Transcatheter aortic valve replacement (TAVR) has become an alternative to surgical aortic valve replacement in patients with symptomatic severe aortic stenosis (1–4). However, unlike standard open surgical aortic valve replacement in which the surgical valve is sewn into place, implantation of the transcatheter heart valve can incur the risk of malposition during TAVR, which if significant may lead to severe aortic regurgitation or valve embolization and potentially necessitate implantation of a second prosthesis. According to the Valve Academic Research Consortium-2 (VARC-2) definitions, valve malposition (VM) can be categorized as migration, embolization, or ectopic deployment (5). Despite a low incidence, VM has been related to a higher mortality after TAVR (6, 7). Several predictors of VM in patients undergoing TAVR have been reported, including highly calcified aortic valves and the use of self-expandable (as compared with balloon-expandable) prostheses (6, 7), but the anatomical features associated with VM in self-expandable TAVR have not been thoroughly evaluated and specifically not evaluated in the Chinese population, which tends to have a significantly higher calcium burden (8). Therefore, we aimed to further investigate the anatomical predictors of VM during self-expandable TAVR with the Venus A-Valve.

From April 2016 to February 2020, 203 consecutive patients underwent TAVR with Venus A-Valve in five medical centers across Guangdong, China; 23 (11.3%) had VM, and all of them were valve migration toward the left ventricle. The incidence of VM was lower in patients with tricuspid aortic valve as compared with those with bicuspid aortic valve (7.1 vs. 16.7%, p = 0.032, Appendix 2). A total of 61 patients (30.0%) met the criteria of optimal position of the prostheses. Thus, a total of 84 patients were included in this analysis.

The baseline characteristics are shown in Table 1. The percentage of male was higher (p = 0.039), and the left ventricular ejection fraction was lower (p = 0.033) in patients with VM. The MDCT characteristics were shown in Table 2. In patients with VM, the dimensions of the annulus, LVOT, and STJ were generally larger; the left main coronary artery was higher (p = 0.031); and AL ratio (p = 0.003) was smaller. The procedural characteristics are shown in Table 3. The sizes of prostheses were significantly different between two groups (p = 0.009), and the prosthesis perimeter/LOVT perimeter ratio was lower (p = 0.026) in the malposition group. Correlation analysis showed that all the variables with significant difference between two groups were strongly correlated to either AL ratio or STJ height. Thus, only these two anatomical variables were selected in the stepwise regression, and both of them were found to be independent predictors of VM.

The ROC curve indicated a moderate strength of AL ratio to predict VM [area under the curve (AUC) 0.71, 95% confidence interval (CI) 0.59–0.84, p = 0.003, Figure 2]. The cutoff value of AL ratio was 0.96 with a sensitivity of 70% and a specificity of 71%. The ROC curve showed a weak strength of STJ height to predict VM (AUC 0.69, 95% CI 0.55–0.83, p = 0.011, Figure 2). The cutoff value of STJ height was 23.8 mm with a sensitivity of 52% and a specificity of 85%. The combination of both predictors revealed a higher predictive value of VM (AUC 0.77, 95% CI 0.64–0.89, p < 0.001, Figure 2).

In multivariate analysis (Table 4), AL ratio <0.96 [odds ratio (OR) 3.98, 95% CI 1.31–12.14, p = 0.015] and STJ height ≥23.8 mm (OR 4.63, 95% CI 1.48–14.46, p = 0.008) were both strong independent anatomical predictors of VM. The presence of both predictors was associated with a very high risk of VM (OR 10.67, 95% CI 2.45–46.57, p = 0.002).

There was no mortality at 30 days after TAVR in the included population. The rate of moderate-to-severe paravalvular regurgitation was higher in patients with VM (26.1 vs. 4.9%, p = 0.011) than in those with optimal position of the prostheses. No myocardial infarction was observed. The incidences of disabling stroke, major bleeding, major vascular complications, acute kidney injury (stage 2 or 3), permanent pacemaker implantation, and new-onset atrial fibrillation were all comparable between the two groups (Table 5).

In this multicenter retrospective study of TAVR, we utilized the modified definition of optimal position and malposition of the self-expandable Venus A-Valve; and we compared the baseline, MDCT, and procedural characteristics between those two groups. We determined that AL ratio <0.96 and STJ height ≥23.8 mm were both strong independent anatomical predictors of VM during self-expandable TAVR (Figure 3).

In clinical practice, VM could account for a majority of acute paravalvular regurgitation during TAVR and may necessitate implantation of a second prosthesis (10). However, the anatomical feature associated with VM has not been extensively investigated, particularly in a population with high calcium burden (6, 11). The recent TRAVEL registry has shed some new light on this topic and reported that the self-expandable valve was more susceptible to VM than the balloon-expandable valve in TAVR (7). Importantly, the study of new-generation devices also found a higher rate of implantation of a second valve with the self-expandable Evolut R than with the balloon-expandable Sapien 3 (2.7 vs. 0.5%) (12). The higher risk of VM during self-expandable TAVR advocated a further investigation of the anatomical predictors of VM in this setting.

To our knowledge, this is the first report that the AL ratio and STJ height are related to VM. The presence of both AL ratio <0.96 and STJ height ≥23.8 mm can increase the risk of VM at ~10-fold. Therefore, we propose a concept of the “conical LVOT,” which indicates a larger LVOT than the annulus. It is reasonable that a “conical LVOT” may not provide enough upward support force for the prosthesis during the course of release, which could prevent malposition of the prosthesis to the ventricle. It is also explainable that a higher STJ usually accompanies a larger sinus of Valsalva and therefore has a weaker radial force for the outflow of the prosthesis after release. While only the Venus A-Valve was included in this study, the AL ratio and STJ height might be helpful in evaluating the risk of VM when using other self-expandable valves with similar geometric design and releasing process (e.g., Medtronic CoreValve).

Some potentially important risk factors discovered by previous studies were not reproduced in our study, which could be the results of limited statistical power. On the one hand, bicuspid aortic valve was found to be an independent predictor (OR 3.43, p < 0.001) of VM in the TRAVEL registry (7). Similarly, the rate of bicuspid aortic valve was higher in the malposition group than in the optimal position group in our study, but without significant difference (65.2 vs. 54.1%, p = 0.358). On the other hand, in a sub-study of the PARTNER trial, heavy leaflet calcification was identified as an important cause of valve embolization or valve-in-valve procedure (6), which was not found in our study. As the problem of bicuspid anatomy and severe calcification was prominent in Chinese patients (8), both of these two variables were introduced in the final multivariate regression model to minimize the underlying bias.

The incidence of VM requiring a second prosthesis was around 11% in a Chinese population with Venus A-Valve in this study, which is much higher than that (3~4%) in western population with CoreValve (13, 14). This may be attributed to the more complex anatomy and higher degree of calcification of the aortic root in Chinese population (8) as well as a higher radial force at the inflow section of Venus A-Valve, which could enhance the pushing force downward during releasing of the prosthesis and cause VM in complicated cases (9). It has been reported that the new-generation repositionable self-expandable Evolut R and Evolut PRO have lower rates of valve migration, moderate-to-severe paravalvular regurgitation, and implantation of a second prosthesis than the earlier generation CoreValve (15–17). Similarly, it is anticipated that the second-generation repositionable Venus A-Plus Valve will likely have a better performance in terms of VM (18).

As VM relates to clinical outcome, while there was no mortality at 30 days after TAVR in our study, the rate of moderate-to-severe paravalvular regurgitation was higher in patients with VM, which was in accordance with the TRAVEL registry (7). Although the implantation of a second valve can be a safe and effective therapeutic option (13, 14), VM has been related to increased morbidity and mortality (6, 7). Therefore, we hope that the new anatomical predictors of VM could further improve the outcomes after TAVR.

In this multicenter retrospective study, we found that AL ratio <0.96 and STJ height ≥23.8 mm were strong independent anatomical predictors of VM during self-expandable TAVR in a Chinese population with high risk of VM (because of high percentage of bicuspid anatomy and severe calcification of the aortic valves). These results needed further corroboration in the next-generation self-expandable valves.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Research Ethics Committee of Guangdong Provincial People's Hospital. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

JLi, YS, and SZ contributed equally to study design, data acquisition, statistical analysis, and drafting of the manuscript. JLu approved the submission of the final version. GL, HD, MF, YM, YL, HL, ZX, LZ, YC, RF, and DL contributed greatly to the revision of the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.