The role of extracorporeal membrane oxygenation (ECMO) remains ill defined in pulmonary embolism (PE). We investigated outcomes in patients with high-risk PE undergoing ECMO according to initial therapeutic strategy.

From 01 January 2014 to 31 December 2015, 180 patients from 13 Departments in nine centres with high-risk PE were retrospectively included. Among those undergoing ECMO, we compared characteristics and outcomes according to adjunctive treatment strategy (systemic thrombolysis, surgical embolectomy, or no reperfusion therapy). Primary outcome was all-cause 30-day mortality. Secondary outcome was 90-day major bleeding. One hundred and twenty-eight patients were treated without ECMO; 52 (mean age 47.6 years) underwent ECMO. Overall 30-day mortality was 48.3% [95% confidence interval (CI) 41–56] (87/180); 43% (95% CI 34–52) (55/128) in those treated without ECMO vs. 61.5% (95% CI 52–78) (32/52) in those with ECMO (P = 0.008). In patients undergoing ECMO, 30-day mortality was 76.5% (95% CI 57–97) (13/17) for ECMO + fibrinolysis, 29.4% (95% CI 51–89) (5/17) for ECMO + surgical embolectomy, and 77.7% (95% CI 59–97) (14/18) for ECMO alone (P = 0.004). Among patients with ECMO, 20 (38.5%, 95% CI 25–52) had a major bleeding event in-hospital; without significant difference across groups.

In patients with high-risk PE, those with ECMO have a more severe presentation and worse prognosis. Extracorporeal membrane oxygenation in patients with failed fibrinolysis and in those with no reperfusion seems to be associated with particularly unfavourable prognosis compared with ECMO performed in addition to surgical embolectomy. Our findings suggest that ECMO does not appear justified as a stand-alone treatment strategy in PE patients, but shows promise as a complement to surgical embolectomy.