Validation of the prognostic scoring system for in-hospital mortality prediction in cardiogenic shock patients requiring veno-arterial extracorporeal membrane oxygenation.

Abstract

A few prognostic scoring systems have been developed for predicting mortality in patients with cardiogenic shock requiring veno-arterial extracorporeal membrane oxygenation (VA-ECMO), albeit with variations in performance. This study aimed to assess and compare various mortality prediction models in a cohort of patients receiving VA-ECMO following cardiogenic shock or arrest. We retrospectively analyzed 77 patients with cardiogenic shock who were placed on VA-ECMO support between March 2014 and August 2021. The APACHE II, SAPS II, SAVE, Modified SAVE, ENCOURAGE, and ECMO-ACCEPTS scores were calculated for each patient to predict the in-hospital mortality. Fifty-six (72.7%) patients died. All prediction model scores, except the ECMO-ACCEPTS, differed significantly between non-survivors and survivors as follows: ENCOURAGE, 23 versus 16 (p < 0.001); SAVE, -6 versus -3 (p = 0.008); Modified SAVE, -5 versus 0 (p = 0.005); APACHE II, 32 versus 22 (p = 0.009); and SAPS II, 67 versus 49 (p = 0.002). The ENCOURAGE score demonstrated the best discriminatory ability with an area under the receiver-operating characteristic curve of 0.81 (95% confidence interval: 0.7-0.81). All prognostic scoring systems possessed limited calibration ability. However, the SAPS II, SAVE, and ENCOURAGE scores had lower Akaike and Bayesian information criteria values, which were consistent with the results of the Hosmer-Lemeshow C statistic test, indicating better performance than the other scores. The ENCOURAGE score can help predict in-hospital mortality in all subsets of VA-ECMO patients, even though it was originally designed to predict intensive care unit mortality in the post-acute myocardial infarction setting.