Predicting survival with scales for in-hospital extracorporeal cardiopulmonary resuscitation

Abstract

Abstract Funding Acknowledgements None. Background Extracorporeal cardiopulmonary resuscitation (ECPR) is a rescue option in cardiac arrest (CA) refractory to advanced life support. Patient selection is challenging, and anecdotal scores are specific to this situation. Purpose Our objective was to analyse the usefulness of different scales used in shock and venoarterial extracorporeal membrane oxygenator (VA-ECMO) support in in-hospital ECPR. Methods Retrospective analysis of consecutive ECPR cases in a referal center. The ECPR alert includes in-hospital cardiologist/intensivist, on-call interventional cardiologist/cardiac surgeon for cannulation, and on-call perfusionist. We analysed hospital survival based on our own and other literature-based mortality scales calculated prior to VA-ECMO implantation: SCAI, INTERMACS, CardShock, CSS, SAVE-ECMO, RESCUE-IHCA. Results Out of 222 VA-ECMO implanted, we included 66 ECPR cases between 2014-oct 2023. Baseline, admission and ECPR characteristics are collected in the table. Overall in-hospital survival was 16.7% (n=11) (cerebral performance cathegory 1-2). We developed a sumative mortality scale including factors related to survival in our population and the literature (lactate >10.05, >60 minutes CA, peripheral arterial disease (PAD), and on-call time for ECPR alert activation) (figure A). This scale was significantly related to survival to discharge (figure A). Regarding the scales and their association with survival, only RESCUE-IHCA showed a trend towards statistical significance (16 [15] points in survivors vs 19 [15] in died patients, p=0.096). The remaining scores showed no relationship with survival to discharge: SAVE-ECMO -15,73±4.8 in survivors vs -14,07±4.1 in died patients (p= 0.301), CSS 10,09±1.9 vs 10.40±2.7 (p=0.651), CardShock 4.82±1.3 vs 5.27±1.2 (p=0.287). INTERMACS (all patients 1) and SCAI classification (all patients stage E) are not valid for discriminating survival in this situation. We compared the scales using the ROC curve for hospital survival, and we found that our own scale had the best discriminatory ability (Figure B). Conclusions Scales developed for shock or VA-ECMO support were not significantly related to survival in patients undergoing in-hospital ECPR. The specific ECPR scale RESCUE-IHCA, and our own scale developed in this population showed better discriminatory ability in survival to discharge.Table.Baseline, admission and ECPRFigure.A: Own scale. B: ROC curves