The world’s population is aging and, in parallel, prolonged life expectancy after cardiovascular surgery with the limited durability of bioprosthetic valves has actively contributed to a continuous increase in the prevalence of symptomatic severe aortic stenosis and the number of patients with failed bioprosthetic aortic valves (1–3). Accordingly, the number of candidates for cardiac reinterventions is expected to increase in the near future. Transcatheter aortic valve replacement (TAVR) was initially introduced as a less invasive approach for high-surgical-risk aortic stenosis patients. It was then proposed as a safe alternative therapeutic option for those who are at intermediate and even low risk of needing cardiac surgery (4–6). A few years ago, TAVR was also revealed as an important treatment for bioprosthetic structural valve degeneration. To date, there have been no randomized clinical trials that compare the outcomes and survival outcomes of redo-surgical aortic valve replacement versus TAVR in patients with failed bioprosthetic valves. Available data in literature are limited to observational retrospective cohorts, registries, and meta analysis. An advantage of valve-in-valve TAVR (VIV-TAVR) over repeat surgical aortic valve replacement in early survival outcomes and post-survival complications has been reported in patients with degenerated bioprosthetic valves (7–9). However, higher rates of myocardial infarction, hospital readmissions, and patient-prosthesis mismatch have been observed after VIV-TAVR (7–9). Based on the latest guidelines, VIV-TAVR is considered a complex intervention associated with the high risk of per-procedural and post-procedural complications and should be performed by an experienced heart team following a multidisciplinary decision (10, 11). In the present study, we compare the in-hospital outcomes and long-term survival outcomes of VIV-TAVR patients versus native valve TAVR (NV-TAVR) patients.

Of the 1,658 patients included in this study, 69 (4.2%) study participants underwent TAVR for failed surgically implanted bioprosthetic aortic valves (VIV-TAVR group), while 1,589 (95.8%) study participants underwent TAVR for severe aortic stenosis with native aortic valves (NV-TAVR group). The brands of valves implanted during the index cardiac surgery were Mitroflow (N = 30, 43.5%), Carpentier Edwards (N = 26, 37.7%), St Jude Trifecta (N = 6, 8.7%), Hancock (N = 2, 2.8%), and Mosaic (N = 1, 1.4%) (Figure 1). The size and brand of degenerated bioprosthesis were unidentified in four study participants (5.8%). Characteristics of the study population are shown in Table 1. The mean age of the study population was 83.9 ± 7.1-years and 50.2% of participants were women. Pre-TAVR coronary angiography revealed significant coronary artery disease defined by a more than 50% reduction in coronary lumen in 33.3% of the study population, and 23.2% were subsequently treated with PCI during pre-TAVR work-up. There were no significant differences between the VIV-TAVR and NV-TAVR groups accounting for gender, comorbidities, and cardiovascular risk factors. However, study participants in the VIV-TAVR group were younger (80.4 ± 12.2 vs. 84.1 ± 6.8), with more history of cardiovascular events, and expressed a higher predicted mortality risk. The success rate of the VIV-TAVR procedure was as high as NV-TAVR (98.6 vs. 98.8%). Most TAVR procedures were performed through the common femoral artery (92.8%), and Proglide was mainly used as a vascular access closure device (89%) (Table 2). Unlike the NV-TAVR group, where a balloon-expandable valve was commonly used (54.8%), the majority of implanted devices in the VIV-TAVR group were self-expandable (95.7%). The mean duration of VIV-TAVR interventions was longer than that of NV-TAVR procedures (87 ± 24 vs. 76 ± 24 min, p < 0.05), but VIV-TAVR was performed with a low mean contrast volume (114 ± 49 vs. 138 ± 49 ml, p < 0.05). The mean of residual aortic gradient tends toward a significantly higher level in association with VIV-TAVR (12.3 ± 6.7 vs. 10.6 ± 5.2, p = 0.053). Otherwise, the prevalence of per-TAVR complications, post-TAVR echocardiographic paravalvular leak, and length of hospital and CCU stays did not differ between the two study groups. Focusing on in-hospital post-TAVR complications, the single significant difference was less permanent pacemaker implantation in patients with degenerated bioprosthetic valves (2.9 vs. 11.7%, p = 0.024) (Table 3). Note that stroke did not occur in the VIV-TAVR group (0 vs. 1.8%). The multivariable logistic regression confirmed the negative association of pacemaker implantation with VIV-TAVR [OR = 0.235 95%CI (0.056–0.990), p = 0.048] and revealed the positive association of residual aortic gradient with VIV-TAVR [OR = 1.139, 95%CI (1.097–1.182), p = 0.001] (Table 4). Thus, no significant differences in vascular complication, bleeding, acute heart failure, stroke, and acute kidney injury were observed. In contrast to the predictions, the in-hospital mortality rate was lower in the VIV-TAVR group than in the NV-TAVR group (1.4 vs. 2.6%) but did not reach a statistically significant level (Table 3). Over a mean follow-up period of 3.44 ± 1.67 years, the observed rate of all-cause death was 22%, and it was non-significantly higher in the VIV-TAVR group than in the NV-TAVR group (30.4 vs. 21.6%, p = 0.085) (Figure 2). Lastly, the Kaplan-Meier curve and the Log-rank test showed no difference in survival outcome analysis over time (p = 0.074) (Figure 3).

The main finding of the present study is that VIV-TAVR interventions result in a similar success rate to NV-TAVR, as well as fewer permanent pacemaker implantations and a higher residual aortic gradient. Furthermore, VIV-TAVR patients are less likely to be exposed to different post-TAVR complications, at least during the hospital stay, accounting for vascular complications, bleeding events, and stroke. Far from the predicted mortality, VIV-TAVR showed a better early mortality rate and no significant difference in long-term, real-world survival.

In our study, the self-expandable device was nearly implanted in almost all VIV-TAVR patients. A recently published study evaluating the outcomes of self-expandable valves versus balloon-expandable valves in both VIV-TAVR and NV-TAVR patients has reported better results from the self-expandable device implant for degenerated aortic bioprostheses (13). The overall findings in our study are in line with previously reported data from valve-in-valve registries (14, 15) and published cohort studies (13, 16). For example, the prevalence of new permanent pacemaker implantation in the present study was closer to that reported in a large cohort including 1,150 patients with failed surgical aortic valve replacement in the VIV-TAVR (2.9 vs. 3%) and NV-TAVR (11.7 vs. 10.9%) groups, respectively (16). In comparison with NV-TAVR, the TAVR device is placed within the bioprosthetic valve during VIV-TAVR intervention. Thereby, the limited contact with the myocardium results in less stress on the cardiac conduction system and is likely to explain why fewer patients are requiring permanent pacemaker implantation following the VIV-TAVR procedure (17). It is of note that pacemaker implantation after TAVR is associated with higher risks of all-cause mortality and rehospitalization for heart failure but is not associated with stroke or endocarditis (18). However, a high residual gradient following the VIV-TAVR procedure remains a real concern. Pre-existing prosthesis-patient mismatch, small effective orifice area, and deep valve implant were identified as strong predictors, while supra-annular valve type and high transcatheter heart valve implant were associated with risk reduction (19). In addition, a decrease in mean residual gradient and an improvement in mean aortic valve area were associated with bioprosthetic valve fracture in the context of VIV-TAVR for patients with degenerated bioprosthetic valves (20). The present study shows a significantly higher mean residual aortic gradient after VIV-TAVR procedures. However, the ratio of residual aortic gradient within the study groups was lower than that previously reported (16), largely due to a larger number of self-expandable valve implants being used in patients with failed surgical bioprostheses. A higher residual gradient after VIV-TAVR is highly likely due to a small effective orifice area after the procedure, and it is highly dependent on the valve implanted previously during the index cardiac surgery. We emphasize that the success of VIV-TAVR depends on the size of the valve implanted during the index cardiac surgery and, sometimes, surgeons end up implanting very small valves that increase the risk of mismatch. This could be solved by using techniques to enlarge the aortic annulus/root to implant larger valves during index cardiac surgery (21, 22). Note, the benefit of aortic annulus enlargement on reducing moderate to severe patient-prosthesis mismatch after surgical aortic valve replacement should be assessed against a potential increase in perioperative mortality (21). Avoiding mismatch during index cardiac surgery is of paramount importance to reducing mortality after surgical aortic valve replacement and to the success of future VIV-TAVR (21–24). Furthermore, the size and brand of the surgically implanted bioprosthesis, in conjunction with the final result of the index cardiac surgery, are relatively important in determining the outcomes of VIV-TAVR. Lastly, VIV-TAVR procedures showed a similar safety and efficacy profile to NV-TAVR. In terms of survival, it was associated with better early outcomes but a higher long-term all-cause mortality rate. In parallel, an immediate protective effect was also revealed while comparing VIV-TAVR to redo-surgery as a therapeutic option for the management of degenerated aortic bioprostheses (9, 25). However, a recent meta analysis noticed the reverse of this observed early advantage of VIV-TAVR on all-cause death in favor of redo surgery after 6 months of follow-up (26), whereas a propensity-matched cohort study reported a better 5-year survival outcome after VIV-TAVR versus redo surgery (25).

In the absence of randomized clinical trials comparing VIV-TAVR to redo surgical aortic valve replacement, this study provides insights into the safety and efficacy of VIV-TAVR procedures, compared to the standard NV-TAVR, and the treatment modality of choice in high-risk severe aortic stenosis patients. The present findings suggest that high-risk patients with failed bioprosthetic valves do reasonably well after a VIV-TAVR procedure in a real-world setting.

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

All authors listed have made a substantial, direct, and intellectual contribution to the work, and approved it for publication.