Optimal oversizing degree of self-expanding transcatheter valve in patients with bicuspid vs. tricuspid aortic stenosis

To the Editor: Transcatheter aortic valve replacement (TAVR) is a well-established treatment for severe aortic stenosis (AS) across all surgical risk categories, with an increasing proportion of candidates having bicuspid aortic valve (BAV).[1–3] BAV anatomy, characterized by elliptical annulus, asymmetrical cusps, heavy calcification, and variable coronary origins, poses unique challenges in TAVR procedures. These anatomical features contribute to higher rates of complications such as stroke, paravalvular regurgitation (PVR), and permanent pacemaker implantation (PPI), despite advances in transcatheter heart valve (THV) technology.[4] Among the complications, significant PVR is associated with worse clinical outcomes even for mild cases, while new-onset conduction disturbances (NOCD) after TAVR are linked to increased mortality and impaired ventricular recovery.[5,6] To reduce both PVR and NOCD, optimal sizing of THVs is critical. However, current guidelines for sizing remain limited, particularly for BAV patients, and the impact of annular oversizing ratio (OSr) strategy remains controversial. To address this issue, we investigated the association between computed tomography (CT)-based annular OSr and the incidence of PVR and NOCD in patients undergoing self-expanding valve (SEV) implantation, aiming to identify the optimal OSr strategy that balances the risks of PVR and NOCD in bicuspid and tricuspid aortic valve (TAV) anatomies. This single-center, retrospective study included consecutive patients with severe AS who underwent TAVR using SEVs at the Second Affiliated Hospital, Zhejiang University School of Medicine between March 2013 and December 2020. A total of 580 patients met the inclusion and exclusion criteria included in the analysis [Supplementary Figure 1, https://links.lww.com/CM9/C737]. All procedures were approved by the Medical Ethics Committee of the Second Affiliated Hospital of Zhejiang University School of Medicine (approval No. 2016[032]), and all patients provided written informed consent. Baseline and follow-up assessments included transthoracic echocardiography and 12-lead electrocardiogram at baseline, discharge (or 30 days post-procedure), and annually thereafter. The severity of PVR was evaluated according to a previously described 3-class grading scheme (0 = none or trace, 1 = mild/mild-to-moderate, 2 = moderate/moderate-to-severe, and 3 = severe).[7] For analytical purposes, these grades were reclassified into two clinically meaningful binary categories: ≥mild or ≥moderate. NOCD were defined as conduction disturbances persisting for more than 30 days post-procedure. Pre-procedural multidetector CT (MDCT) scans (SOMATOM Definition Flash, Siemens Medical Solutions, Germany) were analyzed using 3mensio software (Bilthoven, the Netherlands) to assess aortic root anatomy. Bicuspid morphology was classified using Sievers’ criteria.[8] The OSr was calculated as: OSr = (prosthesis nominal perimeter – annulus perimeter)/prosthesis nominal perimeter × 100. Multivariable logistic regression analyses were performed to identify predictors of ≥mild PVR and new-onset atrioventricular block (AVB). Receiver operating characteristic (ROC) curve analyses were used to determine the optimal OSr cutoffs for each endpoint, with comparisons of the area under the curve (AUC) between BAV and TAV patients. Detailed methods are described in the Supplementary Material, https://links.lww.com/CM9/C737. The median age of patients was 75 (interquartile range [IQR]: 70–80) years, with 51.9% (301/580) male. Patients with BAV accounted for 67.1% (389/580) of the cohort, comprising 215 type 0, 161 type 1, and 13 type 2 [Supplementary Table 1, https://links.lww.com/CM9/C737]. The median OSr was 8.1% (IQR: 3.0–12.4%), and participants with larger valve sizes had greater OSr, with median OSr values of 5.1% for 23 mm, 7.8% for 26 mm, 9.7% for 29 mm, and 10.7% for 31/32 mm valves (P <0.001) [Supplementary Tables 1–2, https://links.lww.com/CM9/C737]. A total of 570 patients with echocardiography data were included for PVR assessment [Supplementary Figure 1, https://links.lww.com/CM9/C737]. At 30 days, ≥mild PVR occurred in 54.2% (309/570) of patients, with 8.6% (49/570) having ≥moderate PVR. The highest incidence of ≥mild PVR (71.4% [55/77]) occurred in patients with undersized THVs (OSr <0). Lower incidence of ≥mild PVR (P <0.001) and ≥moderate PVR (P <0.001) were found in patients with larger OSr [Supplementary Table 3 and Supplementary Figure 2, https://links.lww.com/CM9/C737]. We included 571 patients with electrocardiogram data for NOCD assessment [Supplementary Figure 1, https://links.lww.com/CM9/C737]. New-onset AVB occurred in 15.2% (87/571) of patients, with a higher incidence in the group with larger OSr (P = 0.021), particularly in BAV patients [Supplementary Tables 3 and Supplementary Figure 2, https://links.lww.com/CM9/C737]. No significant correlation was found between OSr and the incidence of new-onset left-bundle branch block (LBBB) or PPI requirement (P = 0.247 and P = 0.141, respectively; Supplementary Tables 3, https://links.lww.com/CM9/C737). BAV patients had a higher rate of ≥mild PVR than TAV patients, though this difference was not statistically significant (56.8% [216/380] vs. 48.9% [93/190], P = 0.075). The incidence of new-onset AVB (14.7% [56/382] vs. 16.7% [31/189], P = 0.586) was comparable between BAV and TAV groups [Supplementary Table 4 and Supplementary Figure 2, https://links.lww.com/CM9/C737]. In the multivariate logistic regression models, BAV was associated with ≥mild PVR (adjusted odds ratio [OR]: 2.130, 95% confidence interval [CI]: 1.184–3.832, P = 0.012), but not related to new-onset AVB (adjusted OR: 1.136, 95% CI: 0.665–1.941, P = 0.640. Moderate or greater annular calcification was an independent predictor of ≥mild PVR (adjusted OR: 1.637, 95% CI: 1.137–2.355, P = 0.008), while peripheral artery disease was independently associated with new-onset AVB (adjusted OR: 1.917, 95% CI: 1.031–3.567, P = 0.040; Supplementary Tables 5–6, https://links.lww.com/CM9/C737). Multivariate logistic regression models identified that OSr was not independently associated with PVR after adjustment for confounding factors (adjusted OR: 1.050 per 5% OSr, 95% CI: 0.894–1.299, P = 0.651). Conversely, OSr was associated with increased risk of new-onset AVB (adjusted OR: 1.250 per 5% OSr, 95% CI: 1.039–1.504, P = 0.018; Supplementary Tables 5–6, https://links.lww.com/CM9/C737). In BAV patients, each 5% increase in OSr was associated with a 27% reduction in ≥mild PVR (adjusted OR: 0.732, 95% CI: 0.617–0.869, P <0.001), but a 47% increase in new-onset AVB (adjusted OR: 1.473, 95% CI: 1.154–1.881, P = 0.002; Supplementary Tables 7–8, https://links.lww.com/CM9/C737). In TAV patients, no significant association was found between OSr and ≥mild PVR (adjusted OR: 1.032 [95% CI: 0.821–1.298]; P = 0.787) and new-onset AVB (adjusted OR: 1.077 [95% CI: 0.771–1.503], P = 0.655; Supplementary Tables 9–10, https://links.lww.com/CM9/C737). The optimal OSr cutoff for predicting ≥mild PVR was 7.37% overall (AUC 0.621, P <0.001) and 7.09% in BAV patients (AUC 0.647, P <0.001). For new-onset AVB, the optimal cutoff was both 5.30% overall (AUC 0.565, P = 0.053) and in BAV patients (AUC 0.595, P = 0.011). No significant cutoff was identified in TAV patients for predicting either ≥mild PVR (P = 0.251) or new-onset AVB (P = 0.825). Furthermore, predictive performance was comparable across subgroups stratified by THV skirt design, prior atrial fibrillation, and gender [Supplementary Table 11, https://links.lww.com/CM9/C737]. The main results are presented in Figure 1.Figure 1: Optimal oversizing strategy of self-expanding transcatheter aortic valve in AS patients with BAV versus TAV: Balancing PVR and AVB Risk. (A) Study population flowchart. (B–C) Incidence of PVR and NOCD in overall, BAV, and TAV cohort. (D) Perimeter-based annular OSr inversely correlates with PVR and positively with AVB incidence, particularly in BAV patients. (E) Risk factor for ≥mild PVR and new-onset AVB, showing a significant interaction between BAV morphology and OSr in predicting PVR. (F) In BAV patients, an OSr of 7.09% optimally predicted ≥mild PVR, and an OSr of 5.30% optimally predicted new-onset AVB. No optimal OSr was identified for TAV patients. AS: Aortic stenosis; AVB: Atrioventricular block; BAV: Bicuspid aortic valve; NOCD: New-onset conduction disturbances; ns: Not significant; OSr: Oversizing ratio; PVR: Paravalvular regurgition; TAV: Tricuspid aortic valve; TAVR: Transcatheter aortic valve replacement; THV: Transcatheter heart valve.Our study demonstrates that OSr inversely correlates with PVR severity and positively correlates with new-onset AVB, particularly in BAV patients. Specifically, we found that BAV patients require a more conservative oversizing strategy than TAV patients to reduce PVR risk while minimizing AVB. A 5% increase in OSr in BAV patients was associated with a 27% reduction in ≥mild PVR but a 47% increase in AVB, indicating the need for a balanced approach to valve sizing in this population. Early-generation THVs were associated with relatively high rates of PVR (10–15%). However, advances in CT-based pre-procedural planning and new-generation devices have reduced the incidence of significant PVR to below 5%. In our cohort, the incidence of ≥mild and ≥moderate PVR was 54.2% and 8.6%, respectively, likely influenced by the high proportion of BAV patients, the use of early-generation devices, and a broader patient risk spectrum that included both inoperable and low-risk patients. In this study, 8.9% (34/380) BAV patients developed ≥moderate PVR, registering better outcomes than previous reports. This finding is consistent with recent studies showing reduced PVR in BAV patients treated with newer-generation devices, likely attributable to advances such as repositionability and sealing skirts. Nevertheless, PVR remains a major challenge, as even mild PVR has been associated with adverse long-term outcomes, underscoring the importance of early recognition and intervention. Our study further identified moderate or severe annular calcification as an independent predictor of ≥mild PVR. Moreover, the relationship between BAV and OSr highlights the distinct pathophysiological mechanisms in BAV versus TAV patients. Whereas aortic valve calcification volume was the strongest predictor of PVR in TAV patients, PVR in BAV patients was more strongly influenced by factors such as left ventricular outflow tract calcification, THV oversizing, and malpositioning. New-onset AVB occurred in 15.2% (87/571) of our cohort within 30 days. Higher OSr was correlated with increased incidence of AVB, particularly in BAV patients. Consistent with previous studies, these findings support the hypothesis that radial forces exerted by THVs on surrounding conduction tissues may contribute to the development of AVB. In contrast, no such correlation was found in TAV patients, likely due to the lower event rates and differences in device design. Optimal THV sizing is critical to minimize complications such as PVR and AVB. Our findings support current recommendations for a 12–25% OSr for SEVs. In BAV patients, we identified an optimal OSr of ≥7.09% to minimize ≥mild PVR and ≤5.30% to reduce new-onset AVB. In TAV patients, no optimal OSr could be determined, likely reflecting the low incidence of complications. This study has several limitations. First, only SEV recipients were included, which may limit the generalizability of the findings. Second, baseline differences between BAV and TAV patients, together with the relatively small sample size, restricted robust statistical adjustment. Third, oversizing was assessed solely by annular perimeter, without accounting for other relevant factors such as calcification distributions, septal or membranous septum characteristics, implantation depth, or supra-annular geometry. Finally, follow-up was limited to 30 days, precluding long-term outcome assessment. In conclusion, this study provides important insights into the relationship between annular OSr and the incidence of PVR and AVB in TAVR patients. In BAV patients, a relatively conservative oversizing strategy appeared favorable, with an optimal OSr of 7.09% for minimizing ≥mild PVR and 5.30% for reducing AVB. No clear optimal OSr cutoff was identified for TAV patients. Further prospective studies are warranted to validate these findings, refine valve sizing strategies, and evaluate their long-term clinical impact, particularly for BAV patients. Funding This work was supported by Noncommunicable Chronic Diseases-National Science and Technology Major Project (No. 2024ZD0527100), State Key Laboratory of Transvascular Implantation Devices, and the National Natural Science Foundation of China (Nos. 82271606 and 82425007). Conflicts of interest None.