Extracorporeal membrane oxygenation for COVID-19: a systematic review and meta-analysis.

BACKGROUND: There are several reports of extracorporeal membrane oxygenation (ECMO) use in patients with coronavirus disease 2019 (COVID-19) who develop severe acute respiratory distress syndrome (ARDS). We conducted a systematic review and meta-analysis to guide clinical decision-making and future research. METHODS: We searched MEDLINE, Embase, Cochrane and Scopus databases from 1 December 2019 to 10 January 2021 for observational studies or randomised clinical trials examining ECMO in adults with COVID-19 ARDS. We performed random-effects meta-analyses and meta-regression, assessed risk of bias using the Joanna Briggs Institute checklist and rated the certainty of evidence using the GRADE approach. Survival outcomes were presented as pooled proportions while continuous outcomes were presented as pooled means, both with corresponding 95% confidence intervals [CIs]. The primary outcome was in-hospital mortality. Secondary outcomes were duration of ECMO therapy and mechanical ventilation, weaning rate from ECMO and complications during ECMO. RESULTS: We included twenty-two observational studies with 1896 patients in the meta-analysis. Venovenous ECMO was the predominant mode used (98.6%). The pooled in-hospital mortality in COVID-19 patients (22 studies, 1896 patients) supported with ECMO was 37.1% (95% CI 32.3-42.0%, high certainty). Pooled mortality in the venovenous ECMO group was 35.7% (95% CI 30.7-40.7%, high certainty). Meta-regression found that age and ECMO duration were associated with increased mortality. Duration of ECMO support (18 studies, 1844 patients) was 15.1 days (95% CI 13.4-18.7). Weaning from ECMO (17 studies, 1412 patients) was accomplished in 67.6% (95% CI 50.5-82.7%) of patients. There were a total of 1583 ECMO complications reported (18 studies, 1721 patients) and renal complications were the most common. CONCLUSION: The majority of patients received venovenous ECMO support for COVID-19-related ARDS. In-hospital mortality in patients receivin