An association between early rescue left heart unloading during VA-ECMO treatment and prognosis of cardiogenic shock; the post-hoc study of the EARLY-UNLOAD trial

Abstract

Abstract Funding Acknowledgements None. Background/introduction Recently published studies showed prophylactic left heart unloading does not improve mortality in cardiogenic shock (CS) treated with venoatrial extracorporeal membrane oxygenation (VA-ECMO). However, significant proportion of patients on VA-ECMO treatment for CS require early rescue left heart unloading. Purpose To compare characteristics and prognosis according to rescue-unloading timing in patients with CS treated with VA-ECMO. Methods This study is a post-hoc analysis of the EARLY-UNLOAD trial (NCT04775472), the trial evaluated an impact of an early prophylactic unloading strategy using left atrial cannulation (LAC) within 12 hours from the initiation of VA-ECMO treatment. This analysis was performed per-protocol basis for patients without violation (N = 108). Patients were classified into four groups according to timing of left heart unloading: prophylactic (<12h, N = 53), early rescue (<30h, N = 16), late rescue (>30h, N= 10), and no unloading (N = 29). The primary endpoint was in-hospital mortality, and secondary endpoint was 30-day all-cause death. Logistic regression was used to search factors related to the primary endpoint. Results Among patients in the conservative strategy arm, rescue unloading was performed in 47% of patients, and 29% of patients received left heart unloading within 30 hours. Median time from ECMO pump on to rescue unloading was 23 hours. Compared to other groups, the early rescue group had the highest initial arterial lactate level (6.2 vs 11.4 vs 4.3 vs 6.3 mmol/L, prophylactic, early rescue, late rescue, and no unloading, respectively, P = 0.011) and SAPS2 score at baseline (52.7 vs 68.2 vs 50.6 vs 55.7, P = 0.017). Inotropes requirement on day 1 was also highest in the group (vasoactive-inotropic score, 12.4 vs 31.2 vs 7.5 vs 13.8, P = 0.014). The early rescue unloading group showed the highest in-hospital mortality (Figure 1, 87.5%, P = 0.004) and 30-day mortality (75%, P = 0.029). Survival analysis using Kaplan-Meier analysis and log-rank test showed significant difference of probability of all cause death at 30 days among these 4 groups (Figure 2, P by log-rank test = 0.049). In the logistic regression models for in-hospital mortality, rescue unloading within 30 hours was associated with increased odds ratio (OR) for in-hospital mortality (OR = 6.87, 95% confidence interval [CI] = 1.31-35.92.15, P = 0.022). Conclusions In this post-hoc analysis of the EARLY-UNLOAD trial, early-rescue unloading within 30 hours was associated with increased in-hospital mortality in patient treated with VA-ECMO for CS.