Abstract
Emerging evidence suggests prolonged use of noninvasive respiratory support may increase mortality of patients with coronavirus disease 2019 (COVID-19)-associated acute respiratory distress syndrome who require extracorporeal membrane oxygenation (ECMO). Using a database of adults receiving ECMO for COVID-19, we calculated survival curves and multivariable Cox regression to determine the risk of death associated with pre-ECMO use of high-flow nasal oxygen (HFNO), noninvasive ventilation (NIV), and invasive mechanical ventilation (IMV) days. We investigated the performance of a novel variable, advanced respiratory support days (composite of HFNO, NIV, and IMV days), on Respiratory ECMO Survival Prediction (RESP) score. Subjects (N = 146) with increasing advanced respiratory support days (<5, 5-9, and ≥10) had a stepwise increase in 90 day mortality (32.2%, 57.7%, and 75.4%, respectively; p = 0.002). Ninety-day mortality was significantly higher in subjects (N = 121) receiving NIV >4 days (81.8% vs. 52.4%, p < 0.001). Each additional pre-ECMO advanced respiratory support day increased the odds of right ventricular failure (odds ratio [OR]: 1.066, 95% confidence interval [CI]: 1.002-1.135) and in-hospital mortality (1.17, 95% CI: 1.08-1.27). Substituting advanced respiratory support days for IMV days improved RESP score mortality prediction (area under the curve (AUC) or: 0.64 vs. 0.71). Pre-ECMO advanced respiratory support days were associated with increased 90 day mortality compared with IMV days alone. Adjusting the RESP score for advanced respiratory support days improved mortality prediction.