Case Volume and Outcomes After TAVR With Balloon-Expandable Prostheses: Insights From TVT Registry

Background: Given conflicting findings of previous studies, much remains to be understood regarding a volume-outcomes relationship (VOR) in transcatheter aortic valve replacement (TAVR). Objectives: The purpose of this study was: 1) to determine if, after the initial learning curve (LC), a VOR for balloon-expandable (BE) TAVR persisted; and 2) to determine if LCs and VORs differed across different device generations. Methods: Data collected by the TVT registry for BE valve implants from November 2011 through January 2017 were included in this analysis (n = 61,949). Primary outcomes included 30-day all-cause mortality, stroke, and major vascular complications. For each center, all implants were ordered chronologically according to case sequence number (CS#). To determine where the learning curve terminated (LCT), a grid search analysis was applied across a range of CS# from 10 to 300 by increments of 1. After LCT, the VOR was assessed by examining case volume/month by center. This analysis was performed separately for: 1) all BE valve types; 2) Sapien 3 (S3) only; and 3) S3 in BE valve naïve sites. Results: In experience with all commercially available BE valve types, there was an initial LC that terminates around case #201. After the initial LC, a volume–outcomes relationship was no longer evident. In analysis limited to S3, there was no demonstrable LC or VOR. Likewise, there was no demonstrable LC or VOR with S3 for BE valve naïve sites. Conclusions: After a case experience of 200 cases, there was LCT; subsequent to initial learning, a VOR was no longer evident. In the S3-only analysis, there was no LC or no demonstrable VOR. With current-generation BE-TAVR, centers should expect to achieve consistently excellent outcomes even during early case experience.