Optimizing Save Score for Intraoperative Complications

Abstract

Veno-arterial extracorporeal membrane oxygenation (VA ECMO) is an effective support therapy for refractory cardiogenic shock and can be used as a bridge to more definitive therapy. However, VA ECMO has high rates of complications and requires considerable financial and human resources. Selecting patients most likely to benefit from ECMO is controversial. The Survival After Veno-arterial ECMO (SAVE) Score is a tool developed by Schmidt et al. to predict survival for patients receiving ECMO therapy for cardiogenic shock. We conducted a single-center, retrospective study to validate the relationship between SAVE scores and survival rates, and to identify additional factors that can predict prognosis. We retrospectively reviewed all cardiogenic shock patients who received VA-ECMO therapy at our institution between 2016 - 2019. We collected patient variables/outcomes, ECMO data, and calculated SAVE scores. We identified 66 cardiogenic shock patients treated with VA-ECMO and stratified them by SAVE score risk categories and measured survival to discharge with the following results: 33.33% (4/12) in category V (SAVE score ≤-10; 18% predicted survival); 29.63% (8/27) in category IV (SAVE score -9 to -5; 30% predicted survival); 36.84% (7/19) in category III (SAVE score -4 to 0; 42% predicted survival); 57.14% (4/7) in category II (SAVE score 1 to 5; 58% predicted survival); 0% (0/1) in category I (SAVE score >5; 75% predicted survival). Overall survival was 34.85% (23/66). When patients with an intraoperative complication causing hemodynamic compromise as the etiology of their cardiogenic shock were isolated (18 total) and a -2 SAVE score modifier applied, the following changes to overall survival patterns were seen: 28.57% (4/14) in category V patients, and 41.18% (7/17) in category III patients. Overall, the survival rates predicted by SAVE score risk categories matched well when applied to the patient population at our institution. Of note, risk category I was not considered in analysis due to limited sample size (n = 1). Applying a negative score modifier to intraoperative complication patients brought measured survival rates closer to predicted, suggesting this etiology of cardiogenic shock portends worse survival than reflected in original SAVE score. Further study to identify patterns of complications and survival may aid in more accurate risk stratification.