (4.6%) cases, including lack of consultant handover at arrival (3; 1.7%), medication administration issues (1; 0.6%) and lack of interpreter (1; 0.6%).

Retrieval index

There was a statistically significant negative correlation between the referral index and retrieval distance (ANOVA p = 0.004248) (Figure 2).

Figure 2.

Retrieval index as a function of roundtrip distance from the specialist ECMO centre. ECMO: extracorporeal membrane oxygenation.

Cannulation procedure

Right femoral vein was the preferred cannulation site for the drainage cannula (92; 74%) and right internal jugular vein for the return cannula (108; 87%). There were two (1.6%) femoro-femoral cannulations and two (1.6%) dual-lumen cannulas inserted in the right jugular vein. There were 16 bedside cannulations (14.4%). Complications were reported in 53 (21%) cannulations, mainly due to guidewire kinking (32; 64%).

Location of cannulation was reported in 114 cases, of which 34 (15%) procedures were performed at the bedside and 194 (85%) in theatres. Complications were encountered in 9 (26%) of bedside cannulation procedures and 41 (21%) of theatre cannulations. Of the 50 complicated cannulations, 32 were due to guidewire kinking and in 12 cases the cause not specified. There was no statistically significant difference between number of reported complications during procedures performed at bedside as opposed to in theatre (p = 0.4879).

At the ECMO centre

On arrival, 111 (90%) of patients retrieved on ECMO had a whole-body CT before transfer to the intensive care unit.

Discussion

All issues identified in this evaluation were discussed during team debrief and addressed in subsequent practice. Prevalence of critical events reported in our evaluation was relatively low compared with results previously reported by Blecha et al.,¹⁵ who reported 9 transportation delays, 5 equipment-related and 3 transportation vehicle-related issues during retrieval of 431 patients, 52 of whom were retrieved on ECMO.

We found no difference in the proportion of complications reported in the overall number of cannulations performed at the bedside as opposed to in theatre. Our current practice to preferably cannulate patients in theatre was introduced in 2015 as a result of a shared perception that bedside cannulations were more challenging. Our data set is too small to confidently evaluate the impact of this change in practice on the number of reported complications. Since 2015, the decision to cannulate at the bedside is rarely undertaken and only after careful consideration of patient safety and potential complications.^14,16 The consideration is often prompted by concurrent unavailability of an operating theatre.

Similarly, our current protocol includes a whole-body CT scan on arrival to the specialist ECMO centre. In 10% of retrievals, the scan was delayed owing to a critical condition of the patient that required immediate transfer to the critical care unit or unavailability of the CT scanner.

The large variability and seemingly small average number of ECMO retrievals per team member does not account for over-time changes in staffing and any experience acquired by the team members prior to October 2013, during training or as part of a different ECMO centre. All members of the retrieval team receive regular training in both all aspects of ECMO support to ensure safety and quality of the service.

Reported communication issues as well as issues such as delays on arrival to the referring centre and unavailability of blood products or operating theatre suggest a need for change in our protocol to improve communication and information sharing with the referring team with the aim to prevent complications arising from miscommunication and insufficient experience.

We identified a large variability in retrieval index between the referring hospitals ranging from 0.3 to 5.4 retrievals per 1000 ITU patients. Contrary to previous reports,⁸ our findings suggest that retrieval distance may partially explain this variability. This may simply be due to potential risks associated with a longer journey counterbalancing the added benefit of ECMO. Charlesworth et al.¹³ also proposed that in addition to geography, characteristics and experience of the referring centre may have an impact on the decision-making process regarding ECMO retrievals. Owing to variability in practice, workload and capacity among the referring centres and the five specialist ECMO centres, no generally pertinent conclusions can be made at this point about the reason for variation in retrieval index among the referring centres.

Recognition of futility in relatively young critically ill patients is a difficult task especially when it comes to highly advanced life-sustaining treatment.^8,9 Retrospective nature of our study enables clear segregation of the 173 patients assessed on-site into two distinct categories: suitable and unsuitable for ECMO illustrated in Figure 1. This seemingly straightforward flow-chart fails to appreciate the complexity of each decision process. Each of the 16 decisions not to establish the patient on ECMO due to perceived futility were a result of a detailed discussion with the referring team and the patient’s family taking into account the current clinical status as well as patient’s physiological and psychological reserve, current scientific evidence and clinical experience of the specialist ECMO team.^8,9 Each decision was also consulted with the on-call intensivist at the specialist ECMO centre highlighting the critical value of a specialist opinion and holistic approach in clinical decision making especially with regard to unique, controversial and highly specialist services such as ECMO.^17–19

Limitations

This study had several limitations. Firstly, an evaluation of six years of experience in a single retrospective observational study fails to appreciate change of practice as a result of technological developments, increased experience of the retrieving team and advancement of knowledge about aetiology and management of severe refractory ARDS, such as introduction of protective ventilation and prone positioning.¹

Secondly, the study was conducted with a relatively small dataset from a single tertiary ECMO centre. Each of the five ECMO centres in England has a unique retrieval routine and distinctive patient demographics; hence, our results may not be representative of the national ECMO service.

Finally, the reports were introduced in 2013 with no rigid format or style as an opportunity for the retrieving consultant to reflect on the retrieval and identify areas for improvement. While excellent for the original purpose, the variable style and contents of the reports resulted in an incomplete dataset which negatively impacted the statistical accuracy and power of our study. Reflective learning should remain the main purpose of the reports; however, our findings suggest a possibility of introducing a list of details to be included in the reflection to ensure consistency. This will hopefully facilitate use of this great resource to re-evaluate the service in the future, while maintaining the case-to-case flexibility of the format.

Conclusion

Provision of ECMO as part of treatment escalation in severe acute respiratory failure in England remains limited to five specialist ECMO centres to ensure safety and maintain high quality. Our evaluation of the service provided by one of the five centres emphasizes the necessity of no-less-than excellent communication with the local team and highlights the need for regular reflection-based evaluation of the service in order to ensure continued provision of safe and efficient service.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.