Abstract
BACKGROUND: Veno-arterial extracorporeal membrane oxygenation (VA-ECMO) has been used as extracorporeal cardiopulmonary resuscitation (ECPR) to support further resuscitation efforts in patients with cardiac arrest, yet its clinical effectiveness remains uncertain. OBJECTIVES: This study reviews the role of ECPR in contemporary resuscitation care compared to no ECPR and/or standard care, e.g. conventional CPR, and quantitatively summarize the rates of long-term neurologically intact survival after adult in-hospital cardiac arrest (IHCA) or out-of-hospital cardiac arrest (OHCA). METHODS: We searched the following databases on January 31 st, 2020: CENTRAL, MEDLINE, Embase, and Web of Science. We followed PRISMA guidelines and used PICO format to summarize the research questions. Risk of bias was assessed using the ROBINS-I tool. Pooled risk ratios (RRs) for each outcome of interest were calculated. Quality of evidence was evaluated according to GRADE guidelines. RESULTS: Six cohort studies were included, totaling 1750 patients. Of these, 530 (30.3%) received the intervention, and 91 (17.2%) survived with long-term neurologically intact survival. ECPR compared to no ECPR is likely associated with improved long-term neurologically intact survival after cardiac arrest in any setting (risk ratio [RR] 3.11, 95% confidence interval [CI] 2.06-4.69; p < 0.00001) (GRADE: Very low quality). Similar results were found for long-term neurologically intact survival after IHCA (RR 3.21, 95% CI 1.74-5.94; p < 0.0002) (GRADE: Very low quality) and OHCA (RR 3.11, 95% CI 1.50-6.47; p < 0.002) (GRADE: Very low quality). Long-term time frames for neurologically intact survival (three months to two years) were combined into a single category, defined a priori as a Glasgow-Pittsburgh cerebral performance category (CPC) of 1 or 2. CONCLUSIONS: VA-ECMO used as ECPR is likely associated with improved long-term neurologically intact survival after cardiac arrest. Future evidence from randomized trials is very likely to have an important impact on the estimated effect of this intervention and will further define optimal clinical practice. Review registration: PROSPERO CRD42020171945.