Echocardiographic evaluation of ventricular performance in VA-ECMO weaning: results from a observational multicentric prospective study

Abstract Funding Acknowledgements None. Background Although temporary mechanical circulatory support devices are increasingly used in the management of cardiogenic shock (CS), several open question regarding the kind of support and timing of institution/weaning remains open. Echocardiography is widely applied to drive the weaning process from VA-ECMO although there are no standardized protocols. Purpose We sought to evaluate the accuracy of echocardiographic indices in predicting successful weaning from VA-ECMO in CS patients. We tested total isovolumic time (tIVT), an index of ventricular performance(1,2), and longitudinal function (MAPSE) on top of the commonly used parameters (left ventricular ejection fraction, LVEF; E/e’ at mitral valve; left ventricular velocity time integral, LV VTI; TAPSE; right ventricular S’ at tissue Doppler imaging, RV S’). Methods Multicentric prospective study on adults receiving VA-ECMO due to refractory CS and/or cardiac arrest (CA) from 2012 to 2020. Echocardiography was performed at VA-ECMO placement and during the weaning trials. Waning trials were performed daily starting from 48 hours after VA-ECMO institution. A stepwise decrease of VA-ECMO blood flow was applied, until reaching 1 or 1.5 L/min. Successful weaning was defined as removal of VA-ECMO with no requirement for further MCS in the following 30 days. Results 76 patients were enrolled (23.8% female, 54.2±10.6 years-old), 52.6% due to CA. Median duration of VA-ECMO was 58 hours; 47 (61.8%) had intra-aortic balloon pump (IABP) inserted upfront. In-hospital mortality was 62.4% overall, 40% for CS without CA. 6 of the 32 weaned patients died in hospital for multiorgan failure due to septic shock. The baseline echocardiographic parameters associated with successful weaning were (Figure 1-A): stroke volume (SV) >13.2 ml (HR 4.12 [1.71-9.95], p 0.002), tIVT >23.6 min/sec (HR 0.14 [0.02-1.03], p 0.053), LV VTI >6.6 cm (HR 4.99 [1.94-12.87], p<0.001), MAPSE >6.15 mm (HR 4.42 [1.90-10.24], p 0.001), TAPSE >11 mm (HR 10.13 [1.37-7.02], p 0.023), RV S’ >7 cm (HR 5.40 [1.45-20.12], p 0.012). LVEF, E/e’ and S’ at mitral valve did not reach the statistical significance. Figure 1-B shows Kaplan-Meier analysis on echocardiographic indices collected during the last weaning trial; figure 2 details HR of all these parameters. Conclusions Echocardiography is pivotal in defining cardiac recovery and monitoring the weaning trial. Indices reflecting ventricular performance and delivered flow from LV were associated with weaning success both at VA-ECMO institution and throughout the weaning trials. Parameters reflecting ventricular activation and systo-diastolic interaction already demonstrated superiority in defining ventricular performance. Considering the known limitations of LVEF (HR, loading and LV dyssynchrony dependence), tIVT and MAPSE should be introduced in the standard evaluation in the acute setting.