In-hospital mortality and mode-specific risk factors in adults receiving emergency ECMO in three Chinese tertiary emergency centers: a multicenter retrospective cohort
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background Emergency extracorporeal membrane oxygenation (ECMO) is increasingly applied for refractory cardiac or respiratory failure, yet in-hospital mortality remains substantial. Comprehensive multicenter evidence from departments of emergency medicine is limited. Methods We conducted a retrospective, multicenter study of consecutive adult patients receiving ECMO between January 2019 and February 2025 at three tertiary emergency centers. Baseline characteristics, including demographics, comorbidities, therapeutic interventions, vital signs, and laboratory parameters, were collected. Independent predictors of in‑hospital mortality were determined using multivariable logistic regression, and model performance was evaluated by receiver operating characteristic (ROC) curve anal-ysis. Survival outcomes were analyzed using the Kaplan–Meier method and compared between the VA-ECMO and VV-ECMO subgroups. Results Among 304 included patients (VV-ECMO, 105 [34.5%]; VA-ECMO, 199 [65.5%]), overall in-hospital mortality was 36.2% (110/304), 41.3% with VA-ECMO and 26.2% with VV-ECMO. More than half of deaths (57.3%) occurred within 5 days after ECMO initiation. In the overall cohort, higher APACHE II scores (aOR 1.081, 95% CI 1.028–1.137; p = 0.002), systolic blood pressure (aOR 1.024,95% CI 1.003–1.046; p = 0.027), and lactate (aOR 1.108, 95% CI 1.019–1.205; p = 0.017) independently predicted in-hospital death, whereas diastolic blood pressure (aOR 0.948, 95% CI 0.916–0.980; p = 0.002) and albumin (aOR 0.951, 95% CI 0.905–1.000; p = 0.050) were protective. In subgroup models, APACHE II score, arterial pH, albumin, diastolic pressure, and white cell count were significant for VA-ECMO (AUC 0.845) and lactate and APTT for VV-ECMO (AUC 0.844). Kaplan–Meier curves showed significantly better survival with VV-ECMO than VA-ECMO (log-rank p < 0.001) Conclusions Over half of deaths occurred within 5 days after ECMO initiation. Mode-specific risk profiles support practical stratification and individualized management in emergency ECMO.