Transcatheter Versus Surgical Aortic Valve Replacement in Low‐Risk Patients: Puzzle Solved?

The approval of transcatheter aortic valve replacement (TAVR) by the US Food and Drug Administration for patients with severe aortic stenosis at low risk for surgical aortic valve replacement (SAVR) after the PARTNER 3 and Evolut Low Risk trials ¹ , ² has led to a paradigm shift in the treatment of aortic stenosis worldwide. In the United States, low‐risk patients made up nearly 12% of all patients undergoing TAVR in 2019. ³ A more recent report showed that TAVR was performed in 87.5% of patients between 65 and 80 years of age and 47.5% of those <65 years of age who required aortic valve replacement for isolated severe aortic stenosis in 2021. ⁴ With this quick uptake of TAVR, longer‐term outcomes in low‐risk patients have been eagerly awaited.

In this issue of the Journal of the American Heart Association (JAHA), Sá and colleagues performed a meta‐analysis of 3 randomized controlled trials and 5 propensity score‐matched studies, involving a total of 5444 patients and comparing midterm outcomes with TAVR versus SAVR in patients deemed at low surgical risk. ⁵ Low perioperative procedural risk of mortality was determined by the Society of Thoracic Surgeons (STS) Risk Model in 6 studies, and the European System for Cardiac Operative Risk Evaluation (EuroSCORE II) in 2 studies. All‐cause mortality, the main and only outcome examined, was similar with either approach at 2 years of follow‐up (hazard ratio [HR], 1.08 [95% CI, 0.89–1.31], P=0.448), but in favor of SAVR beyond 2 years in a landmark analysis (HR, 1.51 [95% CI, 1.14–2.00], P=0.004). Interestingly, improved midterm survival with SAVR was observed only in propensity‐score matched studies but not in randomized controlled trials.

To our knowledge, the current study is the largest pooled analysis to date and is a commendable effort to address a complex question. The complexity of this question stems from multiple aspects. One aspect is the definition of “low risk.” While STS score and EuroSCORE II are widely used and effective, they come with limitations. These scores include chronological age and many relevant clinical comorbidities, but unfortunately, critical variables such as frailty, cognitive impairment, mobility, and social support are not well represented in either score. Frailty was shown in several studies to strongly predict worse outcomes after cardiac surgery ⁶ and TAVR. ⁷ , ⁸ At times, physicians may rely on these factors in their risk stratification and choice of aortic valve replacement strategy among patients within the same surgical‐risk category. Hence, performing risk stratification for patients based solely on these scores in propensity‐matched studies precludes the role of physician judgment and clinical expertise, and is likely to underestimate the perioperative risk. In the study by Schaefer et al, one of the propensity score‐matched studies included in the current meta‐analysis, the authors attributed the inferior midterm survival with TAVR versus SAVR at 5 years to the sicker population undergoing TAVR despite a well‐performed propensity score matching. ⁹ That was also observed in the study from the Aortic Valve Replacement in Elective Patients From the Aortic Valve Multicenter (AVALON) Registry, ¹⁰ where despite a comprehensive propensity‐score matching, patients in the TAVR group had higher EuroSCORE II and poor mobility, defined by the authors as severe impairment of mobility secondary to musculoskeletal or neurological dysfunction. These markers of frailty are key factors in determining not only the early postoperative surgical risk but also long‐term outcomes. In addition, the predictability of mortality can vary with STS score versus EuroSCORE II in different patient populations, ¹¹ and thus examining outcomes in a pooled analysis of studies using different risk scores is suboptimal and should be interpreted with caution.

Another aspect of complexity is making sure we examine outcomes in an adequate sample size that rejects the play of chance. In the current study, the authors performed robust time‐to‐event analysis, however, unfortunately only 2 studies, 1 randomized controlled trial, ¹² and 1 propensity score‐matched study, ¹⁰ provided follow‐up beyond 5 years, with a significant drop in the number of patients in both arms at 6 years (TAVR n=85, SAVR n=83) (Figure). ⁹ , ¹⁰ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷

Figure . Studies included in the meta‐analysis with the corresponding follow‐up (years).

AVALON indicates the aortic valve replacement in elective patients from the aortic valve multicenter registry; EVOLUT Low‐Risk, medtronic evolut transcatheter aortic valve replacement in low risk patients; FinnValve, the nationwide finnish registry of transcatheter and surgical aortic valve replacement for aortic valve stenosis; NOTION, nordic aortic valve intervention; OBSERVANT, observational study of effectiveness of SAVR–TAVI procedures for severe aortic stenosis treatment; PARTNER, placement of aortic transcatheter valve trial; PSM, propensity score‐matched study; and RCT, randomized controlled trial.

Finally, given that mid‐ and long‐term all‐cause mortality after TAVR or SAVR can either be secondary to a cardiovascular or noncardiovascular cause, data about clinical events (eg, heart failure hospitalization, stroke, myocardial infarction) and structural valve function (eg, transvalvular gradient, paravalvular leak, need for reoperation) are critical elements in performing a fair head‐to‐head comparison of both therapies. In the current study, Sá and colleagues took the analysis a step further and performed comprehensive meta‐regression analyses of difference covariates, including valve function such as prosthesis‐patient mismatch and paravalvular leak, and did not find a modulating impact on outcomes. This reassuring finding goes in line with the results of the Nordic Aortic Valve Intervention trial that showed similarly low rates of bioprosthetic valve failure with TAVR versus SAVR at 8 years follow‐up (8.7% versus 10.5%, respectively; P=0.61).

The meta‐analysis by Sá et al ⁵ reports interesting results that add to the ongoing debate, but the complex question is far from being answered. Given the limitations of analysis and variability in outcomes among randomized controlled trials and propensity score‐matched studies, the authors correctly cautioned against definitive conclusions, emphasizing the importance of a heart team approach while we wait for the longer‐term outcomes of randomized controlled trials comparing TAVR versus SAVR in low‐risk patients with severe aortic stenosis.

Disclosures

Dr Saad is a consultant for Boston Scientific. Dr Imran has no disclosures to report.