Perioperative Veno-Venous Extracorporeal Membrane Oxygenation in Non-Elective Thoracic Surgery: A Propensity-Matched Bicentric Comparison With Medical Indication.

OBJECTIVES: Perioperative veno-venous extracorporeal membrane oxygenation (VV-ECMO) can be used during non-elective thoracic surgery to manage hypoxaemia and facilitate complex surgical procedures. However, outcome data in this setting remain limited, particularly regarding bleeding and thrombotic risks. We aimed to evaluate whether perioperative VV-ECMO during non-elective thoracic surgery is associated with increased 90-day all-cause mortality compared with VV-ECMO used for medical indications, using a propensity score-matched analysis. Secondary outcomes included major bleeding and thrombotic events. METHODS: This retrospective bi-centre cohort study included adults treated with VV-ECMO for acute respiratory failure between January 2011 and January 2025 in 2 university intensive care units. Patients were categorized according to ECMO indication: perioperative thoracic surgery or medical. A propensity score was built using multivariable logistic regression, followed by 1:4 nearest-neighbor matching without replacement. The primary endpoint was 90-day mortality; secondary endpoints included major bleeding and thrombotic events. RESULTS: Among 372 patients, 44 (12%) received perioperative VV-ECMO. Ventilator-associated pneumonia (78% vs 57%, P = .01) and haemothorax (36% vs 2%, P < .001) were more frequent in the perioperative group. Ninety-day mortality did not differ between groups, both before (54% vs 51%, P = .71) and after matching (log-rank P = .95). Multivariable Cox analysis confirmed no association between surgical ECMO and mortality (HR = 1.003, 95% CI [0.64-1.57]). No significant differences were found in major bleeding (48% vs 47%, P = 1.00) or thrombotic events (41% vs 32%, P = .31) between groups. CONCLUSIONS: Perioperative VV-ECMO during non-elective thoracic surgery was not associated with increased 90-day mortality, bleeding, or thrombotic complications compared to medical VV-ECMO, both before and after matching. These findings support its use in carefully selected patients. CLINICAL TRIAL REGISTRATION NUMBER: NCT07016685.