A retrospective analysis of inpatient cardiac arrests over one year at a tertiary heart attack and cardiothoracic centre identifying potential candidates for an inpatient extracorporeal cardiopulmonary resuscitation service

Abstract

Introduction

Extracorporeal cardiopulmonary resuscitation (ECPR) is an internationally recognised treatment for refractory cardiac arrest, with evidence of improved outcomes in selected patient groups from cohort studies and case series. In order to establish the clinical need for an in-hospital extracorporeal cardiopulmonary resuscitation service at a tertiary cardiac centre, we analysed the inpatient cardiac arrest database for the previous 12 months.

Methods

Evidence-based inclusion criteria were used to retrospectively identify the number of patients potentially eligible for extracorporeal cardiopulmonary resuscitation over a 12-month period.

Results

A total of 261 inpatient cardiac arrests were analysed with 21 potential extracorporeal cardiopulmonary resuscitation candidates meeting the inclusion criteria (1.75 patients per month, or 8% of inpatient cardiac arrests (21/261)). The majority (71%) of these cardiac arrests occurred outside of normal working hours. Survival-to-discharge within this sub-group with conventional cardiopulmonary resuscitation was 19% (4/21).

Conclusion

Sufficient numbers of refractory inpatient cardiac arrests occur to justify an extracorporeal cardiopulmonary resuscitation service, but a 24-h on-site extracorporeal membrane oxygenation team presents a significant financial and logistical challenge.

Introduction

Extracorporeal cardiopulmonary resuscitation (ECPR) refers to the use of veno-arterial extracorporeal membrane oxygenation (VA-ECMO) to support a patient’s circulation during cardiac arrest.^1,2 This involves inserting a cannula into a large vein, pumping venous blood through a membrane lung, which adds oxygen and removes carbon dioxide, and returning the blood under pressure through a second cannula placed in a large artery. The normal configuration in ECPR is to place a drainage cannula in the femoral vein and a return cannula in the femoral artery. This oxygenated blood is returned at such a rate that the arterial system is pressurised and the blood flows to the patient’s vital organs. It is possible to provide blood flows akin to those of a normal cardiac output. In conventional CPR (CCPR), chest compressions (either manual or mechanical) provide much lower blood flows to the vital organs. The duration of CCPR is inversely associated with survival, with neurologically intact survival falling to below 10% after 15–20 min of CPR.³ The ‘low-flow’ state struggles to generate coronary perfusion pressures required for return of spontaneous circulation (ROSC)⁴ but also leads to organ hypoperfusion and ischaemia. One organ especially sensitive to ischaemia is the brain with the risk of a hypoxic brain injury increasing with longer periods of CCPR.⁵ The justification for ECPR is that by providing pressurised and oxygenated blood flow during a cardiac arrest it may prevent organ damage until a reversible feature of the cardiac arrest can be identified and managed. Although there are no randomised controlled trials comparing the effectiveness of ECPR to CCPR, there are case series that suggest that ECPR may confer significant survival and neurologically intact survival benefits over CCPR in selected patient populations.^2,6–8 Consideration of ECPR, for selected patients in refractory cardiac arrest, is now recommended by several large organisations, including the Extracorporeal Life Support Organization (ELSO),⁹ the American Heart Association (AHA)¹⁰ and the UK Resuscitation Council Advanced Life Support guidelines.¹¹

Survival rates to discharge from in-hospital cardiac arrests are approximately 18.4% in the UK.¹² This means more than 80% of patients that have a cardiac arrest in-hospital never go home. ECPR is one intervention that may confer a survival benefit in carefully selected patients in centres that have the facilities to provide extracorporeal mechanical circulatory support.

To assess the feasibility of setting up an inpatient ECPR service we identified the criteria that would be used to select suitable patients, before examining the cardiac arrest database to determine how many patients might fulfil this ECPR criteria in a year. Our hospital is a large specialist cardiovascular centre with primary Percutaneous Coronary Intervention (PCI) and Heart Attack Centre (HAC) servicing North East London with approximately 2100 emergency HAC admissions a year.¹³ There are 227 inpatient beds in the hospital facilitating approximately 50 cardiac surgery cases and 225 cardiology procedures a week. There are 60 critical care beds with an established extracorporeal membrane oxygenation (ECMO) service with more than 40 patients treated with extracorporeal support in 2017–2018. The hospital also provides oncology, endocrine and respiratory services.

Methods

A literature and internet search was conducted to identify and compare various eligibility criteria for ECPR in in-hospital cardiac arrest in other centres and around the world. Through this search several categories were identified: age, witnessed arrest, ‘no-flow’ time, ‘low-flow’ time, arresting rhythm and end tidal carbon dioxide. We then researched the evidence base for these individual categories to identify best-evidence ranges and criteria. These were then combined with established contra-indications – known severe aortic regurgitation, lack of potential for reversibility and severe co-morbidities or active Do Not Attempt Resuscitation (DNAR) order.

The combined ECPR eligibility criteria was circulated to all ECMO consultants (defined as members of the 24-h Barts ECMO consultant rota) for review and was then circulated to cardiology, perfusion and critical care nursing for comment. The final ECPR selection criteria are shown in Table 1.

Table 1.

ECPR selection criteria.

1	<65 yearsa
2	Witnessed cardiac arrest
3	<1 min from arrest to chest compressions
4	ECMO flows possible <60 min from arrest
5	Electrical activity on ECG
6	Persistent EtCO2 > 1.3 kPa during resuscitation or signs of life during resuscitation
7	No known severe aortic regurgitation
8	Likely cardiac cause with reversibility
9	Refractory arrest at 15 min
10	No severe co-morbiditiesb or active DNAR

DNAR: Do Not Attempt Resuscitation; ECMO: extracorporeal membrane oxygenation.

^aAge over 65 a relative contra-indication depending on other patient factors.

^bCo-morbidities that may be significant include peripheral vascular disease, preceding frailty/poor performance, contra-indications to systemic anticoagulation, multi-organ failure and other life-limiting diseases.

In order to assess the potential need for an ECPR service, the ‘Cardiac Arrest Call’ database at St Bartholomew’s Hospital was retrospectively examined between September 2016 and September 2017 inclusive to identify potential ECPR candidates. This database includes all calls made to activate the Cardiac Arrest Team; data are prospectively collected by Resuscitation Officers and Audit Nurses. Patients fulfilling the agreed selection criteria for ECPR were extracted. The following data were collected: patient demographics, the time of arrest (in hours (08.00 am–17.00 pm) vs. out of hours), location of arrest, duration of arrest (<15 min or >15 min), initial rhythm, presumed cause (cardiac/non-cardiac), ‘low-flow’ time, witnessed or not, and medical co-morbidities.

Results

Epidemiology of cardiac arrests

Over this 12-month period, there were 261 cardiac arrests with an average age of 66.6 years old. Of these, 84 were female and 177 were male. The most common ethnicities were White British (63%), Asian or Asian British (22%), White Other (9%) and Black Caribbean (4%). Only 25% of cardiac arrests occurred in-hours Monday–Friday. The most common initial rhythm in patients identified as candidates for ECPR was PEA (12/21, 57%) followed by VT (5/21, 24%) and VF (4/21, 19%).

Inclusion criteria for ECPR candidates

In order to find potential suitable candidates for ECPR, we applied our agreed selection criteria, initially removing all patients over the age of 65 from the database (Figure 1). This left us with 109 patients with an average age of 53 – the majority of patients were male (73%) with 39% of patients in a shockable rhythm. Half of these arrests occurred in either a level 2 or 3 ward area (55/109) and a further quarter happened in the angiography suite. Direct ‘HAC’ admissions were included if they experienced an in-hospital cardiac arrest, but we did not include patients who arrived in cardiac arrest.

Figure 1.

Retrospective selection of ECPR candidates.

Of these 109 patients aged 65 or younger, the average downtime was 13 min with 31% of patients found to be in refractory cardiac arrest at 15 min. ROSC was achieved in 67% of cases with conventional CPR. Due to specialist services available, further two patients were put on VA ECMO and two patients on cardiopulmonary bypass (and are included in the analysis if they met the selection criteria).

After excluding cardiac arrests that were not refractory at 15 min, we were left with 28 patients. A further four patients were excluded for having no electrical activity on ECG during cardiac arrest (i.e. asystole) and a further three patients were excluded for unwitnessed arrest and unknown ‘arrest to chest compressions <1 minute criteria’. This left 21 (21/261, 8%) patients that would meet our initial ECPR selection criteria – the majority (16/21) occurred outside of normal working hours (8 am–5 pm).

Although co-morbidity information was available for each patient and originally tabulated in our results, no patients were excluded from consideration after the listed co-morbidities were reviewed by an ECMO specialist consultant. Patient information is normally available immediately through electronic documentation, and where it is not (i.e. acute primary PCI admission) an admission clerking or handover is used.

Of patients that met ECPR criteria, survival to discharge was 19% (4/21) with three out of the four survivors classified as good neurological outcome, discharged from hospital with CPC scores of 1–2. Four post-operative cardiothoracic surgery patients received cardiac surgery advanced life support protocol resuscitation, all with ROSC. Post-ROSC, two of these patients went onto cardiopulmonary bypass and one required VA ECMO for cardiogenic shock. Three further patients were placed on VA ECMO, one post-ROSC and two during CPR. ECPR in these two patients was ad hoc secondary to an ECMO consultant being present, as there was no formal ECPR service at this time. These three patients did not survive to discharge.

Discussion

Outcomes of cardiac arrest with conventional CPR are poor with observational studies estimating that in-hospital cardiac arrest carries a survival to discharge rate of 15–18%.^10,12 In our analysis, we identified 21 potential candidates over a 12-month period that met our initial criteria for consideration of ECPR. Our outcomes within this cohort are in line with the literature, with three out of the four survivors classified as good neurological outcome, discharged from hospital with CPC scores of 1–2.

Prognostic patient factors

In order to select potential candidates for ECPR, we used criteria which can be assessed during the early stages cardiac arrest. It is difficult to draw definitive conclusions from studies that select ECPR candidates, as they are predominantly retrospective with small cohorts and variable selection criteria. In a recent systematic review, the only prognostic factors associated with improved outcome in ECPR were the presence of electrical cardiac rhythm and shorter low-flow duration.⁷

Age

A retrospective analysis of 111 ECPR patients in the Emergency Department found that on average survivors were younger patients (<56 years old) with a shorter duration of low-flow time (mean CPR duration 51 vs. 61 min).¹⁴ This study found an optimum cut-off age of 56 years old; however, other studies have found the optimum to be higher at <66.¹⁵ In large studies of CCPR, there is a progressive decline in both survival and neurological outcome after 64 years.¹⁶

A recent retrospective study on neurological outcome after ECPR looked at the interplay between age and low-flow time.¹⁷ Four hundred and eighty-two patients were split into three age categories – group 1 (<65 years old), group 2 (65–75 years old) and group 3 (>75 years old). Low-flow time was comparable between the groups. Results showed that neurological outcomes were non-significantly poorer in the >75 years old group. Sub-group analysis showed that poor outcomes increase with age, even with similar low-flow times. This study concludes that ECPR may be indicated in older patients with short low-flow times, or younger patients with long low-flow times. No patients older than 75 years old with low-flow time of >60 min survived.

Although these data suggest poorer outcome in older patients, there is a lack of high-quality evidence supporting a definitive age cut-off, and NICE guidelines state that treatment cannot be restricted based on age alone. The average age of patients we identified as candidates for ECPR was 51.7, compared to an average age of 66.6 years for all arrests at our centre during the study period.

Witnessed arrest, no-flow time and low-flow time

Although intuitive, there is little evidence for witnessed arrest and no-flow time as prognostic indicators in ECPR. This is probably due to the lack of randomised and large outcome studies. They are both recognised prognostic risk factors in conventional CCPR¹⁸ and therefore seem reasonable to include when trying to select patients with the lowest risk of hypoxic brain injury prior to ECPR commencing. Low-flow time is a recognised prognostic indicator for outcome in ECPR.⁷ This makes physiological sense as the longer the period of chest compressions and low flow, the higher the time of organ hypoperfusion and potential ischaemia. Some studies have suggested a significant fall in survival if ECPR is commenced more than 60 min after cardiac arrest¹⁹ and this is a criterion in other centres.²⁰ The average low-flow time of patients we identified as candidates for ECPR was 35.1 min, and 19/21 patients had a documented witnessed arrest, although for 2/21 patients this was not recorded.

Cardiac origin and initial rhythm

The cause of cardiac arrest and initial rhythm have both been shown to be useful predictors of survival to discharge. A three-year prospective observational study of over 160 eligible patients² demonstrated ECPR was associated with survival benefit versus conventional CPR for in-hospital cardiac arrest (IHCA) of cardiac origin. The use of ECPR allows a bridge to further interventional procedures such as revascularisation of the coronary arteries or medium term mechanical circulatory support. Shockable rhythms are associated with better outcomes⁷ and may indicate a cardiac arrest of cardiac aetiology and a relatively short low-flow time, both of which increase the probability of a positive outcome.²

Co-morbidities and MDT approach

A protocolised approach, with broad input from specialists across the multi-disciplinary team (MDT), can facilitate the necessary careful patient selection based on extent of co-morbidities. Guidance for writing safe and efficient ECPR protocols has been published based on expert experience.²¹ The CHEER study was a multicentre observational study wherein patient selection and subsequent treatment of both in-hospital and out-of-hospital ECPR were protocolised.²² Outcomes demonstrated 54% of patients surviving to discharge with complete neurological recovery, although recently published data suggest outcomes may be worse, with the percentage of neurologically intact survivors closer to 30%.^23,24 At our centre, a ‘Shock Call’ protocol has been created and implemented, allowing anyone to identify a patient who may be suitable for ECMO/ECPR, and rapidly convene via real-time conference call a highly skilled MDT comprising senior specialists across Intensive Care, anaesthetics, interventional and heart failure cardiology, cardiothoracic surgery, resuscitation and specialist nursing. In the original observational study published by Chen et al.,² the decision to call ECMO specialists was made ‘by the attending doctors in charge’. However, we would propose that a senior member of the ECPR team attends all cardiac arrests, saving time to decision and reducing cannulation time in candidates in which ECPR is appropriate.

Identifying candidates for ECPR from our database

Almost half of all cardiac arrests occur on level 2 or 3 dependency ward areas with another 25% occurring in the cardiac angiography suites. This demonstrates that the majority of cardiac arrests are in highly monitored patients in predictable locations with immediate access to advanced life support trained staff and resuscitation equipment. This makes it easier to target an ECPR service to these locations. It is also less of a logistic challenge to provide immediate bedside cannulation and ECMO initiation in critical care units and angiography suites as compared to wards due to the availability of extra personnel, equipment and imaging.

Using our selection criteria, we identified 21 potential ECPR candidates; the majority (16/21) occurred outside of normal working hours (8 am–5 pm). In order to achieve the best chance of neurologically intact survival, the ECPR team should be able to initiate ECPR within a 60-min target. Although there appear to be enough potential ECPR candidates to justify an in-hospital ECPR service, the majority of these cardiac arrests occurred out-of-hours and would require a 24-h resident on-site ECMO team. Such a service would likely be most useful and successful in large specialist cardiac centres, with a very high ratio of monitored and higher care level beds, a patient population more likely to experience IHCA, around the clock access to perfusion services and senior staff with relevant experience.

Conclusion

ECPR is a salvage technique for patients in refractory cardiac arrest recommended for consideration in selected patients by the ELSO, the AHA and the UK Resuscitation Council Advanced Life Support guidelines. We have shown that in specialist cardiac centres it is likely there is a cohort of patients that can benefit from ECPR, in our case 1.75 candidates per month. Selection criteria for ECPR exist but are based on poor quality evidence from retrospective trials with small cohorts and variable criteria. Systematic review shows that rhythm and low-flow duration are the best predictors of survival in ECPR. In our analysis, cardiac arrests occurred predominantly in high dependency areas and the cardiac catheterisation lab, meaning ECPR resources could be located in these areas. The majority of cardiac arrests occurred after 17.00 pm or on a weekend, which shows an ECPR service would benefit from 24-h and seven day coverage and therefore resident on-calls for ECMO consultants. The resource implications of this intervention, along with the lack of any randomised control trial outcome data for in-hospital ECPR to enable prediction of absolute benefit, mean that further research, including health economic data and service evaluation of existing ECPR programmes, is required before a general recommendation for the provision of in-hospital ECPR services in tertiary heart attack and cardio-thoracic centres can be made.

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.