Impact of self-expanding valve selection on hemodynamics and clinical outcomes in patients with borderline annular dimensions for balloon-expandable valve: insights from the multicenter-TAVI registry

Background Optimal valve size selection for balloon-expandable transcatheter heart valves (BEVs) in borderline annulus cases remains uncertain. We hypothesized that self-expanding valves (SEVs) may provide more favorable hemodynamic and clinical outcomes in patients with borderline annulus dimensions for BEV sizing. This study aimed to compare the hemodynamic performance and clinical outcomes of BEVs and SEVs in this population. Methods From the multicenter TAVI registry, we analyzed patients who underwent TAVI between 2010 and 2025 (n = 5,797). Borderline annulus for BEVs was defined as an estimated BEV oversizing rate ≤–5% or ≥30%. Propensity score matching was performed between BEV and SEV cohorts, adjusting for age, sex, body surface area, annulus area, annulus perimeter, and STS score. Outcomes were compared between the BEV group (SAPIEN 3 Ultra RESILIA, S3UR) and the SEV group (Evolut Pro/Pro+/FX/FX+). Results Among eligible patients, 964 (36.4%) met the borderline criteria and were treated with either BEV (41.1%) or SEV (58.8%). Mean age was 84 years, and 40% were male. At baseline, the SEV group had a smaller pre-procedural aortic valve area and higher mean gradient, suggesting more advanced disease severity. Device success (90.7% vs. 83.7%, p = 0.008) and early safety (94.9% vs. 89.2%, p=0.012) were significantly higher in the BEV group (figure). Post-procedural mean pressure gradient was significantly lower in the SEV group (8.8 vs. 7.3 mmHg, <0.001), whereas ≥mild paravalvular leak was more frequent. Permanent pacemaker implantation was also more common in the SEV group (1.9% vs. 5.4%, p=0.016). Rates of vascular complications, bleeding, and annular rupture were comparable between groups. Two-year mortality and heart failure hospitalization were similar (17.0% vs. 16.8%, log-rank p = 0.37). In the small annulus subgroup (annulus area 312–345 mm²), the SEV group demonstrated a lower incidence of mean gradient ≥20 mmHg (3.1% vs. 9.9%, p=0.007) and Vmax ≥3.0 m/s (3.6% vs. 10.6%, p=0.009), with a trend toward higher device success (92.3% vs. 86.6%, p=0.10). Conclusions In patients with borderline annulus dimensions for BEV sizing, SEVs were associated with superior hemodynamic performance but higher rates of paravalvular leak and pacemaker implantation, without differences in mid-term clinical outcomes. Selection bias due to preferential use of SEVs in anatomically or clinically more complex cases may have influenced these findings.