Reoperative Surgical Aortic Valve Replacement Versus Transcatheter Valve-in-Valve Replacement for Degenerated Bioprosthetic Aortic Valves

Abstract

Background: Bioprosthetic aortic valve use has increased steadily according to The Society of Thoracic Surgeons (STS) database analyses. One of the momentums toward this trend is the future utilization of transcatheter valve-in-valve (TViV) techniques when bioprosthetic valves fail. We compared the results of reoperative TViV to surgical aortic valve replacement (SAVR) for degenerated bioprosthetic valves.

Methods: From January 2002 to January 2015, we identified 91 patients with degenerated bioprosthetic valves who underwent isolated AVR (SAVR n = 69, TViV n = 22). Patients with prior homografts or active endocarditis were excluded. The STS risk score was used to create 22 matched pairs of SAVR and TViV for comparison.

Results: Before matching, mean STS risk scores were 4.36 ± 3.1 and 7.54 ± 3.0 for SAVR and TViV, respectively (p = 0.001), but were 7.70 ± 3.4 and 7.54 ± 3.0, respectively (p = 0.360), after matching. Mean age was 74.5 ± 10.4 years for SAVR and 75.0 ± 9.6 years for TViV (p = 0.749). Operative mortality was 4.3% (1 of 22) in the SAVR group and zero for TViV (p = 1.00). Mean postoperative gradient was 13.5 ± 13.2 mm Hg for SAVR and 12.4 ± 6.2 mm Hg for TViV (p = 0.584). There was no coronary obstruction in either group, but 22% of TViV (5 of 22) had mild paravalvular leaks versus none in the SAVR group (p = 0.048). Postoperative stroke rate was 9% (2 of 22) for SAVR and zero for TViV (p = 0.488). The TViV group had shorter median length of stay (5 versus 11 days, p = 0.001). Actuarial survival at 3 years was 76.3% (95% confidence interval: 58.1 to 94.5) versus 78.7 (95% confidence interval: 56.2 to 100) for SAVR and TViV, respectively (p = 0.410).

Conclusions: For degenerated bioprosthetic aortic valves, TViV has similar operative mortality, strokes rates, and survival as SAVR in this high-risk cohort. Therefore, TViV is a viable alternative to SAVR, although studies using registry data are needed to establish noninferiority.