Association between annular eccentricity and initial post-deployment paravalvular leak before corrective maneuvers in...

The relationship between annular eccentricity and paravalvular leak (PVL) after transcatheter aortic valve replacement (TAVR) in bicuspid aortic valve (BAV) patients remains insufficiently defined, particularly for Chinese self-expanding valve platforms. This study evaluated the association between the CT-derived eccentricity index (EI) and initial post-deployment moderate-or-greater PVL after self-expanding TAVR in BAV patients.

We retrospectively analyzed 194 consecutive BAV patients who underwent TAVR with the first-generation self-expanding VenusA valve. The primary endpoint was moderate-or-greater PVL assessed immediately after valve release and before corrective maneuvers such as post-dilatation. EI was measured by two independent observers with excellent interobserver agreement (ICC = 0.943). Multivariable logistic regression, ROC analysis, sensitivity analyses, and bootstrap internal validation were performed. NRI and IDI were analyzed as secondary measures of incremental information.

Initial post-deployment moderate-or-greater PVL occurred in 62 patients (32.0%). Although EI was modestly higher in patients with moderate-or-greater PVL than in those with no/mild PVL (median 0.23 [IQR 0.19-0.26] vs. 0.20 [0.17-0.25]; P = 0.005), its stand-alone discrimination was poor (AUC = 0.625; 95% CI 0.539-0.711). After adjustment for clinical and anatomical covariates, EI ≥0.19 remained independently associated with moderate-or-greater PVL (OR = 2.99; 95% CI 1.44-6.23; P = 0.0034), whereas semi-quantitative leaflet calcification and raphe calcification were not. Adding EI increased the C-index only marginally (apparent 0.572 → 0.666; optimism-corrected 0.609), indicating limited overall model discrimination. Predischarge transesophageal echocardiography was available for all 194 patients: moderate-or-greater PVL fell from 62 patients (32.0%) at deployment to only 6 (3.1%) at predischarge, and 57 of the 62 patients with initial moderate-or-greater PVL (91.9%) had regressed to mild or less (paired Wilcoxon signed-rank, P < 0.001), underscoring that initial post-deployment PVL does not equate to clinically relevant residual PVL.

In BAV patients undergoing TAVR with the first-generation self-expanding VenusA valve, EI ≥0.19 was independently associated with initial post-deployment moderate-or-greater PVL, but its stand-alone and model-level discriminative performance were limited. EI should be considered a supplementary dimension in integrated preprocedural risk assessment rather than an isolated decision-making criterion.