Abstract 19540: US Center Variability in the Incidence of Stroke for the Berlin Heart EXCOR Pediatric Ventricular Assist Device

Abstract

Background: Ventricular assist device (VAD)-related stroke is an important performance benchmark for adult VAD programs. No benchmarks are available for pediatric VAD programs where the stroke rate is higher. We sought to describe center variability in the stroke rate for the Berlin Heart EXCOR® Pediatric VAD, the most commonly used durable VAD in US children, to support pediatric VAD quality improvement efforts. Methods: All children age <18 years supported with a Berlin Heart EXCOR Pediatric VAD at a site participating in the EXCOR IDE or FDA Post-Approval Study were identified using the Sponsor’s database. The incidence of neurological dysfunction, defined by INTERMACS criteria, was estimated for patients at centers with at least 300 support days. Results: Between 2007 and March 2014, a total of 320 children underwent implantation of a Berlin Heart EXCOR Pediatric VAD at 42 centers, of which 271 patients at 25 centers met the study inclusion criteria. The baseline characteristics of the cohort have been reported previously. Overall, there were a total of 64 neurological dysfunction episodes in 53 unique patients over 20,875 days of support, yielding an overall stroke incidence of 0.31 events per 100 days of support and a prevalence of 19%. Centers varied considerably (ten-fold) in their incidence rate of stroke ranging from 0 events to 1.7 events per 100 days of support (Figure). Among 13 centers with at least 10 total implants (totaling 205 patients over 14,405 support days), the overall stroke prevalence was 20% (range 6 to 46%) (Incidence rate 0.33 events per 100 days of support). Centers with ≥10 implants did not have fewer strokes than lower volume centers (P=0.30). Conclusion: The incidence of stroke among pediatric VAD centers using the Berlin Heart EXCOR Pediatric VAD was 0.31 events per day of support and varied considerably across pediatric centers. These data may be useful for developing performance benchmarks for neurological dysfunction in pediatric VAD support.