This is a multicenter, retrospective unmatched and matched cohort study of patients admitted between April 2006 and December 2013. Generalized logistic mixed-effects models and Cox proportional hazards models were used to determine the association between treatment with ECMO that was started within 3 days of intensive care unit (ICU) admission and ICU and hospital mortality and length of stay, respectively.
A total of 2440 patients who had severe hypoxemic respiratory failure due to various etiologies were included, 46 who received early veno-venous ECMO and 2394 unmatched and 398 matched controls who received conventional ventilation alone. Compared to matched controls, ECMO was associated with a lower odds of ICU (odds ratio [95% confidence interval], 0.30 [0.13-0.67]) and inhospital death (odds ratio 0.30 [0.14-0.67]). In addition, ECMO was associated with longer times to discharge from ICU and hospital (hazard ratio, 0.42 [0.37-0.47] and 0.53 [0.38-0.73], respectively).
In this observational study, use of early ECMO compared to conventional mechanical ventilation alone in patients who had severe hypoxemic respiratory failure was associated with a lower risk of mortality and a longer length of stay.