Mortality in Patients with Cardiogenic Shock Supported with Veno-Arterial Extracorporeal Membrane Oxygenation: A Systematic Review and Meta-Analysis Evaluating the Impact of Etiology

Purpose Veno-arterial extracorporeal membrane oxygenation (VA-ECMO) is associated with improved outcomes in patients with cardiogenic shock (CS). However, studies evaluating its benefit have reported contradictory results. One of the main factors impacting outcomes is likely related to etiology of CS. This meta-analysis aimed to evaluate mortality after VA-ECMO across multiple etiologies of CS. Methods In June 2019, we systematically searched for studies in electronic databases and references of previous meta-analyses. We selected observational studies with ≥10 adults patients reporting on short-term mortality (30 days or discharge) after the initiation of VA-ECMO by CS etiology published after 2009. We meta-analyzed mortality using random effect models and performed meta-regression to evaluate the impact of CS etiology and other study characteristics. We evaluated the quality of the evidence using the GRADE tool. Results We included 342 studies (29,515 patients) with CS due to myocarditis (11 studies, 846 patients), acute myocardial infarction - AMI (66 studies, 3,849 patients), decompensated heart failure-HF (30 studies, 3,251 patients), post-heart transplant (23 studies, 703 patients), post-cardiotomy (93 studies, 12,076 patients), and post-cardiac arrest-CA (119 studies, 8,790 patients). Studies on myocarditis included younger populations (median age 47 years, 25th - 75th percentile 43-53), and higher proportion of female patients (48%). Studies on AMI and post-cardiotomy reported older population age (61 years, 56-65). Pooled mortality was significantly different by CS etiology: in post-heart transplant 36% (95%CI 29%-43%), myocarditis 41% (34%-47%), HF 51% (44%-58%), post-cardiotomy 58% (55%-62%), post-AMI 60% (56%-64%), post-in-hospital CA 63% (56%-69%), and post-out-of-hospital CA 76% (69%-82%). Differences in age, sex and recruitment time across studies did not explain the differences in mortality in patients with different CS etiology. The quality of the evidence was deemed moderate. Conclusion The benefit of VA-ECMO differs significantly depending on CS etiology. Further studies evaluating other factors impacting mortality in each patient population may further refine patient selection and management to improve short-term outcomes.