Interventional procedure overview of VA ECMO for extracorporeal cardiopulmonary resuscitation (ECPR) in adults in refractory cardiac arrest
Closed for comments This consultation ended on at Request commenting lead permission
Evidence summary
Population and studies description
This interventional procedure overview is focused on VA ECMO use in ECPR for in hospital cardiac arrest (IHCA) and out of hospital cardiac arrest (OHCA). Two additional overviews have been developed focusing on VA ECMO in acute heart failure and postcardiotomy cardiogenic shock.
This overview is based on approximately 129,000 people from 8 systematic reviews (Low 2023, Low 2024, Zhong 2024, Gomes 2023, Cheema 2023, Kiyohara 2023, Scquizzato 2023, Pagura 2024), 1 long-term RCT follow-up study (Rob 2024), 1 retrospective registry study (Inoue 2022), and 1 single-centre retrospective PSM study (Shih 2024). This a rapid review of the literature, and a flow chart of the complete selection process is shown in figure 1. This overview presents 11 studies as the key evidence in table 2 and table 3, and lists 31 other relevant studies in appendix B, table 5.
All 8 systematic reviews (Low 2023, Low 2024, Zhong 2024, Gomes 2023, Cheema 2023, Kiyohara 2023, Scquizzato 2023, Pagura 2024), included the same 3 RCTs; ARREST (USA, 2020), Prague OHCA (Czech Republic, 2022), and INCEPTION (Netherlands, 2023). Of these systematic reviews, 3 included a fourth pilot RCT: EROCA (USA 2021) (Cheema 2023, Kiyohara 2023, Scquizzato 2023).
OHCA and IHCA studies
The selected key evidence includes 3 publications reporting on 2 systematic reviews and meta-analyses comparing ECPR and CCPR in people with IHCA and people with OHCA from studies done in Asia, Europe and North America (Low 2023, Low 2024, Zhong 2024). All other key evidence included people with OHCA only.
The systematic review and meta-analyses by Low (2023), included 11 studies (3 RCTs and 8 PSMs) reporting on adults with OHCA or IHCA. This was updated in the publication by Low (2024) to include a further 2 PSM studies. Overall, both Low et al. publications report on 14,048 adults, of whom 6,336 had ECPR. Of the studies included in the review, 8 (3 RCTs and 5 PSMs) included people with OHCA, 4 PSM studies included people with IHCA, and 1 PSM included people with OHCA and IHCA. All RCTs were noted to be either at 'low risk' or 'some concerns' for bias, while all PSMs were noted to be of 'high quality'. Across the included studies, the mean age ranged from 55 to 73 years and 77% were male. The most common cause of cardiac arrest was acute coronary syndrome (33 to 77%), and of those treated with ECPR, 57% had an initial shockable rhythm compared to 60% in the CCPR group.
The systematic review and meta-analyses by Zhong (2024) included 17 studies (3 RCTs, 5 prospective studies, and 9 retrospective studies) reporting on 167,728 people with OHCA or IHCA. Of the 17 included primary studies, 11 were also included in the Low systematic reviews, including the same 3 RCTs. Of the studies included by Zhong (2024), 10 included people with OHCA, 4 included people with IHCA, and 3 included people with OHCA and IHCA. Using the Cochrane Collaboration's risk assessment tool, no RCT study was considered "high risk of bias" in each domain. Using the Newcastle Ottawa Scale, most of the 10 cohort studies were considered medium quality. Across the included studies, the mean age ranged from 50 to 75 years. People who had ECPR were considerably younger and more likely to have an initial shockable rhythm compared to those in the CCPR group.
Both systematic reviews reported on short-term (30 days from CA) and long-term (90 or more days from CA) outcomes, as well as survival up to 1 year.
OHCA studies
The systematic review and meta-analyses by Gomes (2023) included 3 RCTs comparing ECPR and CCPR in 418 people with OHCA. In the risk of bias assessment of the 3 RCTs, 2 were considered to have 'some concerns' due to crossover between the study groups. However, this potential bias was minimised by balanced deviations from intended interventions between study groups and by blinded outcomes assessment. The mean age across the ECPR groups ranged from 54 to 59 years compared to 57 to 58 in the CCPR groups and proportion of males was 82 to 93% and 73 to 89%, respectively. The primary cause of cardiac arrest was acute myocardial infarction in 50% of people in the Prague OHCA trial, and 77% in the INCEPTION trial. In 2 of the included RCTs (ARREST and INCEPTION) most people enrolled had an initial shockable rhythm (98 to 100%), whereas the Prague OHCA trial recruited a high proportion with non-shockable rhythm (mean 61% both arms), a subset known to have worse outcomes.
The systematic reviews and meta-analyses by Cheema (2023), Kiyohara (2023) and Scquizzato (2023) all included the same 3 RCTs and an additional pilot trial (EROCA). The number of people included in each systematic review ranged from 433 to 435 depending on whether they included the intention to treat, per protocol or as-treated analysis set. The mean age across the ECPR groups ranged from 54 to 62 years, and 57 to 61 years in the CCPR group. The proportion of males ranged from 67 to 93% in the ECPR groups and 67 to 89% in the CCPR groups. Each systematic review reported slightly different meta-analysis results as they each used different statistical methods, pooled outcome endpoints differently, and did different subgroup analyses. All 3 systematic reviews conducted risk of bias assessments using the Cochrane Risk of Bias (RoB2) tool. Kiyohara (2023) considered all studies to be low risk of bias, Cheema (2023) considered 3 studies to have a low risk of bias and 1 (EROCA) to be high risk of bias. Scquizzato (2023) considered all studies to be intermediate risk of bias due to the absence of blinding of the treating team in all studies and unblinded assessors of neurological outcome in one study. All 3 systematic reviews reported on outcomes at hospital discharge or 30 days, and at 3 or 6 months.
The systematic review and meta-analyses by Pagura (2024) included 18 studies (3 RCTs and 15 observational studies [6 with PSM design]) reporting on 21,877 people with refractory OHCA. The mean age across the ECPR groups ranged from 46 to 66 years, and 54 to 77 years in the CCPR group. The proportion of males ranged from 70 to 93% in the ECPR groups and 62 to 89% in the CCPR groups. The mean proportion of people in the ECPR groups with initial shockable rhythm was 61% and 60% in the CCPR group. Outcomes were reported up to 6 months.
In the long-term follow-up study of the Prague OHCA RCT, 255 people who survived following the initial trial period were followed for a median of 5.3 years (3.8 to 7.2) after initial cardiac arrest. The median age of people at recruitment was 58 years and 82% were male. More people in the CCPR arm had an initial shockable rhythm (VF) 64%, compared to the ECPR arm 58%. The authors noted that there was no formal power analysis for the long-term follow-up outcomes (Rob 2024).
The retrospective SAVE-J II multicentre registry study by Inoue (2022) included 1,644 adults with OHCA who had ECPR in Japan. The median age was 60 years and 85% of the population were male. Initial shockable rhythm was reported in 69% of people and 59% had a primary cause of CA of acute coronary syndrome. Outcomes were reported until hospital discharge.
The single centre retrospective PSM study by Shih (2024) reported on 1,193 people with OHCA resuscitated with ECPR or CCPR in a high volume emergency department in Taiwan. The PSM cohort included 231 people, of whom 77 had ECPR. The median age was 57 years, 81% were male and the proportion of people with an initial shockable rhythm was 86% and 88% in the ECPR and CCPR arms, respectively. Outcomes were reported until hospital discharge.
Table 2 presents study details.
Study no. | First author, date country | Characteristics of people in the study (as reported by the study) | Study design | Inclusion criteria | Intervention | Follow up |
|---|---|---|---|---|---|---|
1 | Low, 2023 Low, 2024 Europe, Asia, North America | Low, 2023 n=9,192 (4,595 ECPR) Mean age (years):
Males:
CA cause:
Shockable rhythm
Low, 2024 n=14,048 (6,336 ECPR) Mean age ranged from 55 to 73 years Males: not reported | Low, 2023 Systematic review and meta-analysis of 11 studies: 3 RCTs and 8 PSMs Search date: April 2023 Low, 2024 Systematic review and meta-analysis of 13 studies with 14 pairwise comparisons: 3 RCTs and 10 PSMs Search date: November 2023 | Low, 2023 RCTs and PSMs comparing ECPR with CCPR in adults (aged 18 years and over) with OHCA and IHCA. CA location:
Low, 2024 CA location:
|
Mean low flow time
| Low, 2024 In-hospital, 30-day post-discharge, 3mo, 6mo, 1-year Low, 2024 In-hospital 30-day post-discharge |
2 | Zhong, 2024 Czech Republic, Belgium, France, Germany, Japan, Korea, Netherlands, Taiwan, USA | n=167,728 (2,308 ECPR) Mean age (years, range):
Males: not reported Shockable rhythm
| Systematic review and meta-analysis of 17 studies: 3 RCTs, 5 prospective studies, 9 retrospective studies. Search date: July 2023 | Studies including people over 16 years with IHCA or OHCA comparing CCPR and ECPR. CA location:
|
| |
3 | Gomes, 2023 Czech Republic, Netherlands, USA | n=418 (208 ECPR) Mean age (years):
Male (%)
CA cause:
Shockable rhythm
| Systematic review and meta-analysis of 3 RCTs (ARREST, Prague OHCA, INCEPTION) Search date: March 2023 All RCTs were open label. | RCTs which compared ECPR with standard CPR for OHCA. |
Time from arrest to ECPR
| In-hospital, 6 months |
4 | Cheema, 2023 Czech Republic, Netherlands, USA | n=434 (220 ECPR) Mean age:
Male (%):
| Systematic review and meta-analysis of 4 RCTs (ARREST, Prague OHCA, INCEPTION, EROCA) Search date: March 2023 All RCTs were open label. | RCTs comparing ECPR with conventional CPR in people with OHCA. |
| In-hospital or 30 days 3 or 6 months |
5 | Kiyohara, 2023 Czech Republic, Netherlands, USA | n=435 (221 ECPR) Mean age (years):
Male (%):
| Systematic review and meta-analysis of 4 RCTs (ARREST, Prague OHCA, INCEPTION, EROCA) Search date: February 2023 All RCTs were open label. | RCTs comparing the clinical outcomes of ECPR and CCPR for people with OHCA |
Time from arrest to ECPR
| In-hospital or 30 days 6 months |
6 | Scquizzato, 2023 Czech Republic, Netherlands, USA | n=433 (220 ECPR) Mean age:
Male:
| Systematic review and meta-analysis of 4 RCTs (ARREST, Prague OHCA, INCEPTION, EROCA) Search date: February 2023 All RCTs were open label. | RCTs enrolling adults with refractory OHCA randomised to have ECPR or CCPR alone |
| In-hospital or 30 days Longest follow-up available |
7 | Pagura, 2024 Asia, Europe, US | n=21,877 (3,129 ECPR) Mean age (years):
Male (%):
CA cause ACS:
Shockable rhythm
| Systematic review and meta-analysis of 18 studies: 3 RCTs (ARREST, Prague OHCA, INCEPTION) and 15 observational studies (6 with PSM design). Search date: April 2023 | Observational and RCTs, comparing the effect of ECPR and CCPR in refractory OHCA. |
| In-hospital or 30 days 6 months |
8 | Rob, 2024 Czech Republic | n=255 (123 ECPR) Median age at randomisation (years):
Male (%):
Shockable rhythm (VF)
| Long-term follow-up of the Prague OHCA RCT. The original RCT was a single-centre, prospective, open label RCT. Randomisation was done on-scene (pre-hospital) by web-based system. Functional assessments during follow-up were done by evaluators who were blinded to group allocation. | Adults aged 18 to 65 years with witnessed OHCA of presumed cardiac aetiology, who had received a minimum of 5 minutes of advanced cardiac life support without ROSC. Patients with unwitnessed, non-cardiac cause CA, and had suspected or confirmed pregnancy, had ROSC within 5 minutes, had obvious life-limiting comorbidities, bleeding diathesis, DNR order or prearrest CPC ≥3 were excluded. |
Comparator: CCPR 20/256 patients (7.8%) were crossed over (11 crossovers from the CCPR to the ECPR group and 9 from the ECPR group to the CCPR group) | Median 5.3 years (IQR 3.8 to 7.2) from CA. |
9 | Inoue, 2022 Japan | n=1,644 Median age (years): 60 (18 to 93) Male (%): 84.6% Initial shockable rhythm: 69.4% CA cause ACS: 59% | Retrospective SAVE-J II multicentre registry study Search date: 2013 to 2018 | Adults with OHCA who had ECPR. |
| In-hospital |
10 | Shih, 2024 Taiwan | n=1,193 (85 ECPR) PSM cohort=231 (77 ECPR) Median age (years):
Male (%):
Initial shockable rhythm:
| Single centre retrospective PSM study (Taiwan) Search date: 2016 to 2021 | Adults over 20 years old with refractory OHCA resuscitated in the emergency department (ED). Excluded people with non-cardiac causes for arrest, and those who achieved sustained ROSC within 15 mins at the ED. |
Estimated low flow time <100 min. | In-hospital or 30 days |
Procedure technique
Of the 11 studies, none detailed the ECMO device or combination of devices used. Only a few of the studies detailed ECMO procedures. ECPR was initiated at the hospital (either catheterisation laboratory or emergency department) in all studies reporting on the use of ECPR for OHCA (Gomes 2023, Cheema 2023, Kiyohara 2023, Scquizzato 2023, Pagura 2024, Inoue 2022, Shih 2024). Six OHCA studies included in the systematic reviews by Low et al. also reported ECPR initiation in hospital, but one did not specify the location of ECPR initiation. One systematic review detailed that targeted temperature management after cardiac arrest was used in 45% of people having ECPR and 15% of patients having CCPR (Low 2023).
Efficacy
Short-term survival with favourable neurological outcomes
Short-term survival with favourable neurological outcome was reported in 10 out of 11 studies included in the key evidence. Most often favourable neurological outcome was defined as CPC 1 or 2 on the cerebral performance category. One systematic review also included studies using the MGOS (Zhong 2024), and 2 systematic reviews included studies using the mRS (Scquizzato 2023, Pagura 2024). Mostly, short-term was defined as hospital discharge or 30 days after initial CA.
OHCA and IHCA studies
Three publications reporting on 2 systematic reviews presented meta-analyses for short-term survival with favourable neurological outcomes in both OHCA and IHCA (Low 2023, Low 2024, Zhong 2024). Both systematic reviews defined short-term as to hospital discharge or 30 days after initial CA.
In the systematic review of 11 studies reporting on adults with OHCA or IHCA, ECPR was associated with improved short-term survival with favourable neurological outcomes compared to CCPR in a meta-analysis of 7 studies (OR 1.65, 95% CI: 1.02 to 2.68; p=0.042) (Low 2023). The updated analysis of 10 studies also found ECPR was associated with favourable neurological outcomes at short-term follow-up (OR 1.57, 95% CI 1.14 to 2.15) (Low 2024). A meta-analysis of 11 studies in people with OHCA and IHCA (including 1,442 patients who had ECPR and 51,221 who had CCPR), showed improved short-term survival with favourable neurological outcomes with ECPR compared to CCPR (RR 2.88, 95% CI: 1.96 to 4.23, I2=76%; p<0.0001; Zhong 2024). This result was consistent in a subgroup analysis using RCTs and PSM data only (RR 1.67, 95% CI: 1.16 to 2.40, I2=51%; p=0.005; Zhong, 2024).
After stratifying based on the location of cardiac arrest, subgroup meta-analysis of 4 studies, showed that statistically significantly more people having ECPR for IHCA had short-term favourable neurological outcomes compared to people having CCPR for IHCA (OR 2.37, 95% CI: 1.34 to 4.19; p=0·0031, Low 2023). Similar outcomes were also seen in a subgroup meta-analysis of 10 studies (RR 2.18, 95% CI: 1.24 to 3.81, I2=9%; Zhong 2024). In a subgroup meta-analysis of 3 studies, no significant differences in the rate of short-term favourable neurological outcome were observed in people with OHCA treated with ECPR or CCPR (OR 1.24, 95% CI: 0.65 to 2.36; p=0.51). Similar outcomes were also seen in a subgroup meta-analysis of 10 studies (RR 1.50, 95% CI: 0.98 to 2.29, I2=55%; Zhong 2024).
OHCA studies
Five systematic reviews (Gomes 2023, Cheema 2023, Kiyohara 2023, Scquizzato 2023, Pagura 2024), 1 retrospective registry study (Inoue 2022), and 1 single-centre retrospective PSM study (Shih 2024) presented short-term survival with favourable neurological outcomes in people with OHCA. Short-term was defined as hospital discharge or 30 days after initial CA (although this was labelled 'mid-term' in the Cheema (2023) systematic review).
In the systematic review of 3 RCTs including 418 people with OHCA, ECPR was associated with a non-statistically significant higher rate of survival with a favourable neurological outcome at the shortest follow-up (26%) compared with standard CPR (17%), RR 1.47 (95% CI: 0.91 to 2.40; p=0.12; Gomes 2023). A subgroup analysis of patients with an initial shockable rhythm showed similar findings to the main analysis, with a non-statistically significant benefit of ECPR (RR 1.62, 95% CI: 0.95 to 2.76; p=0.07; Gomes 2023).
Across 3 systematic reviews and meta-analyses including the same 4 RCTs reporting on people with OHCA, the rate of short-term survival with favourable neurological outcome was statistically higher in the ECPR groups compared to CCPR (Cheema 2023, Kiyohara 2023, Scquizzato 2023). Pooled survival with favourable neurological outcome was 25% for those who had ECPR compared to 16% who had CCPR (Scquizzato 2023). Using a Mantel-Haenszel, random-effects meta-analysis, the odds ratios reported were 1.82, 95% CI: 1.13 to 2.92; p=0.01, I2=0% (Scquizzato 2023) and 1.84, 95% CI: 1.14 to 2.99, I2=0%; p=0.01 (Kiyohara 2023). The risk ratio (RR) using an inverse-variance, random effects meta-analysis was 1.59, 95% CI: 1.09 to 2.33, I2=0%; p=0.02 (Cheema 2023). A subgroup analysis of patients with an initial shockable rhythm showed higher rate of short-term survival with favourable neurological outcome in the ECPR (31%) groups compared to CCPR (21%), OR 1.93, 95% CI: 1.16 to 3.23, I2=0%; p=0.012; Scquizzato 2023).
In a meta-analysis of 14 studies (3 RCTs and 11 observational studies) in people with OHCA, the rate of short-term survival with favourable neurological outcome was significantly improved with ECPR (14%) compared with CCPR (7%); OR 2.35, 95% CI: 1.61 to 3.43, I2=80%; p<0.0001. This was consistent in subgroup analysis including only RCTs OR: 1.83, 95% CI: 1.13 to 2.96, I2=0%; p=0.01 (Pagura 2024).
In the retrospective SAVE-J II multicentre registry study which included 1,644 adults with OHCA who had ECPR in Japan, the overall rate of survival with favourable neurological outcome at hospital discharge was 14% (Inoue 2022). This rate was higher in those with an initial shockable rhythm (17%), and lower in those with a non-shockable rhythm (9% PEA, 4% asystole) (Inoue 2022). Multivariable analysis showed that younger age, initial shockable rhythm at the scene, and location of cardiac arrest were significantly associated with both favourable outcome and survival to hospital discharge (p<0.01) (Inoue 2022).
In the single centre retrospective PSM study reporting on people with OHCA resuscitated with ECPR or CCPR in a high volume emergency department in Taiwan, the rate of survival with favourable neurological outcome was 18% among those who had ECPR, compared to 5% in those who had CCPR. The PSM multivariate analysis reported an adjusted odds ratio (aOR) of 13.31, 95% CI: 1.61 to 109.9, p=0.016 (Shih 2024). Higher rates of favourable neurological outcome were associated with younger age (48 versus 59 years, p=0.001), CPR duration (37 min versus 51 min, p=0.006) and collapse to ECMO flow initiation time (76 versus 98.0 min, p=0.031).
Long-term survival with favourable neurological outcomes
Long-term survival with favourable neurological outcome was reported in 8 out of 11 studies included in the key evidence. Most often favourable neurological outcome was defined as CPC 1 or 2 on the cerebral performance category. One systematic review also included studies using the MGOS (Zhong 2024), and 2 systematic reviews included studies using the mRS (Scquizzato 2023, Pagura 2024). Mostly, long-term was defined as 3 or 6 months after initial CA.
OHCA and IHCA studies
Two systematic reviews presented meta-analyses for long-term survival with favourable neurological outcomes in both OHCA and IHCA (Low 2023, Zhong 2024). Both systematic reviews defined long-term as 3 months or more after initial CA.
In the systematic review of 11 studies reporting on adults with OHCA or IHCA, ECPR was associated with statistically significantly improved long-term survival with favourable neurological outcomes compared to CCPR in a meta-analysis of 8 studies (OR 2.04, 95% CI: 1.41 to 2.94; p=0.0001) (Low 2023). A meta-analysis of 11 studies in people with OHCA and IHCA (including 896 patients who had ECPR and 1,977 who had CCPR), showed improved long-term survival with favourable neurological outcomes with ECPR compared to CCPR (RR 2.11, 95% CI 1.40 to 3.19, I2=69%; p=0.0004; Zhong 2024). This result was consistent with a subgroup analysis using RCTs and PSM data only (RR 1.83, 95% CI: 1.32 to 2.53, I2=14%; p=0.0003; Zhong, 2024).
After stratifying based on the location of cardiac arrest, subgroup meta-analysis of 3 studies, showed a statistically significantly higher rate of long-term favourable neurological outcomes in ECPR groups compared to CCPR in people with IHCA (OR 2.80, 95% CI: 1.31 to 6.00; p=0.008), and in people with OHCA (4 studies: OR 1.96, 95% CI: 1.02 to 3.79; p=0.045; Low 2023). Similar outcomes were also seen in a subgroup meta-analysis of 3 studies in people with IHCA (RR 2.17, 95% CI: 1.19 to 3.94, I2=0%), and of 5 studies in people with OHCA (RR 1.95, 95% CI: 1.01 to 3.59, I2=11%; Zhong 2024).
OHCA studies
Five systematic reviews (Gomes 2023, Cheema 2023, Kiyohara 2023, Scquizzato 2023, Pagura 2024) and 1 long-term RCT follow-up study (Rob 2024), presented long-term survival with favourable neurological outcomes in people with OHCA. Long-term was defined as 3 or 6 months in 2 systematic reviews (Cheema 2023, Scquizzato 2023), or 6 months in 3 systematic reviews (Gomes 2023, Kiyohara 2023, Pagura 2024). Median follow-up was 5.3 years after initial CA in the RCT follow-up study (Rob 2024).
In the systematic review of 3 RCTs including 418 people with OHCA, ECPR was associated with a non-statistically significant higher rate of survival with a favourable neurological outcome at 6 months (28%) compared with standard CPR (19%), RR 1.48, 95% CI: 0.88 to 2.49, I2=28%; p=0.14 (Gomes 2023). A subgroup analysis of patients with an initial shockable rhythm showed similar findings to the main analysis, with a non-significant benefit of ECPR (RR 1.50, 95% CI 0.90 to 2.50; p=0.12; Gomes 2023).
In 2 systematic reviews and meta-analyses including the same 4 RCTs reporting on people with OHCA, unlike short-term survival with favourable neurological outcomes, the rate of long-term survival with favourable neurological outcomes was not statistically significantly higher in the ECPR groups compared to CCPR (Cheema 2023, Kiyohara 2023). In a systematic review using a Mantel-Haenszel, random-effects meta-analysis, the OR was 1.74, 95% CI: 0.86 to 3.51, I2=35%; p=0.12 (Kiyohara 2023) and in another, the risk ratio (RR) using an inverse-variance, random effects meta-analysis was 1.47, 95% CI: 0.89 to 2.43, I2=25%; p=0.13 (Cheema 2023). However, in a systematic review and meta-analysis of the same 4 RCTs, pooled survival with favourable neurological outcome was 27% for those treated with ECPR compared to 18% on CCPR (OR 1.72, 95% CI: 1.09 to 2.70, I2=26%; p=0.02; Scquizzato 2023). Although trial sequential analysis confirmed the statistically significant beneficial effect of ECPR, the sample size included in the meta-analysis did not reach the required information size (n=520; Scquizzato 2023). The systematic review by Scquizzato (2023) also reported no difference in patients surviving with poor neurological outcomes at the longest follow-up available (0% ECPR, 2% CCPR [OR 0.24, 95% CI: 0.05 to 1.26, I2=0%; p=0.780]).
Two systematic reviews considered long-term survival with favourable neurological outcomes in subgroup meta-analyses in patients with an initial shockable rhythm. The rate of long-term survival with favourable neurological outcomes was 34% in the ECPR group compared with 23% in the CCPR group, OR 1.90, 95% CI: 1.16 to 3.13, I2=23%; p=0.011 (Scquizzato 2023), and OR 1.91, 95% CI: 0.90 to 4.03, I2=36%; p=0.09 (Kiyohara 2023).
One systematic review did a subgroup meta-analysis of studies by design (single-centre or multi-centre). A statistically significant difference in long-term survival with favourable neurological outcome with ECPR compared to CCPR was confirmed among single-centre studies (30% ECPR compared to 19% CCPR; OR 1.88, 95% CI: 1.11 to 3.19, I2=48%; p=0.02; Scquizzato 2023).
In a meta-analysis of 6 studies (including RCTs and observational studies) in people with OHCA, the rate of long-term survival with favourable neurological outcome was statistically significantly improved with ECPR (16%) compared with CCPR (8%); 2.72, 95% CI: 1.47 to 5.04, I2=47%; p=0.002; Pagura 2024).
In the long-term follow-up study of the Prague OHCA RCT, 255 people who survived following the initial trial period were followed for a median of 5.3 years (3.8 to 7.2) after initial cardiac arrest. In the ITT analysis, the rate of survival with favourable neurological outcome was 27% in those with ECPR and 19% in those with CCPR (RR 0.90, 95% CI: 0.79 to 1.03, p=0.13). In the per protocol and as-treated analysis, RR were 0.86, 95% CI: 0.75 to 0.99, p=0.035 and RR 0.84, 95% CI: 0.73 to 0.96, p=0.007, respectively (Rob 2024). Only 1 person in each trial arm in the ITT analysis was reported as surviving with a poor neurological outcome (Rob 2024).
Short-term survival
OHCA and IHCA studies
Three publications reporting on 2 systematic reviews presented meta-analyses for short-term survival in both OHCA and IHCA (Low 2023, Low 2024, Zhong 2024). Short-term is defined as survival to discharge or 30 days after initial CA.
In the systematic review of 11 studies reporting on adults with OHCA or IHCA, ECPR was associated with improved 30-day survival compared to CCPR in a meta-analysis of 7 studies (OR 1.45, 95% CI: 1.08 to 1.96; p=0.015) (Low 2023). In the updated analysis of 9 studies (1 RCT and 8 PSM studies), 30-day survival rate was 24% in the ECPR group and 18% in the CCPR group (OR 1.70, 95% CI: 1.29 to 2.26, I2=71%; Low 2024). A meta-analysis of 3 RCTs and 9 PSM studies in people with OHCA and IHCA, showed improved short-term survival with ECPR compared to CCPR (RR 1.25, 95% CI 1.00 to 1.56, I2=57%; p=0.05; Zhong 2024).
After stratifying based on the location of cardiac arrest, subgroup meta-analysis of 3 studies showed higher rates of survival at discharge or 1 month in ECPR groups compared to CCPR for people with IHCA (RR 2.03, 95% CI: 1.30 to 3.18, I2=0%), than for people with OHCA (RR 1.10, 95% CI: 0.91 to 1.34, I2=44%; Zhong 2024).
OHCA studies
Three systematic reviews (Cheema 2023, Scquizzato 2023, Pagura 2024), 1 retrospective registry study (Inoue 2022), and 1 single-centre retrospective PSM study (Shih 2024) presented short-term survival outcomes in people with OHCA. Short-term was defined as hospital discharge or 30 days after initial CA.
In the 2 systematic reviews and meta-analyses including the same 4 RCTs reporting on people with OHCA, no statistically significant difference in short-term survival was observed in people who had ECPR compared to CCPR. In the inverse variance random effects meta-analysis of the 4 RCTs the RR was 1.21, 95% CI: 0.64 to 2.28, I2=48%; p=0.55 (Cheema 2023). In the Mantel-Haenszel random effects meta-analysis, short term survival was 33% in the ECPR group and 27% in the CCPR group (OR 1.35, 95% CI: 0.55 to 3.29, I2=0%; p=0.514; Scquizzato 2023).
In the systematic review of RCTs and observational studies, meta-analysis of 13 studies showed a higher rate of short-term survival with ECPR (20%) than with CCPR (10%) for people with OHCA (OR 1.71, 95% CI: 1.18 to 2.46, I2=81%; p=0.004; Pagura 2024).
In the retrospective SAVE-J II multicentre registry study which included 1,644 adults with OHCA who had ECPR in Japan, the overall rate of survival to hospital discharge was 27% (Inoue 2022). This rate was higher in those with an initial shockable rhythm (32%), and lower in those with a non-shockable rhythm (19% PEA, 11% asystole). Multivariable analysis showed that shorter low flow time was significantly associated with survival to hospital discharge (p<0.001) (Inoue 2022).
In the single centre retrospective PSM study reporting on people with OHCA resuscitated with ECPR or CCPR in a high volume emergency department in Taiwan, the in-hospital or 30-day survival rate was 29% among those with ECPR, compared to 8% in those with CCPR. The PSM multivariate analysis reported an adjusted OR (aOR) of 6.02, 95% CI: 2.19 to 16.52 (Shih 2024).
Long-term survival
OHCA and IHCA studies
Two systematic reviews presented meta-analyses for long-term survival in both OHCA and IHCA (Low 2023, Zhong 2024). Long-term is defined as 3- to 6-month survival after initial CA. Both studies also report on survival up to 1 year.
In the systematic review of 11 studies reporting on adults with OHCA or IHCA, ECPR was associated with improved survival compared to CCPR at 3 months (3 studies [OR 3.98, 95% CI: 1.12 to 14.16; p=0.033]), 6 months (6 studies [OR 1.87, 95% CI: 1.36 to 2.57; p=0.0001]), and 1 year (5 studies [OR 1.72, 95% CI: 1.52 to 1.95; p<0.0001]) (Low 2023). The meta-analysis of 3 RCTs and 9 PSM studies in people with OHCA and IHCA, also showed improved long-term survival with ECPR compared to CCPR at 3 to 6 months (RR 2.73, 95% CI: 1.67 to 4.48, I2=0%; p<0.0001) and at 1 year (RR 1.92, 95% CI: 1.14 to 3.25, I2=0%; p=0.01; Zhong 2024).
After stratifying based on the location of cardiac arrest, subgroup meta-analysis of 3 studies, showed higher rates of survival to 3 to 6 months or 1 year in ECPR groups compared to CCPR for people with OHCA (RR 3.16, 95% CI: 1.36 to 7.38, I2=0%), than for people with IHCA (RR 1.92, 95% CI: 1.14 to 3.25, I2=0%; Zhong 2024).
OHCA studies
Three systematic reviews (Cheema 2023, Kiyohara 2023, Scquizzato 2023) and 1 long-term RCT follow-up study (Rob 2024), presented long-term survival in people with OHCA. Long-term was defined in 1 systematic review as 3 or 6 months (Cheema 2023), in 1 systematic review as 6 months (Kiyohara 2023), and in 1 systematic review as 30 days to 6 months. In Scquizzato (2023), survival was reported at the longest follow-up available. Median follow-up was 5.3 years after initial CA in the RCT follow-up study (Rob 2024).
In the 3 systematic reviews and meta-analyses including the same 4 RCTs reporting on people with OHCA, no statistically significant difference in long-term survival was observed in people who had ECPR compared to CCPR (Cheema 2023, Kiyohara 2023, Scquizzato 2023). In the inverse variance random effects meta-analysis of the 4 RCTs the RR for survival at 3 or 6 months was RR 1.32, 95% CI: 0.18 to 9.5, I2=64%; p=0.79 (Cheema 2023). In the Mantel-Haenszel random effects meta-analyses, survival at the longest follow-up was 28% in the ECPR group and 22% in the CCPR group (OR 1.31, 95% CI: 0.49 to 3.49, I2=58%; p=0.592; Scquizzato 2023) and the OR for survival at 6 months was 1.50, 95% CI: 0.67 to 3.36, I2=50%; p=0.33 (Kiyohara 2023).
In the long-term follow-up study of the Prague OHCA RCT, 255 people who survived following the initial trial period were followed for a median of 5.3 years after initial cardiac arrest. In the ITT analysis, Kaplan-Meier estimates of survival were 28% in those with ECPR and 20% in those with CCPR (log rank p=0.01). In the per protocol and as-treated analysis, the difference between groups were log rank p=0.008 and p<0.001, respectively (Rob 2024).
Mortality
OHCA and IHCA studies
Two publications reporting on 1 systematic review presented meta-analyses for in-hospital mortality in both OHCA and IHCA (Low 2023, Low 2024).
In the meta-analyses of 3 RCTs and 8 PSMs of adults with OHCA or IHCA (n=9,192), ECPR was associated with significant reduction in mortality (OR 0.67, 95% CI: 0.51 to 0.87, I2=42%; p=0.0034 (Low 2023). In-hospital mortality in the updated analysis of 3 RCTs and 11 PSMs of adults with OHCA or IHCA (n=14,048) was 77% in ECPR group and 83% in the CCPR group (OR 0.63, 95% CI: 0.50 to 0.79, I2=64%; Low 2024).
After stratifying based on the location of cardiac arrest, subgroup meta-analysis of 4 studies, showed a significant reduction in mortality in people having ECPR compared to CCPR in those with IHCA (OR 0.42, 95% CI: 0.25 to 0.70, I2=0%; p=0.0009; Low 2023), and in people with OHCA (3 RCTs and 6 PSMs: OR 0.67, 95% CI: 0.51 to 0.88, I2=74%; Low 2024). Pooled HRs across studies showed that longer duration of CPR was associated with increased mortality (HR per min 1.01, 95% CI 1.00 to 1.01; p=0·0001), and an initial presentation with a shockable rhythm was associated with reduced mortality (HR 0.52, 95% CI 0.32 to 0.86; p=0.011). However, it did not show an association between age and mortality (HR per year 1.02, 95% CI 0.98 to 1.06; p=0.41; Low 2023).
OHCA
One systematic review reported in-hospital mortality in people with OHCA. In the systematic review of 3 RCTs including 418 people with OHCA, the mean absolute rate of in-hospital mortality was not significantly lower in the ECPR group (71%) compared to the CCPR group (79%; RR 0.89, 95% CI: 0.74 to 1.07, I2=53%; p=0.23) (Gomes 2023).
Safety
Of the 11 included studies in the key evidence, only 4 studies reported on any safety outcomes (2 systematic reviews [Low 2023, Cheema 2023], 1 registry study [Inoue 2022] and 1 long-term RCT follow-up study [Rob 2024]).
Bleeding
One systematic review and 1 registry study reported bleeding events. In the systematic review of 11 studies including adults with OHCA or IHCA, people who had ECPR were more likely to have bleeding than those who had CCPR (meta-analysis 4 studies: OR 4.84, 95% CI: 1.91 to 12.24; p=0.0009) (Low 2023). In the retrospective SAVE-J II multicentre registry study which included 1,644 adults with OHCA who had ECPR in Japan, rates of cannulation-related bleeding were 16% (268 of 1,644), and rates of haemorrhage were 8.5% (139 of 1,644; Inoue 2022).
Adverse events
One systematic review of 4 RCTs in people with OHCA reported the comparative rates of adverse events between ECPR and CCPR. Of the 4 RCTs included in the systematic review, only 2 two studies reported the rate of adverse events. Meta-analysis indicated that ECPR was associated with higher rate of adverse events than CCPR (RR: 3.22, 95% CI: 1.18 to 8.80, I2=63%; p=0.02; Cheema 2023).
In the retrospective SAVE-J II multicentre registry study which included 1,644 adults with OHCA who had ECPR in Japan, the overall reported complication rate during ECPR was 33% (535 of 1,644). This included procedure-related complications (21%, 346 of 1,644) such as cannula malposition and unsuccessful cannulation, ECMO-related complications (3%, 50 of 1,644) and ischaemia (2%, 26 of 1,644) as well as the bleeding complications reported in the section above (Inoue 2022).
Long-term adverse events
In the long-term follow-up study of the Prague OHCA RCT, 255 people who survived following the initial trial period were followed for a median of 5.3 years after initial cardiac arrest. During the follow-up, 39 people (32%) in the ECPR group and 30 (23%) in the CCPR group were discharged from the hospital or long-term hospital facilities after the initial CA event (median time to discharge 19.5 days, IQR 12.5 to 32 days). Of these, 10% (4 of 39) patients in the ECPR group and 20% (6 of 30) in the CCPR group died during the follow-up (RR 0.51, 95% CI: 0.16 to 1.66, p=0.26). At least one rehospitalisation after discharge occurred in 77% (30 of 39) of the ECPR group and 60% (18 of 30) of the CCPR group (RR 1.28, 95% CI: 0.91 to 1.8, p=0.15). At least one cardiovascular rehospitalisation occurred in 64% (25 of 39) of people in the ECPR group and 50% (15 of 30) of those in the CCPR group (RR 1.28, 95% CI: 0.84 to 1.97, p=0.26. One person in each group reported myocardial infarction after discharge, and 1 person treated with CCPR had a stroke. Hospitalisation for heart failure occurred in 5% (2 of 39) of people in the ECPR group and 10% (3 of 30) in the CCPR group, and for ventricular arrhythmia in 3% (1 of 39) people in the ECPR group and 10% (3 of 30) in the CCPR group (Rob 2024).
Anecdotal and theoretical adverse events
Expert advice was sought from consultants who have been nominated or ratified by their professional society or royal college. They were asked if they knew of any other adverse events for this procedure that they had heard about (anecdotal), which were not reported in the literature. They were also asked if they thought there were other adverse events that might possibly occur, even if they had never happened (theoretical).
They listed the following anecdotal and theoretical adverse events:
Left ventricle overloading
Deep vein thrombosis
Arteriovenous fistula
Pseudoaneurysm
Harlequin syndrome
Haemolysis
Intracerebral haemorrhage
Major pulmonary bleed
Failure to cannulate during cardiac arrest
Malposition of the cannula
Device clotting
Air entrainment/embolus
Embolism
Oxygenator failure
Consumption coagulopathy
Acquired Von Willebrand syndrome
Systemic inflammatory response syndrome (SIRS)
Multi-organ failure including kidney, liver, and pancreas.
Sixteen professional expert questionnaires were submitted. Find full details of what the professional experts said about the procedure in the specialist advice questionnaires for this procedure.
Validity and generalisability
Most of the key papers included are systematic reviews with meta-analyses. There was a significant amount of overlap identified across the systematic reviews included in the overview; much of the available evidence identified in this review is based on the same RCTs. Evidence was mainly for adult patients resuscitated from OHCA; two systematic reviews included studies done in IHCA.
No RCTs were available for the IHCA population.
Recent systematic reviews of RCTs in OHCA have varying conclusions of the benefit of ECPR compared to CCPR. This is reflected in the conflicting conclusions of the RCTs themselves.
All 3 RCTs and 1 pilot RCT were small and had a high rate of crossovers and deviations from the intended intervention.
Authors of the systematic reviews noted that the RCTs may have insufficient power to observe significant differences between intervention arms.
There was a high heterogeneity across RCTs, such as the proportion of people with non-shockable rhythms enrolled, the definition of refractory OHCA (5 minutes or 15 minutes without ROSC), the time between cardiac arrest and ECMO initiation, whether there was a standardised emergency medical services protocol in place, whether the RCT was a single or multicentre study, the centre's level of experience with ECPR, and outcome endpoints.
In the 3 RCTs reported in 8 systematic reviews (Low 2023, Low 2024, Zhong 2024, Gomes 2023, Cheema 2023, Kiyohara 2023, Scquizzato 2023, Pagura 2024), ECPR was initiated in 64% (Prague OHCA), 66% (INCEPTION) and 86% (ARREST) of the patients randomised to the intervention arm. In the additional pilot RCT included in 3 systematic reviews (Cheema 2023, Kiyohara 2023, Scquizzato 2023) 42% randomised to the intervention arm received ECPR (Scquizzato 2023). In the ARREST trial, none of the patients randomised to the standard CPR group received ECPR, while the rate of crossover was 8% in Prague OHCA and 5% in INCEPTION (Scquizzato 2023).
Some systematic reviews included observational studies, which are at higher risk of bias, particularly in ECPR, as the decision to treat is based on clinician's evaluation of comorbidities and prognostic factors which may have strong impact on the outcome. Authors of the Shih (2024) observational study noted that there may be selection bias for ECPR as there was no protocol for who got ECPR, therefore it was likely patients with better prognosis were selected for ECPR than randomised studies.
Only 1 follow-up study for 1 RCT reported any outcomes with a follow-up longer than 6 months.
Ongoing trials
ON-SCENE Initiation of Extracorporeal CardioPulmonary Resuscitation During Refractory Out-of-Hospital Cardiac Arrest (ON-SCENE) (NCT04620070); RCT; Netherlands; n=390; completion date July 2026
Pre-hospital ECMO or Conventional Resuscitation for Refractory Cardiac Arrest (PACER) (NCT06177730); RCT; Australia and New Zealand; n=10; completion date December 2024
How are you taking part in this consultation?
You will not be able to change how you comment later.
You must be signed in to answer questions