Perioperative Morbidity and Outcomes of Transitioning from ECMO to VAD Support in Pediatric Patients: A STS Study

Abstract

<h3>Purpose</h3> As a bridge to transplant strategy, children transitioned from ECMO to VAD (ventricular assist device) have higher waitlist mortality compared to those who receive de novo VAD. However, the reasons and timing for such additive morbidity and mortality are unclear. We therefore studied the acute perioperative morbidity and mortality in de novo VAD (group 1) to those transitioned from ECMO to VAD (group 2) using the comprehensive data from The Society of Thoracic Surgeons (STS) database. <h3>Methods</h3> Pediatric patients receiving VAD between 2014-19 reported to STS database were identified and selected based on surgical codes and device specifications. Patients were then divided into group 1 and 2 and perioperative data analyzed using appropriate statistical tests. <h3>Results</h3> A total of 735 children underwent VAD placement with 498 in group 1 and 237 in group 2. Patients in group 2 were significantly younger, smaller, were in cardiogenic shock or have CPR. Patients in group 2 also were significantly likely to have hepatic, renal and neurologic dysfunction at baseline and be three times as likely to be on ventilator (all p <0.001, Table 1). Group 2 patients were twice as likely to transition to biVAD compared to group 1. Post-operative course was significant for unplanned reoperations in 8% group 1 versus 34% in group 2 (p<0.001). VAD related complications were similar after VAD implantation in both groups except intracranial bleeding in 4% for group 1 while 13% for group 2 (p<0.01). Overall mortality was 16% for group 1 and 34% for group 2 (p<0.01). Regression analysis showed that ECMO use (OR 2.17), ventilator (OR 2.2), cardiogenic shock (OR1.8) were all independent preoperative predictors of VAD mortality while renal failure, stroke and bleeding were independent post-operative predictors of VAD mortality within 30 days. <h3>Conclusion</h3> There are significant baseline differences in the cohorts undergoing de novo VAD compared to those transitioned from ECMO, however the need for such ECMO is an important predictor of early post VAD mortality. Therefore, early and elective VAD placement in this cohort of patients should be sought to avoid high post-VAD mortality. Comparison of these two strategies should be avoided without accounting for baseline differences while optimal type and timing of support is researched.