By Jeffrey Zimmet, MD, PhD

Associate Professor of Medicine, University of California, San Francisco; Director, Cardiac Catheterization Laboratory, San Francisco VA Medical Center

Dr. Zimmet reports no financial relationships relevant to this field of study.

SYNOPSIS: This large retrospective study of patients undergoing reintervention for failed bioprosthetic aortic valves showed better short-term outcomes with valve-in-valve transcatheter aortic valve replacement vs. redo surgical aortic valve replacement.

SOURCE: Deharo P, Bisson A, Herbert J, et al. Transcatheter valve-in-valve aortic valve replacement as an alternative to surgical re-replacement. J Am Coll Cardiol 2020;76:489-499.

Over time, there has been a progression toward using more bioprosthetic valves instead of mechanical valves in surgical valve replacement. The main drawback of these valves is their limited lifespan, with most valves affected at some point by structural deterioration. The traditional approach to this problem, redo valve surgery, is hindered by complication rates that are consistently well in excess of a first cardiovascular surgery. With the advent of transcatheter aortic valve replacement (TAVR) came early interest in treating failed aortic bioprostheses with valve-in-valve (ViV) TAVR. Several retrospective studies have demonstrated advantages of ViV TAVR over redo surgical aortic valve replacement (SAVR) regarding obvious early outcomes, including bleeding, post-procedural hospital stays, and early mortality. Accordingly, ViV procedures have become more common over time vs. redo SAVR, reflecting increasing enthusiasm for this technique among both cardiologists and surgeons.

Deharo et al reported the results of an analysis of ViV TAVR vs. redo SAVR from a large, administrative database that includes all TAVR and SAVR procedures performed in France. Between 2010 and mid-2019, 4,327 patients were identified as undergoing procedures for failure of aortic bioprostheses, including 1,773 undergoing redo SAVR and 2,554 with ViV TAVR. In the unmatched cohort, patients treated by ViV TAVR were older, were more frail, presented with more comorbidities, and recorded higher expected surgical mortality rates as estimated by EuroSCORE II. After propensity score matching, 717 patients were identified in each group for the comparison.

The 30-day outcomes clearly favored TAVR. All-cause mortality was reported in 26 matched TAVR patients vs. 52 patients undergoing redo SAVR, with cardiovascular death the largest contributor (2.9% vs. 6.6%). The odds ratio for 30-day mortality with TAVR vs. SAVR was 0.48, with a 95% confidence interval of 0.30-0.78. Rates of stroke, myocardial infarction (MI), and major bleeding were not statistically different between groups, while pacemaker rates were higher after TAVR.

Longer-term follow-up was available for a median of 794 days in the redo SAVR group and 786 days in the ViV TAVR group. For the matched sets, there was no significant difference between groups regarding cardiovascular mortality, stroke, MI, and new atrial fibrillation. However, both rehospitalization for heart failure and new pacemaker implantation were more frequent in the TAVR group. The authors summarized their findings by saying short-term outcomes were better with ViV TAVR vs. redo surgery, while major cardiovascular outcomes during longer-term follow-up were not significantly different.


Short-term outcomes are better for TAVR vs. redo open surgery. This is an obvious finding. Even in this analysis dating back to 2010, ViV TAVR outnumbered redo SAVR for the entire dataset, accounting for nearly 60% of all cases. The numbers certainly would be more skewed toward TAVR in more recent years. The outcomes as framed in this paper appear straightforward: ViV TAVR is associated with lower short-term mortality and cardiovascular complications, with long-term outcomes that are reported to be not different. However, this greatly oversimplifies the case.

Closer analysis of the current publication shows in the central illustration exhibiting combined events at long-term follow-up, the lines cross at around 1.5 years. The early advantage of TAVR transitions to a clearly lower event rate for redo SAVR at four years post-intervention, and a non-significant difference for all events reported in the paper.

Hemodynamic data are unavailable in this analysis. Prior publications have shown ViV TAVR is associated with greater degrees of patient-prosthesis mismatch than redo SAVR, with higher transvalvular gradients post-procedure. This may explain the consistent finding, also seen here, of fewer hospitalizations for heart failure among patients undergoing redo SAVR.

Buried in the discussion, with data shown only in the supplemental appendix, is the finding that among lower-risk patients (EuroSCORE < 5), redo SAVR was associated with a strong trend toward lower long-term cardiovascular mortality (2.9% vs. 4.5%; P = 0.06). Notably, these lower-risk patients outnumbered the high-risk patients in the dataset (959 vs. 475). Although the higher-risk patients clearly drove the negative outcomes, lower-risk patients were not rare.

Although ViV TAVR may be the procedure of choice for elderly patients with high surgical risk, the question remains as to whether redo surgery may result in better long-term outcomes for younger patients with fewer comorbidities. The authors of an accompanying editorial argued a randomized trial comparing these two therapies is in order; for now, no such trial is on the horizon. ViV TAVR will continue to grow in use. However, these results provide a note of caution, and should prompt us to consider all options in those patients who would be expected to handle surgery well.