Interventional procedure overview of venoarterial extracorporeal membrane oxygenation (VA ECMO) for acute heart failure
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 acute HF. Two additional overviews have been developed focusing on VA ECMO in post cardiotomy and as extracorporeal cardiopulmonary resuscitation (ECPR). Some of the evidence includes a mix of indications and has been presented in more than one overview.
This overview is based on approximately 32,000 people from 4 systematic reviews (Elsaeidy 2024, Sohail 2022, Alba 2021, Vishram-Nielsen 2023), 3 randomised controlled trials (Thiele 2023, Banning 2023, Ostadal 2023), 1 retrospective registry study (Olson 2020) and 1 single centre retrospective study (Cheng 2019). The 3 randomised controlled trials (Thiele 2023, Banning 2023, Ostadal 2023) were also included in the Elsaeidy 2024 systematic review. There were 15 overlaps accounting for 11,766 people in primary studies included across 3 systematic reviews (Sohail 2022, Alba 2021, Vishram-Nielsen 2023). This is a rapid review of the literature, and a flow chart of the complete selection process is shown in figure 1. This overview presents 9 studies as the key evidence in table 2 and table 3, and lists 30 other relevant studies in appendix B, table 5.
All randomised controlled trials included in the key evidence, and those included in the systematic review by Elsaeidy (2024), were conducted in Europe (Thiele 2023, Banning 2023, Ostadal 2023). The 3 systematic reviews of observational studies included in the key evidence included studies from Asia, Australia, Europe, North America and South America (Sohail 2022, Alba 2021, Vishram-Nielsen 2023). The included registry study used data from the Extracorporeal Life Support Organization (ELSO) which collates data worldwide (Olson 2020). The single centre retrospective study was conducted in the US (Cheng 2019).
Most key evidence studies included people with cardiogenic shock (CS). Of the key studies including people with CS, 2 systematic reviews and 2 randomised controlled trials specifically included people with CS complicating acute myocardial infarction (AMI) (Sohail 2022, Elsaeidy 2024, Banning 2023, Thiele 2023). 1 systematic review included people receiving VA ECMO for fulminant myocarditis (Vishram-Nielsen 2023), and 1 registry study included people receiving VA ECMO for peripartum cardiomyopathy (Olson 2020).
The systematic review by Elsaeidy (2024) included 4 randomised controlled trials including 611 people. Three of these are reported separately in the key evidence (Thiele 2023, Banning 2023, Ostadal 2023), the other study being a pilot study preceding the randomised controlled trial by Ostadal (2023). Half of the studies were deemed to have an overall low risk of bias, and 2 had some concerns of bias overall about deviations from the intended intervention and outcome measurement (Thiele 2023) and about the selection of the reported results (Ostadal 2023). The comparator in all trials was standard medical therapy, however 1 trial also allowed later cross-over of patients to VA ECMO if they continued to be haemodynamically unstable. The mean age of people included in the studies ranged from 60 to 68 years, and the proportion of males ranged from 73 to 95%. All studies reported outcomes at 30 days and 2 studies reported outcomes at 1 year follow-up.
The randomised controlled trial reported by Thiele (2023), which was also included in the Elsaeidy (2024) systematic review, compared VA ECMO to standard medical therapy alone in 417 adults with CS complicating AMI. Two thirds of people included presented with ST-segment elevation myocardial infarction (STEMI). All trial participants received early revascularisation ahead of the intervention. Intraaortic balloon pump (IABP) was permitted as an escalation therapy, and although the trial protocol forbade any cross-over, VA ECMO was initiated in 26 people in the control group (12.5%). The median age was 62 years and 81% of the population were male. Trial outcomes were reported at 30 days.
The randomised controlled trial reported by Banning (2023), which was also included in the Elsaeidy (2024) systematic review, compared early peripheral VA ECMO to standard medical therapy alone in 35 adults with CS complicating AMI. Due to the impact of the COVID-19 pandemic, the trial was stopped before completion of recruitment. The median age was 67 years and 81% of the population were male. Trial outcomes were reported at 30 days and 1 year follow-up.
The randomised controlled trial reported by Ostadal (2023), which was also included in the Elsaeidy (2024) systematic review, compared immediate VA ECMO to early conservative therapy in 117 adults with rapidly deteriorating or severe CS. The most common cause of CS in both arms was STEMI (50%) followed by decompensation of chronic heart failure (23%). The study permitted cross-over from the control group, to receive VA ECMO in the case of worsening haemodynamic stability. 39% of the control group required downstream VA ECMO therapy. The median age was 67 years and 74% of the population were male. Trial outcomes were reported at 30 days.
The systematic review by Sohail (2022) included 72 observational studies reporting on 10,276 adults who had VA ECMO for CS complicating AMI. The median concomitant IABP use across the included studies was 70%. The median age was 60 years and 78% of the population were male. Meta-analyses of the studies pooled short-term outcomes from studies with follow-ups of 7 days, 30 days and hospital discharge.
The systematic review by Alba (2021) included 306 observational studies reporting on 29,289 people with CS of any aetiology. The largest number of studies reported on people with CS after cardiac arrest (ECPR), CS complicating AMI, and postcardiotomy cardiogenic shock. Risk of bias across studies was considered low in 219 (72%), moderate in 81 (26%), and high in 6 (2%) studies. The age of people included in the studies ranged from 47 to 61 years, and 22% to 59% of the population were female. Meta-analyses of the studies pooled short-term outcomes from studies with follow-ups of 30 days and hospital discharge.
The single-centre retrospective study conducted in the US by Cheng (2019) included 149 people who survived VA ECMO (n=118) or CentriMag VAD (n=31) support as a bridge to recovery. The most common indication for ECMO intervention was postcardiotomy CS (36%), followed by allograft failure (27%), AMI (24%) and acute decompensated heart failure (ADHF) (14%). The median age was 59 years and 68% of the population were male. The median follow-up time was 306 days (IQR: 59 to 916 days).
The systematic review by Vishram-Nielsen (2023) included 54 observational studies reporting on 2,388 people with fulminant myocarditis. The median age was 41 years and 50% of the population were male. Meta-analyses of the studies pooled short-term outcomes from studies with follow-ups of 30 days and hospital discharge.
The retrospective ELSO registry study by Olson (2020) reported outcomes for people with peripartum cardiomyopathy treated with VA ECMO. The median age was 31 years and 42% were of white ethnicity. Outcomes were reported for follow-up period 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 | Elsaeidy, 2024 Belgium, Czech Republic, Germany, Latvia, Norway, Slovenia Spain, UK | n=611 Mean age ranged from 60 to 68 years Males: 80% (range 73% to 95%) Type of MI: STEMI (range 0 to 61.9%) NSTEMI (range 6.7% to 61.9%) | Systematic review and meta-analysis of 4 RCTs (Banning, 2023; Thiele, 2023; Ostadal, 2023, Lackermair, 2021) Search date: Sept 2023 All open label RCTs | RCTs that investigate the efficacy and safety of ECMO compared to standard care in managing CS-complicating AMI patients. |
| 30 days (4 studies) 1 year (2 studies) |
2 | Thiele, 2023, Germany, Slovenia ELCS-SHOCK | n=417 (ECLS n=209) Median age (years)
Male (%)
Median LVEF on admission
Two thirds of patients presented with ST-segment elevation myocardial infarction. 77.7% patients underwent CPR before randomisation. PCI was performed in 96.6% patients. | Randomised controlled trial, open label. Randomisation was done by means of a web-based system with the use of randomly changing blocks and stratification according to the trial site. | Patients aged between 18 and 80 with CS-complicating AMI and planned early revascularisation by either PCI or coronary-artery bypass grafting (CABG) CS defined as stage C, D, or E of the SCAI criteria. Excluded were people who had undergone CPR for more than 45 minutes before randomisation or who had a mechanical cause of CS or severe peripheral-artery disease precluding the insertion of cannulae. |
ECLS was not initiated in 17 patients in the ECLS group (8.1%), including in 4 patients who died before initiation. ECLS was initiated in 26 patients in the control group (12.5%), including 22 patients within 24 hours after randomisation and 4 patients thereafter. IABP was permitted as escalation therapy for the control group. | 30 days |
3 | Banning, 2023, Belgium, Germany, Latvia, Norway, Spain, UK EURO SHOCK | n=35 (VA ECMO n=17) Median age (years)
Male (%)
Median LVEF on admission
| Randomised controlled trial, open label. Randomisation was carried out using a web-based randomisation system stratified by out-of-hospital cardiac arrest (OHCA). Due to the impact of the COVID-19 pandemic, the trial was stopped before completion of recruitment. | People presenting with CS-complicating AMI and who had had attempted/ successful primary PCI (PPCI) of the culprit lesion were enrolled if there was persistent CS 30 mins after the procedure. CS defined as BP <90 mmHg or maintained above 90 mmHg with the addition of vasopressor or inotropic support, with evidence of hypoperfusion. |
IABP was permitted as escalation therapy for the control group, or for left ventricular unloading in the VA ECMO group. 5 patients randomised to ECMO did not receive ECMO. | 30 days, 1 year |
4 | Ostadal, 2023 Czech Republic ECMO-CS | n=117 (ECMO n=58) Median age (years)
Male (%)
The most common cause of CS in both arms was STEMI (50.4%) followed by decompensation of chronic heart failure (23.1%). | Randomised controlled trial, open label. An automated, web-based system was used for randomisation with permuted blocks, with stratification according to the type of cardiogenic shock (rapidly deteriorating or severe), and the trial centre. | People over 18 with rapidly deteriorating or severe CS. Rapidly deteriorating CS defined as SCAI stage D to E Severe CS defined as SCAI stage D |
| 30 days |
5 | Sohail, 2022, Asia, Australia, Europe, North America | n=10,276 Median age (years)=60 (IQR 56.35 to 63.94) Male % = 78% | Systematic review and meta-analysis of 72 studies. Search date: August 2020 | Adults (over 18 years) receiving VA ECMO for CS complicating AMI. |
| Short term mortality (7, 30 days, discharge) |
6 | Alba, 2021 Europe, Asia, North America, South America, Australia | n=29,289 Age (years): Range 47 to 61 Female %: Range 22 to 59 Indication
| Systematic review and meta-analysis of 306 observational studies. Search date: June 2019 | Adults (aged 18 and over) with CS of any aetiology, with VA ECMO implantation. |
| 30 day or discharge |
7 | Cheng, 2019 US | n=149 (ECMO n=118) Median age (years)=59 (51-67) Male=67.8% Aetiology:
| Single centre retrospective study (26-bed ICU) Search date: 2010 to 2016 | People who survived VA ECMO or CentriMag VAD support as a short-term MCS as bridge to recovery. |
| Median 306 days (IQR 58.925 to 916.75) |
8 | Vishram-Nielsen, 2023 Asia, Australia, Europe, North America | n=2,388 Median age (years) = 41 (IQR 37 to 47) Male % = 50% | Systematic review and meta-analysis of 54 retrospective studies. Search date: July 2020 | Adult (aged 18 and over) patients with fulminant myocarditis, evaluating short-term mortality after VA ECMO implantation. |
| 30 day, hospital discharge |
9 | Olson, 2020 Worldwide | n=88 Median age (years): 31.1 (IQR 25.4 to 35.2) | Retrospective ELSO registry study Search date: 2007 to 2019 | People with peripartum cardiomyopathy treated with ECMO. |
| Hospital discharge |
Procedure technique
Of the 10 studies, none detailed the ECMO device or combination of devices used. ECMO was started before percutaneous coronary intervention (PCI) in one randomised controlled trial of people with CS complicating AMI (Thiele 2023) but was started within 6 hours of randomisation in people who had already undergone PCI in another (Banning 2023). Left ventricular (LV) venting strategies were detailed in 6 studies (Thiele 2023, Banning 2023, Ostadal 2023, Sohail 2022, Alba 2021, Vishram-Neilsen 2023); 2 RCTs had a predefined criteria for LV venting and permitted insertion of an intra-aortic balloon pump (IABP) or Impella device (Thiele 2023, Banning 2023), another randomised controlled trial permitted LV unloading but strategies were left to the discretion of physicians at participating centres (Ostadal 2023). The median concomitant use of IABP reported in systematic reviews was 70% (Sohail 2022), 20 to 67% (Alba 2021) and 60% (Vishram-Neilsen 2023). Of the 10 studies, 6 detailed the median length of time on ECMO (Thiele 2023, Sohail 2022, Alba 2021, Cheng 2019, Vishram-Neilsen 2023, Olson 2020), which ranged from 2.7 days (Thiele 2023) to 10.5 days (Cheng 2019).
Efficacy
Survival
Survival was reported in 2 retrospective studies (Cheng 2019, Olson 2020). The retrospective study of 88 people with peripartum cardiomyopathy reported a rate of survival to hospital discharge of 64% (Olson 2020). In the randomised controlled trial of 117 people with CS complicating AMI, 12% of people in both the ECMO group and the control group had been discharged home at 30 days (Ostadal 2023).
In the single centre retrospective study, 30% (149 of 502) of people having VA ECMO survived to discharge (Cheng 2019). Of these survivors, the Kaplan-Meier estimate of survival at 3 years was 74% in the overall population but was statistically significantly lower (p<0.001) in people with AMI (40%) compared to those with ADHF (100%), postcardiotomy (86%), allograft failure (74%) (Cheng 2019).
Short-term mortality
Short-term mortality was reported in 7 studies, 4 of which included mortality at 30 days and 3 systematic reviews which pooled mortality results at 30 days and hospital discharge.
In the systematic review of 611 people with CS complicating AMI across 4 randomised controlled trials, the pooled 30-day mortality was 46% for those who had VA ECMO, compared to 48% in the control group. The relative risk (RR) was 0.95 (95% CI: 0.80 to 1.12; p=0.54, I2=0%) (Elsaeidy 2024).
In the randomised controlled trial of 417 people with CS complicating AMI, 47% of people who had ECMO and 49% of people in the control group reported death from any cause at 30 days. The RR was 0.98 (95% CI: 0.80 to 1.19; p=0.81; Thiele 2023). In the randomised controlled trial of 117 people with CS complicating AMI, all-cause mortality at 30 days was 50% in the ECMO group compared to 48% in the control group. The risk difference was 2.5 (95% CI -15.6 to 20.7) and HR 1.11 (95% CI 0.66 to 1.87) (Ostadal 2023). In the randomised controlled trial of 35 people with CS complicating AMI, 30-day all-cause mortality was 44% for those who had ECMO compared to 61% for those who had standard therapy. The hazard ratio (HR) was 0.56 (95% CI: 0.21 to 1.45; p=0.22). At 1 year follow-up, all-cause mortality was 52% in people who had ECMO, and 82% in those who had standard therapy (HR 0.52, 95% CI 0.21 to 1.26, p=0.14; Banning 2023). In the systematic review of 72 studies of CS complicating AMI, the pooled short-term mortality (30-day and in-hospital) was 58% (95% CI: 54 to 61%), I2=88% (Sohail 2022).
In a subgroup analysis by age in the randomised controlled trial of 417 people with CS complicating AMI, the 30-day all-cause mortality rate for those who had ECMO was 40% in those under 65 years, and 59% in those over 65 years (Thiele 2023). In a subgroup analysis by age in the systematic review of CS complicating AMI, age greater than 60 years was associated with increased mortality OR 4.58 (95% CI: 2.71 to 7.72; Sohail 2022).
In the systematic review of 306 studies of CS of any aetiology, the pooled overall short-term mortality (30-day and in-hospital) was 61% (95% CI 59 to 63) (Alba 2021). Pooled short-term mortality by CS aetiology subgroup, showed the highest mortality was in people with ECPR for out of hospital cardiac arrest (OHCA) (76%; 95% CI 69 to 82%, I2=94%, 41 studies). This was followed by ECPR for in hospital CA (IHCA) at 64% (95% CI 59 to 69, I2=81%, 46 studies), post-AMI at 60% (95% CI 59 to 64, I2=87%, 80 studies), postcardiotomy at 59% (95% CI 56 to 63, I2=87%, 64 studies). Pooled short-term mortality for people with acute decompensated heart failure (ADHF) was 53% (95% CI 46 to 59, I2=89%, 33 studies), 52% in people with pulmonary embolism (95% CI 38 to 66, I2=75%, 10 studies). It was lowest in people with myocarditis at 40% (95% CI 33 to 46), I2=65%, 13 studies) and after heart transplant 35% (95% CI 29 to 42, I2=64%, 25 studies). Using multivariate meta regression analysis, differences in short-term mortality across aetiologies remained statistically significant (p<0.01) after adjusting for population age, sex, and recruitment timeframe. Univariate meta regression analysis stratified by aetiology also showed a 7% to 9% increase in mortality per 10-year increase in cohort's age (Alba 2021).
In the systematic review of 54 studies in people with fulminant myocarditis, the pooled short-term mortality (30-day and in-hospital) was 35% (95% CI: 29 to 40, I2=69%, 50 studies). The pooled short-term mortality from 36 studies looking at death on ECMO was 27% (95% CI: 21 to 34, I2=67%; Vishram-Nielsen 2023).
In the single centre retrospective study of 149 people who had survived VA ECMO explantation, 14% died after hospital discharge (median follow-up 306 days) (Cheng 2019).
Bridged to heart transplant
The proportion of people who had a heart transplant after ECMO treatment was reported in 3 systematic reviews and 1 registry study. In the systematic review of 306 studies in people on ECMO for CS of any aetiology, meta-analyses demonstrated the probability of having a heart transplant was higher in people with heart failure (13%), compared to those with myocarditis (5%), AMI (3%), and postcardiotomy CS (less than 1%; Alba 2021). In the systematic review of people with fulminant myocarditis, the pooled probability of heart transplant in a meta-analysis of 23 studies was 4% (95% CI: 0.47 to 8.76, I2=72%; Vishal-Nielsen 2023). In the registry study of 88 people with peripartum cardiomyopathy, 10% were weaned from ECMO to either heart transplant or a VAD (Olson, 2020).
Bridged to long term device
The proportion of people receiving a ventricular assist device (VAD) after ECMO treatment was reported in 2 systematic reviews and 2 registry studies. In the systematic review of 306 studies in people on ECMO for CS, meta-analyses demonstrated the probability of receiving a VAD was higher in people with heart failure (29%), compared to those with AMI (9%), myocarditis (5%), heart transplant (2%) and postcardiotomy (1%; Alba 2021). In the systematic review of people with fulminant myocarditis, the pooled probability of heart transplant in a meta-analysis of 22 studies was 2% (95% CI: 0.47 to 8.76, I2=72%; Vishal-Nielsen 2023). In the registry study of 88 people with peripartum cardiomyopathy, 10% were weaned from ECMO to either heart transplant or a VAD (Olson, 2020).
Reinfarction
In the systematic review of 4 randomised controlled trials for people with CS complicating AMI, pooled 30-day reinfarction rate was 2% in both the VA ECMO and control group (RR 0.87, 95% CI: 0.25 to 3.04; p=0.83, I2=0%) (Elsaeidy 2024). In the randomised controlled trial of 417 people with CS complicating AMI, both intervention and control groups reported myocardial reinfarction rates of 1% (Thiele, 2023).
Rehospitalisation for heart failure
Two randomised controlled trials reported on the number of people who were readmitted to hospital because of heart failure. In the randomised controlled trial of 417 people with CS complicating AMI, 1% of people in both the ECMO group and control group were re-hospitalised within 30 days because of heart failure (Thiele, 2023). The randomised controlled trial of 35 people with CS complicating AMI reported a readmission rate for heart failure of 8% in the ECMO group and 7% in the standard therapy group at 1 year follow-up (Banning 2023).
Quality of life
One randomised controlled trial reported quality of life at 30 days using the EQ-5D-3L questionnaire, however few completed the questionnaire in both the standard therapy (n=2) and VA ECMO (n=4) groups. Among the respondents, the median summary index score for those on ECMO was 0.667 (0.326 to 1.00), and 0.765 (0.739 to 0.790) for those on standard therapy. In the standard therapy group, there were no reported problems with mobility, self-care, or usual activities at 30 days, while half of the respondents from the VA ECMO group reported some difficulties in these domains at 30 days (Banning, 2023).
Safety
Bleeding
In the systematic review of 4 randomised controlled trials, the pooled bleeding event rate was 25% (76 out of 302) in the ECMO group compared to 12% (36 out of 306) in the control group (RR 2.14, 95% CI: 1.49 to 3.07; p<0.0001, I2=0%; Elsaeidy 2024). Moderate or severe bleeding was reported in 23% (49 out of 209) of people on ECMO compared to 12% (20 out of 208) of people in the control group (RR 2.44, 95% CI: 1.50 to 3.95) in the randomised controlled trial of 417 people with CS complicating AMI (Thiele 2023). Major bleeding was reported in 36% (5 out of 14) of people on ECMO and 6% (1 out of 18) of people in the control group in the randomised controlled trial of 35 people with CS complicating AMI (Banning 2023). Bleeding complications were reported in 31% (18 out of 58) of people on ECMO compared to 20% (12 out of 59) of the control group in the randomised controlled trial of 177 people with CS complicating AMI (Ostadal 2023). Pooled bleeding/vascular complication rates were 28% (19.0 to 35.4) in the systematic review of 72 studies with CS complicating AMI (Sohail, 2022).
A pooled bleeding event rate of 40% (95% CI: 22.89 to 58.92, I2=76%) was reported in a meta-analysis of 6 studies in the systematic review of people with fulminant myocarditis on ECMO (Vishram-Nielsen 2023).
Bleeding complications were reported in 49% (31 out of 88) of people on ECMO with peripartum cardiomyopathy. Cannulation site bleeds were reported in 16 out of 88 people and surgical site bleeds in 13 out of 88 people (Olson 2020).
Renal replacement therapy or acute kidney injury
In the systematic review of 4 randomised controlled trials, the pooled acute kidney injury or RRT event rate was 9% (25 out of 281) in the ECMO group compared to 14% (40 out of 285) in the control group (RR 0.65, 95% CI: 0.41 to 1.04; p=0.07, I2=0%; Elsaeidy 2024). RRT was reported in 8% (17 out of 209) of people on ECMO compared to 14% (29 out of 208) of people in the control group (RR 0.58, 95% CI: 0.33 to 1.03) in the randomised controlled trial of 417 people with CS complicating AMI (Thiele 2023). Acute kidney injury was reported in 29% (4 out of 14) of people on ECMO and 44% (8 out of 18) of people in the control group in the randomised controlled trial of 35 people with CS complicating AMI (Banning 2023). Pooled renal failure rates were 40% (16.4 to 41.7) in the systematic review of 72 studies with CS complicating AMI (Sohail, 2022).
A pooled RRT event rate of 35% (95% CI: 11.90 to 62.35, I2=89%) was reported in a meta-analysis of 6 studies in the systematic review of people with fulminant myocarditis on ECMO (Vishram-Nielsen 2023).
Renal complications were reported in 38% (24 out of 88) people on ECMO with peripartum cardiomyopathy. RRT was reported in 16 out of 88 people (Olson 2020).
Stroke
In the systematic review of 4 randomised controlled trials, the pooled stroke event rate was 4% in both ECMO and control groups (12 out of 302 and 11 out of 306) (RR 1.14, 95% CI: 0.52 to 2.49; p=0.75, I2=18%; Elsaeidy 2024). Stroke or systemic embolisation was reported in 4% (8 out of 209) of people on ECMO and 3% (6 out of 208) of people in the control group (RR 1.33, 95% CI: 0.47 to 3.76) in the RCT of 417 people with CS complicating AMI (Thiele 2023). Stroke was not reported in any people on ECMO (0 out of 14) but in 11% (2 out of 18) of people in the control group in the randomised controlled trial of 35 people with CS complicating AMI (Banning 2023). Conversely, in the randomised controlled trial of 177 people with CS complicating AMI, stroke was reported in 5% (3 out of 58) of people on ECMO and none in the control group in the (Ostadal 2023). Pooled stroke rates were 11% (5.0 to 16.7) in the systematic review of 72 studies with CS complicating AMI (Sohail 2022).
A pooled neurological event rate of 7% (95% CI: 3.25 to 12.60, I2=30%) was reported in a meta-analysis of 8 studies in the systematic review of people with fulminant myocarditis on ECMO (Vishram-Nielsen 2023).
Neurological complications were reported in 14% (9 out of 88) of people on ECMO with peripartum cardiomyopathy. CNS infarction was reported in 3 out of 88 people, and CNS haemorrhage in 5 out of 88 people (Olson 2020).
Of the 21 people who died after hospital discharge (median follow-up 306 days) in the single centre retrospective study of people who had initially survived VA ECMO explantation, 1 cause of death was reported as due to stroke (Cheng 2019).
Neurological outcome
One randomised controlled trial reported on the proportion of people with a good neurological outcome at 30 days, assessed as category 1 using the Cerebral Performance Category (CPC 1). In the study of 117 people with CS complicating AMI, the proportion of people assessed as CPC 1 was 24% in the ECMO group compared to 27% in the control group (Ostadal 2023).
One randomised controlled trial reported on the proportion of people with a poor neurological outcome at 30 days, assessed as category 3 or 4 using the Cerebral Performance Category (CPC 3 or 4). In the study of 417 people with CS complicating AMI, the proportion of people assessed as CPC 3 or 4 was 25% (27 out of 109) in the ECMO group compared to 23% (24 out of 106) in the control group (Thiele 2023).
Sepsis
In the systematic review of 4 randomised controlled trials, the pooled sepsis event rate was 18% (54 out of 305) in the ECMO group compared to 17% (51 out of 306) in the control group (RR 1.07, 95% CI: 0.77 to 1.48; p=0.85, I2=0%; Elsaeidy 2024). Post-procedural sepsis was reported in 1 person on ECMO, and septic shock in 1 person in the control group in the randomised controlled trial of 35 people with CS complicating AMI (Banning 2023). Sepsis was reported in 40% (23 out of 58) of people on ECMO compared to 39% (23 out of 59) of the control group in the randomised controlled trial of 177 people with CS complicating AMI (risk difference 0.7, 95% CI: -17.0 to 18.4; p=0.941; Ostadal 2023).
Of the 21 people who died after hospital discharge (median follow-up 306 days) in the single centre retrospective study of people who had initially survived VA ECMO explantation, 4 causes of death were reported as due to sepsis (Cheng 2019).
Infection
The pooled median infection rate was 18% (11.8 to 43.0) in the systematic review of 72 studies with CS complicating AMI (Sohail 2022). A pooled infection event rate of 35% (95% CI: 15.80 to 56.34, I2=79%) was reported in a meta-analysis of 8 studies in the systematic review of people with fulminant myocarditis on ECMO (Vishram-Nielsen 2023). Infectious complications were reported in 8% (5 out of 88) of people on ECMO with peripartum cardiomyopathy. Culture proven infection was reported in 5 out of 88 people, and white blood cell count <1,500/μl in 2 out of 88 people (Olson 2020).
Pneumonia
In the systematic review of 2 randomised controlled trials, the pooled pneumonia event rate was 24% (18 out of 75) in the ECMO group compared to 25% (19 out of 77) in the control group (RR 0.97, 95% CI: 0.57 to 1.65; p=0.90, I2=0%; Elsaeidy 2024). Pneumonia was reported in 31% (18 out of 58 and 18 out of 59) of people in both groups in the randomised controlled trial of 177 people with CS complicating AMI (risk difference 0.5, 95% CI: -16.2 to 17.3; p=0.951; Ostadal 2023).
Limb ischaemia
Peripheral ischaemic vascular complications were reported in 11% (23 out of 209) of people on ECMO compared to 4% (8 out of 208) of people in the control group (RR 2.86, 95% CI: 1.31 to 6.25) in the randomised controlled trial of 417 people with CS complicating AMI (Thiele 2023). Leg ischaemia was reported in 14% (8 out of 58) of people on ECMO compared to 5% (3 out of 59) of the control group in the randomised controlled trial of 177 people with CS complicating AMI (risk difference 8.7, 95% CI: -1.8 to 19.2; p=0.107; Ostadal 2023). Pooled limb ischaemia rates were 9% (7.6 to 15.0) in the systematic review of 72 studies with CS complicating AMI (Sohail 2022). The pooled limb ischaemia event rate of 17% (95% CI: 5.78 to 30.65, I2=69%) was reported in a meta-analysis of 6 studies in the systematic review of people with fulminant myocarditis on ECMO (Vishram-Nielsen 2023). Limb complications were reported in 8% (5 out of 88) of people on ECMO with peripartum cardiomyopathy. Limb ischaemia was reported in 3 out of 88 people, and limb fasciotomy in 2 out of 88 people (Olson 2020).
Cardiac complications
Cardiovascular death was reported in 14% (2 out of 14) of people in the ECMO group, and 33% (6 out of 18) of people in the control group in the randomised controlled trial of 35 people with CS complicating AMI (Banning 2023). The same study reported recurrent MI in 11% (2 out of 18) people in the control group, and none in the ECMO group. It also reported 5 serious cardiac adverse events in people who had ECMO compared to 4 in those who had standard therapy. These included cardiac arrest, cardiac tamponade, ventricular tachycardia, LV thrombus, AV block and atrial fibrillation (Banning 2023). Resuscitated cardiac arrest was reported in 10% (6 out of 58) of people in the ECMO group compared to 14% (8 out of 59) of the control group in the randomised controlled trial of 177 people with CS complicating AMI (risk difference -3.2, 95% CI: -15.0 to 8.5; Ostadal 2023).
A pooled ventricular tachycardia or fibrillation event rate of 23% (95% CI: 1.73 to 50.96; I2=84%) was reported in a meta-analysis of 4 studies in the systematic review of people with fulminant myocarditis who had ECMO. The same systematic review also reported a pooled rate of third degree atrioventricular block of 23% (95% CI: 0.00 to 78.46; I2=93%) from 3 studies (Vishram-Nielsen 2023).
Cardiovascular complications were reported in 48% (30 out of 88) of people on ECMO with peripartum cardiomyopathy. These included inotropes on ECMO, cardiac arrhythmia, hypertension requiring vasodilators and myocardial stun by echocardiogram (Olson 2020).
Of the 21 people who died after hospital discharge (median follow-up 306 days) in the single centre retrospective study of people who had initially survived VA ECMO explantation, 4 causes of death were reported as heart failure related (Cheng 2019).
Respiratory complications
The randomised controlled trial of 35 people with CS complicating AMI reported 1 serious respiratory adverse event in people on ECMO compared to 2 in those on standard therapy. These included pulmonary embolism, aspiration pneumonia and thoracic haemorrhage (Banning 2023).
Pulmonary complications were reported in 10% (9 out of 88) of people on ECMO with peripartum cardiomyopathy. This included pneumothorax requiring treatment and pulmonary haemorrhage (Olson 2020).
GI complications
The randomised controlled trial of 35 people with CS complicating AMI reported 1 serious gastrointestinal (GI) adverse event (intestinal ischaemia) in people on standard therapy (Banning 2023).
Hepatic complications
The randomised controlled trial of 35 people with CS complicating AMI reported 1 serious hepatobiliary adverse event (liver injury) in people on standard therapy (Banning 2023).
A pooled liver failure event rate of 6% (95% CI: 0.41 to 14.20; I2=0%) was reported in a meta-analysis of 2 studies in the systematic review of people with fulminant myocarditis on ECMO (Vishram-Nielsen 2023).
Technical complications
Technical complications were reported in 2% (1 out of 58) of people on ECMO compared to none in the control group in the randomised controlled trial of 177 people with CS complicating AMI (risk difference 1.7, 95% CI: -1.6 to 5.1; p=0.496; Ostadal 2023).
Mechanical complications were reported in 33% (21 out of 88) of people on ECMO with peripartum cardiomyopathy. These included cannula problems, circuit clots, and pump malfunctions (Olson 2020).
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 studies included in the key evidence had a large number of participants from a variety of countries.
Due to the impact of the COVID-19 pandemic, the Banning (2023) trial was stopped before completion of recruitment, and therefore had a small sample size (n=35). Olson (2020) study also included a relatively small population (n=88); however, this was the largest study sample identified for the postpartum cardiomyopathy indication.
Follow-up for most studies were short, reporting key efficacy outcomes at 30 days, or at hospital discharge. One retrospective study reporting on ECMO as a bridge to recovery had a longer follow-up period (Cheng 2019).
The systematic reviews included as key evidence pooled short-term mortality outcomes (30 day, hospital discharge) from included studies.
CS can have many aetiologies with different risk profiles and outcomes. Most studies included in this review focus on CS complicating AMI, however other included studies report populations with mixed CS aetiologies.
People having ECMO for decompensated AHF, myocarditis, peri-partum cardiomyopathy may have better outcomes than people with AMI complicating CS.
The randomised controlled trials included recruited participants with different classifications of CS; Thiele (2023) included those with SCAI classification C, which is considered much lower risk, than SCAI D and E, which were included in the Ostadal (2023) study.
A large proportion of people in the control groups in the included randomised controlled trials were permitted other MCS such as IABP and Impella. In the Ostadal (2023) trial there was also a large amount of cross-over of the control group to receive ECMO (39%).
Ongoing trials
Assessment of ECMO in Acute Myocardial Infarction Cardiogenic Shock (ANCHOR) (NCT04184635); open label randomised controlled trial, France, n=400, completion October 2026.
Left Ventricular Unloading to Improve Outcome in Cardiogenic Shock Patients on VA ECMO (UNLOAD ECMO) (NCT05577195), randomised controlled trial, Germany, n=198, completion December 2025.
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