Treatment of late paravalvular regurgitation after transcatheter aortic valve implantation: prognostic implications

Abstract

Structured Graphical Abstract

Transcatheter treatment of paravalvular regurgitation (PVR) after transcatheter aortic valve implantation (TAVI) using redo-TAVI, plug closure, or balloon valvuloplasty. Out of 201 patients, the failed valve was self-expanding in two-thirds and balloon-expandable in one-third of the patients. There were 87 patients (43%) who underwent redo-TAVI, 79 (39%) plug closure, and 35 (18%) balloon valvuloplasty. At 30 days, aortic regurgitation (AR) ≥ moderate persisted in 33 (17.4%): 8 (9.9%) after redo-TAVI, 18 (25.9%) after plug closure, and 7 (21.9%) after valvuloplasty (P = 0.036). Regardless of treatment strategy, patients with persistent AR ≥ moderate had higher 1-year mortality compared with those in whom AR was reduced to ≤ mild.

Open in new tabDownload slide

See the editorial comment for this article ‘Prevention and management of paravalvular regurgitation after transcatheter aortic valve implantation’, by C.J. Chung and C.M. Otto, https://doi.org/10.1093/eurheartj/ehad151.

Introduction

Aortic paravalvular regurgitation (PVR) after transcatheter aortic valve implantation (TAVI) is associated with increased morbidity and mortality.¹ Although the rate of PVR after TAVI has decreased,² the long-term adverse PVR effect may be clinically more relevant in patients with longer life expectancy.³ Various transcatheter intervention techniques are available to treat late PVR after TAVI, yet supportive data are limited.⁴ The aim of this study was to assess the efficacy and safety of different transcatheter modalities in reducing PVR after the index TAVI using a multicenter registry.⁵

Methods

Registry design

Twenty-two centers participating in the redo-TAVI registry⁵ contributed data on consecutive patients who underwent transcatheter intervention for PVR after TAVI using either redo-TAVI, plug closure, or balloon valvuloplasty. Only patients with an isolated re-intervention (temporal separation between index TAVI and re-intervention) were included. Patients were excluded from the analysis if aortic regurgitation (AR) location at presentation after TAVI (PVR vs. central regurgitation) remained undetermined. Baseline demographics, clinical and echocardiographic features, procedural, and follow-up data were collected by the co-investigators at each institution. Data collection and monitoring regarding the outcomes were assessed according to the Valve Academic Research Consortium-3 (VARC-3) definitions.⁶ Patients were categorized according to the type of applied intervention: redo-TAVI, plug closure, or balloon valvuloplasty. For each patient, the index TAVI and PVR intervention date as well as device model and size were collected. Baseline bioprosthesis valve area, mean and maximal gradients, and degree and location of regurgitation were gathered from echocardiographic studies prior to the PVR intervention, at 30 days and 1 year later. Echocardiographic data were site-reported according to established guidelines.⁶ Inconsistencies were resolved directly by communicating with the local investigators. The inclusion of patients was approved in each center by a local ethics committee.

Endpoints and definitions

The primary endpoints were defined as 30-day AR as well as 30-day and 1-year overall mortality post-PVR intervention. Secondary endpoints included re-intervention success [a 30-day composite procedural success defined by VARC-3,⁶ including freedom from mortality, freedom from intervention related to the device or to a major vascular/cardiac structural complication, and technical success with intended performance of the valve (mean gradient <20 mmHg and less than moderate AR)], stroke, cardiac structural complications, residual aortic valve mean gradient, and patient functional class at 30 days.

Statistical analysis

Descriptive data are reported as mean ± standard deviation or median [interquartile range (IQR)] for continuous variables, number (%) for categorical variables, and ratios with their respective 95% confidence intervals (CIs) as appropriate. Comparison of proportions was performed using Fisher’s exact method, while comparison of continuous variables was performed using a Kruskal–Wallis or a Wilcoxon test as appropriate. Multivariable and univariable predictors of mortality were evaluated using a Cox proportional hazards model ensuring compliance with the proportional hazard assumption by plotting Schoenfeld residuals. Multivariable Cox models were refined using a stepwise procedure based on the Akaike information criterion (AIC) for variable selection. To reduce bias, the multivariable analysis was performed with missing values imputed using the ‘missForest’ R package⁷ utilizing a non-parametric random forest method to impute values; all fields had non-missing values in at least 80% of cases. Kaplan–Meier curves were used to compare survival between specific groups, and the log-rank test was used to assess statistical significance of inter-group differences. As post-procedural residual PVR was evaluated after 30 days, a landmark analysis excluding patients who died or were censored before 30 days of follow-up was used to evaluate survival according to PVR status. Findings were considered statistically significant when P < 0.05. All calculations were performed using R version 4.0.4 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Baseline characteristics

Patient baseline characteristics are presented in Table 1. A total of 201 consecutive patients treated percutaneously for ≥ moderate PVR after index TAVI were identified: 87 (43%) by redo-TAVI, 79 (39%) with plug device, and 35 (18%) by balloon valvuloplasty. Age and sex were comparable between the three cohorts: mean age 78 ± 8 years, approximately two-thirds were female. The Society of Thoracic Surgeons (STS) risk score was 5.1 ± 4.4% overall and higher among the redo-TAVI cohort compared with the plug or valvuloplasty cohorts (6.1 ± 5.6% vs. 3.8 ± 2.3% vs. 4.7 ± 2.9%, respectively; P = 0.009). The failed valve more frequently had a self-expanding (vs. balloon-expandable) mechanism (129, 63.9%). This difference was more pronounced in the redo-TAVI cohort (69, 79.3%) compared with the plug (41, 51.9%) or valvuloplasty (19, 54.2%) cohorts (P = 0.001). A detailed TAVI model scattering is available in Supplementary data online, Table S1. Half of the re-interventions for PVR procedures (n = 100) were performed during 2019–22, one-third (n = 68) during 2014–18, and 15% (n = 33) during 2009–13 (Supplementary data online, Figure S1). Median time elapsing between the index TAVI and PVR re-intervention was 207 (IQR 35; 765) days (between 1 and 3759 days): 428 (72; 1030) in redo-TAVI, 182 (18; 513) in plug closure, and 61 (19; 408) in balloon valvuloplasty (P < 0.001). A similar time distribution remained when patients were stratified according to the index TAVI valve expansion mechanism: balloon- vs. self-expanding (Supplementary data online, Figure S2).

Procedural characteristics

Figure 1 illustrates the main re-intervention features. The majority of cases were done using transfemoral access, excluding three transradial plugs and four alternative access redo-TAVI cases. Redo-TAVI was most frequently done with balloon-expandable Sapien-type valve (Edwards Lifesciences, Irvine, CA; n = 55, 64%) followed by a self-expanding Evolut-type valve (Medtronic, Minneapolis, MN; n = 24, 28%) and other less frequent self-expanding valves. In the plug closure cohort, the most frequent device used was an Amplatzer vascular plug (AVP) II (n = 33, 42%), followed by AVP IV (n = 30, 38%), AVP III (n = 12, 16%), and Amplatzer duct occluder (ADO) (n = 3, 4%) (all Abbott, Abbott Park, IL). Mean plug size was 8.38 ± 2.21 mm (in AVP II: refer to long diameter) with an average of 1.32 ± 0.49 plugs exploited per patient. In the balloon valvuloplasty cohort, the most frequent balloon used was True balloon (BARD, Tempe, AZ; n = 20, 56%) followed by Z-med (B. Braun Medical, Bethlehem, PA; n = 6, 17%) and OSYPKA (OSYPKA AG, Rheinfelden, Germany; n = 5, 14%).

Figure 1

Procedural characteristics of transcatheter interventions for post-transcatheter aortic valve implantation paravalvular regurgitation using either redo-transcatheter aortic valve implantation, plug, or balloon valvuloplasty. True balloon (BARD, Tempe, AZ). Z-med (B. Braun Medical, Bethlehem, PA). OSYPKA (OSYPKA AG, Rheinfelden, Germany). AVP II/III/IV and ADO (Abbott, Abbott Park, IL). Sapien (Edwards Lifesciences, Irvine, CA). Evolut (Medtronic, Minneapolis, MN). BEV, balloon-expandable valve; SEV, self-expanding valve; TAVI, transcatheter aortic valve implantation; PVR, paravalvular regurgitation

Open in new tabDownload slide

Primary outcomes

At 30 days after the re-intervention for PVR, AR ≥ moderate persisted in 33 (17.4%) patients overall: 8 (9.9%) after redo-TAVI, 18 (25.9%) after plug closure, and 7 (21.9%) after balloon valvuloplasty (P = 0.036). In 27/29 (94%) patients, persistent AR was classified as PVR (vs. central regurgitation). Residual AR grade remained stable between 30 days and 1 year (Figure 2).

Figure 2

Aortic regurgitation at baseline, 30 days, and 1 year according to treatment approach. This bar graph presents the degree of aortic regurgitation as assessed by echocardiography. Aortic regurgitation ≥ moderate is stratified as paravalvular or central according to the main regurgitant jet location. PVR, paravalvular regurgitation, AR, aortic regurgitation; TAVI, transcatheter aortic valve implantation

Open in new tabDownload slide

Overall, 10 (5.0%) patients died within 30 days and 29 (14.4%) within 1 year after re-intervention. At 30 days, there were no mortality cases in redo-TAVI group, eight (10.1%) in the plug closure group, and two (5.7%) in the balloon valvuloplasty group (P = 0.010). At 1 year, mortality rates were comparable between the 3 groups: 11 (12.6%) vs. 14 (17.7%) vs. 4 (11.4%), respectively (P = 0.418). Regardless of treatment strategy, patients with AR ≥ moderate persisting after re-intervention had higher mortality at 1 year compared with those in whom AR was reduced to ≤ mild [6 (21.4%) vs. 11 (8.0%); P = 0.007; Figure 3].

Figure 3

Kaplan–Meier survival curves after transcatheter re-interventions for paravalvular regurgitation after index transcatheter aortic valve implantation. (A) Stratified by treatment approach, survival was higher in patients treated with redo-transcatheter aortic valve implantation at 30 days and comparable at 1-year follow-up. (B) Stratified by aortic regurgitation reduction success, survival at 1-year follow-up was higher in patients in whom aortic regurgitation was reduced to ≤ mild compared with those with persistent aortic regurgitation ≥ moderate. PVR, paravalvular regurgitation; AR, aortic regurgitation; TAVI, transcatheter aortic valve implantation

Open in new tabDownload slide

Secondary outcomes

The composite procedural success (Supplementary data online, Figure S3) was achieved in 107 (63.3%) overall: 50 (68.5%) in redo-TAVI, 33 (52.4%) in plug closure, and 24 (72.7%) in balloon valvuloplasty. The rates of other secondary endpoints were relatively low and comparable as detailed in Table 2. Postoperative in-hospital days were overall 3 (1; 6): 5 (2; 7.75) after redo-TAVI, 2 (1; 4) after plug closure, and 1 (1; 4) after balloon valvuloplasty. Compared with baseline (before PVR re-intervention), all 3 re-intervention strategies were associated with symptomatic improvement at 30 days (P < 0.001). This benefit persisted at 1 year and was similar between the cohorts (Supplementary data online, Figure S4 and Supplementary data online, Table S2). Longer-term survival (up to 5 years) also appeared comparable between the cohorts (Supplementary data online, Figure S5).

Discussion

While clinically significant PVR used to occur in approximately 15% of patients treated with early generation TAVI prostheses, the introduction of computed tomography for procedural planning and newer generation devices with external sealing skirts have led to lower PVR incidence.^1,2,8 Nevertheless, PVR remains a key issue in TAVI for several reasons. Firstly, even in the most carefully selected patients and modern devices, PVR occurs more frequently after TAVI than after surgery.⁹ Secondly, numerous studies have found PVR to be associated with significant increase in both short- and long-term morbidity and mortality,^1,8-10 and even mild PVL is arguably morbid.^11,12 Thirdly, with TAVI expansion to younger patients, the clinical impact of longstanding PVR may potentially have further impact on both clinical outcomes and bioprosthesis longevity. Notably, a larger portion of low-risk TAVI patients can potentially be suitable candidates for re-intervention to reduce PVR, if needed.

The current study of 201 patients who underwent transcatheter re-intervention for PVR after index TAVI is the largest and most comprehensive so far. Our main findings are the following (Structured Graphical Abstract): (i) despite re-intervention, persistent AR ≥ moderate was not infrequent (17.4% overall), occurring in 1 in 10 patients treated with redo-TAVI, 1 in 4 patients treated with plug-closure, and 1 in 5 patients treated with balloon valvuloplasty; (ii) overall mortality was 5% at 30 days and 14.4% at 1 year: 0% and 12.6% after redo-TAVI, 10.1% and 17.7% after plug closure, and 5.7% and 11.4% after balloon valvuloplasty, respectively; and (iii) patients with residual AR ≥ moderate despite re-intervention had higher mortality at 1 year compared with patients in whom AR was reduced to ≤ mild [6 (21.4%) vs. 11 (8.0%); P = 0.007].

Literature on transcatheter PVR management is predominantly limited to case series and registries of surgical valve patients.¹³ Given the rigid suture ring of surgical bioprostheses, neither valve-in-valve nor balloon valvuloplasty is of potential benefit, and plug closure plays a key role (even more so in mechanical valves). In the largest series to date, a quarter of the 259 patients treated with plug for post-surgical PVR (of whom only 48% aortic) had persistent AR ≥ moderate, similar to our current findings in post-TAVI patients. Also similar was the association between AR reduction and survival.¹⁴

So far, only anecdotal cases of PVR management after index TAVI have been reported.^4,15,16 The largest series (n = 27) mainly described the use of AVP (80%), where although 24 (89%) of the procedures were determined technically successful, survival was only 61% at 1 year, probably reflecting risk profile and comorbidities of an earlier TAVI patient population.⁴ Redo-TAVI or balloon valvuloplasty were only recently described as treatment options for PVR after TAVI. Our group recently reported a 10% rate of persistent AR ≥ moderate with redo-TAVI applied for ≥ moderate PVR (vs. 2.5% with redo-TAVI for ≥ moderate central regurgitation, P = 0.137).¹⁷ Akodad et al.¹⁸ reported on 11 patients with PVR ≥ moderate after index TAVI in whom balloon valvuloplasty reduced PVR to ≤ mild in all, concluding that in selected patients, late balloon valvuloplasty could be a valuable alternative to plug.

The optimal PVR treatment is prevention. When significant PVR occurs during the index TAVI, several strategies (i.e. balloon post-dilatation, supplementary valve implant) can potentially be applied effectively although not without side effects.^19-21 Once diagnosed later and deemed clinically significant, options include conservative management, surgery, and transcatheter intervention. As conservative management or surgical TAVI valve explantation and surgical aortic valve replacement are associated with poor outcome,^22,23 transcatheter treatments seem attractive.

Balloon valvuloplasty may mitigate PVR by expanding the primary valve in case of under-expansion or under-sizing.^18,24 Redo-TAVI can conceivably act the same yet has the added advantage of preventing possible recoil of the primary valve and can also aid by extending the failed sealing skirt in case of primary valve malpositioning.^17,25 On the other hand, redo-TAVI can increase the risk for patient–prosthesis mismatch and coronary obstruction or access difficulties.²⁶ A plug should be especially effective when a large, asymmetric burden of calcium is causing the sealing failure, yet can properly cover only focal leakages. Recognizing the main PVR mechanism is therefore critical for choosing the most suitable approach, which can become more complex when there are multiple underlying mechanisms. Management strategies undoubtedly depend on more than the primary mechanism (under-expansion, malpositioning, protruding calcification) as other factors are also vital to consider such as patient anatomy (sinotubular junction height/diameter and distance to the coronary arteries), echocardiographic features (jets number and size), and patient characteristics such as suspected infection, frailty, and expected longevity.

It is important to emphasize that although the rate of persistent AR and mortality were more favorable in patients treated with redo-TAVI as compared with those treated with plug closure or valvuloplasty, it is beyond the scope of this retrospective analysis to recognize whether these variances derive from procedural efficacy differences or bias. For instance, using an arbitrary cut-off time (between index TAVI and re-intervention for PVR) of 3 months to separate ‘early’ (n = 76) vs. ‘late’ (n = 125) presenters, the rate of residual AR ≥ moderate was 16 (25.4%) vs. 13 (12.5%) (P = 0.04) and the rate of mortality 8 (10.5%) vs. 2 (1.6%) (P = 0.007) at 30 days after re-intervention, respectively. This is probably because heterogeneity exists between early and late presenters, as does between other unmeasured predispositions. Accordingly, choosing the most suitable treatment requires a detailed pre-procedural planning by an experienced structural heart team and a tailored approach.

Limitations

This is an observational study that has collected a carefully selected cohort of patients. Timing and treatment modality was left to the discretion of the local heart team, probably reflecting suitable anatomy and suitable patient characteristics for a specific transcatheter re-intervention. An independent adjudication of events or an independent core laboratory imaging analysis was lacking, and imaging data (i.e. jet measures and number, valve expansion and position, aortic root measures, root and left ventricular outflow tract calcification) essential to PVR mechanism recognition and therapy assortment were not available. The number of cases and events (persisting AR and mortality) was relatively small, which on one hand limits the power to detect smaller differences in outcomes which may still be clinically significant and on the other hand can lead to spurious findings. Outcomes, especially mortality risk, should not be interpreted as the result of the procedural approach only.

Conclusions

This study describes the efficacy and safety of redo-TAVI, plug closure, and balloon valvuloplasty in treating PVR after TAVI. Patients in whom post-TAVI PVR was reduced had better prognosis regardless of treatment strategy. Patient selection for these therapies needs further investigation.

Acknowledgements

We would like to thank all redo-TAVI registry investigators for their continuous efforts and assistance.

Supplementary data

Supplementary data is available at European Heart Journal online.

Pre-registered clinical trial number

None supplied.

Ethical approval

Ethical Approval was not required.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Funding

All authors declare no funding for this contribution.

References

Author notes