Longer-Term Follow-up of Self-Expanding Transcatheter Versus Surgical Aortic Valve Replacement in Low Risk Patients: What Does the Data Tell Us? Authors: Samuel Heuts, Mateo Marin-Cuartas, Michael A Borger Published: 2026 The Evolut-LR trial The recently published 6-year results of the Evolut low-risk (Evolut-LR) trial, with additional 7-year exploratory analyses, provide important data on the randomized comparison between self-expanding transcatheter aortic valve implantation (TAVI) and surgical aortic valve replacement (SAVR) in low-risk patients [1]. At first glance, the headline message seems reassuring: there is no statistically significant difference in the primary composite outcome of all-cause mortality and stroke (log-rank P-value = 0.43). Yet, event rates for this primary composite outcome were 23.3% for TAVI and 20.4% for SAVR. Therefore, beneath this apparent equipoise may lie a more nuanced interpretation of the data. Predefined statistical analysis The original 1-year evaluation of the Evolut-LR data employed a Bayesian non-inferiority design [2,3], with a non-inferiority margin of 6% absolute risk difference. Starting from the 3-year analyses onwards, the Bayesian non-inferiority analytical framework was shifted towards a frequentist superiority testing framework, using log-rank tests and univariable Cox proportional-hazards models. While not uncommon in longer-term follow-up of initial non-inferiority trials, this shift may have potentially relevant implications. Both the log-rank test and Cox-models rely on the proportional hazards assumption (ie, the relative risk/hazard between groups should remain consistent over time). Based on the visual inspection of the Kaplan–Meier curves of the Evolut-LR trial (and other TAVI vs SAVR trials [4]), this assumption does not hold up. In such cases, P-values derived from a log-rank test (or Cox-model) are uninterpretable. Methodological guidance on the issue of non-proportional hazards suggests employing alternative strategies in these circumstances, such as the performance of a milestone analysis [5]. Avoiding the assumption of non-proportional hazards The reported 2.8% difference in the primary composite outcome has a 95% CI of –1.9% to 7.6% in the milestone analysis. Using available 7-year data, the difference in mortality increased to 3.7% (95% CI –2.4% to 9.9%). The progression in mortality risk seems exponential in the TAVI group between years 4 and 7; from 9.0 to 27.7% (+18.7%) versus from 12.1 to 23.9% (+11.8%) for SAVR. This increasing divergence in event rates is a cause for concern, and shifts the evidentiary balance from perceived early reassurance [2] towards longer-term uncertainty. Re-interventions as signal of late structural failure Perhaps the most striking finding within the Evolut-LR publication is the statistically significant difference in re-interventions, markedly disfavouring the TAVI-group, and increasing with time (6 years: 5.5% vs 3.3%, 7 years: 9.8% vs 6.0%, P = 0.02). It is also important to note that the interventional approach used for these re-interventions was surgical in 65% of patients (ie, an explant-TAVI procedure). Surgical explants of TAVI-prostheses, the most rapidly growing cardiac surgical interventions, are associated with a disproportionally high risk of morbidity and mortality. Together, these findings raise important concerns regarding the long-term durability of SE TAVI using the Evolut platform, and the eventual impact of these re-interventions on patients’ prognosis. The authors note that post-dilation practices, particularly off-guidance balloon oversizing, may have resulted in substantial damage to SE valve leaflets [1]. Still, post-implantation dilatation is performed for reasons such as paravalvular leakage or incomplete TAVI stent unfolding, variables that are associated with worse long-term prognosis if left untreated. Therefore, these practices seem to be an inherent requirement in SE TAVI, particularly in anatomically less favourable circumstances. Consistency across trials The patterns in event rates as observed in Evolut-LR mimic those of previous intermediate-risk and low-risk trials [4]; crossing of survival curves in the longer-term, after an initial phase in which TAVI seems particularly beneficial, followed by subsequent progressive divergence. Figure 1 presents reconstructed all-cause mortality curves from all available randomized literature at 7-year follow-up (Evolut-LR, PARTNER-3, NOTION). These temporal dynamics reinforce the need for statistical approaches that respect time-dependent treatment effects, rather than compressing them into a single hazard ratio (or evaluating them by a log-rank test). Moreover, it is imperative that the late divergence in event rates is quantified properly, and the underlying clinical reasons are elucidated as well. Pooled analysis of reported 7-year follow-up data. The analysis is based on reconstructed 7-year all-cause mortality data (pseudo-individual patient data derived from Kaplan-Meier curves) from Evolut-LR, PARTNER-3, and NOTION. Abbreviations: SAVR, surgical aortic valve replacement; TAVI, transcatheter aortic valve implantation Implications for the Heart Team In younger and low-risk patients with substantial life expectancy, the trade-off between early procedural recovery and late structural risk (and reintervention) must be deliberated in detail. It is also paramount that trials continue their surveillance for at least 10 years in the ongoing low-risk TAVI vs SAVR trials, before the indication for TAVI is formally expanded to low-risk and young patients. Conclusion The Evolut-LR presents much-anticipated data on the longer-term follow-up of TAVI vs SAVR therapy. The totality of available long-term data suggests that the long-term story on the optimal management of aortic stenosis is not yet finished. Stay tuned.