Although outcomes of patients after cardiac arrest remain poor, studies have suggested that extracorporeal cardiopulmonary resuscitation (ECPR) might improve survival and neurological outcomes. We aimed to investigate any potential benefits of using ECPR over conventional cardiopulmonary resuscitation (CCPR) in patients with out-of-hospital cardiac arrest (OHCA) and in-hospital cardiac arrest (IHCA). In this systematic review and meta-analysis, we searched MEDLINE via PubMed, Embase, and Scopus from Jan 1, 2000, to April 1, 2023, for randomised controlled trials and propensity-score matched studies. We included studies comparing ECPR with CCPR in adults (aged ≥18 years) with OHCA and IHCA. We extracted data from published reports using a prespecified data extraction form. We did random-effects (Mantel-Haenszel) meta-analyses and rated the certainty of evidence using the Grading of Recommendations, Assessments, Developments, and Evaluations (GRADE) approach. We rated the risk of bias of randomised controlled trials using the Cochrane risk-of-bias 2.0 tool, and that of observational studies using the Newcastle-Ottawa Scale. The primary outcome was in-hospital mortality. Secondary outcomes included complications during extracorporeal membrane oxygenation, short-term (from hospital discharge to 30 days after cardiac arrest) and long-term (≥90 days after cardiac arrest) survival with favourable neurological outcomes (defined as cerebral performance category scores 1 or 2), and survival at 30 days, 3 months, 6 months, and 1 year after cardiac arrest. We also did trial sequential analyses to evaluate the required information sizes in the meta-analyses to detect clinically relevant reductions in mortality. We included 11 studies (4595 patients receiving ECPR and 4597 patients receiving CCPR) in the meta-analysis. ECPR was associated with a significant reduction in overall in-hospital mortality (OR 0·67, 95% CI 0·51-0·87; p=0·0034; high certainty), without evidence of publication bias (pegger=0·19); the trial sequential analysis was concordant with the meta-analysis. When considering IHCA only, in-hospital mortality was lower in patients receiving ECPR than in those receiving CCPR (0·42, 0·25-0·70; p=0·0009), whereas when considering OHCA only, no differences were found (0·76, 0·54-1·07; p=0·12). Centre volume (ie, the number of ECPR runs done per year in each centre) was associated with reductions in odds of mortality (regression coefficient per doubling of centre volume -0·17, 95% CI -0·32 to -0·017; p=0·030). ECPR was also associated with an increased rate of short-term (OR 1·65, 95% CI 1·02-2·68; p=0·042; moderate certainty) and long-term (2·04, 1·41-2·94; p=0·0001; high certainty) survival with favourable neurological outcomes. Additionally, patients receiving ECPR had increased survival at 30-day (OR 1·45, 95% CI 1·08-1·96; p=0·015), 3-month (3·98, 1·12-14·16; p=0·033), 6-month (1·87, 1·36-2·57; p=0·0001), and 1-year (1·72, 1·52-1·95; p<0·0001) follow-ups. Compared with CCPR, ECPR reduced in-hospital mortality and improved long-term neurological outcomes and post-arrest survival, particularly in patients with IHCA. These findings suggest that ECPR could be considered for eligible patients with IHCA, although further research into patients with OHCA is warranted. None.