Author + information
- Received December 18, 2017
- Revision received April 20, 2018
- Accepted April 23, 2018
- Published online July 16, 2018.
- E. Murat Tuzcu, MDa,∗ (, )@ClevelandClinic,
- Samir R. Kapadia, MDb,
- Sreekanth Vemulapalli, MDc,
- John D. Carroll, MDd,
- David R. Holmes Jr., MDe,
- Michael J. Mack, MDf,
- Vinod H. Thourani, MD, MPHg,
- Frederick L. Grover, MDh,i,
- J. Matthew Brennan, MD, MPHc,
- Rakesh M. Suri, MD, DPhila,
- David Dai, PhDc and
- Lars G. Svensson, MD, PhDb
- aCleveland Clinic, Abu Dhabi, United Arab Emirates
- bCleveland Clinic, Cleveland, Ohio
- cDuke University Medical Center, Duke Clinical Research Institute, Durham, North Carolina
- dUniversity of Colorado, Anschutz Medical Campus, School of Medicine, Aurora, Colarado
- eMayo Clinic, Rochester, Minnesota
- fBaylor Scott and White Health, Plano, Texas
- gMedstar Washington Hospital Center, Washington, DC
- hDepartment of Surgery, School of Medicine, University of Colorado, Anschutz Medical Campus, Aurora, Colorado
- iDenver Department of Veterans Affairs Medical Center, Denver, Colorado
- ↵∗Address for correspondence:
Dr. E. Murat Tuzcu, Department of Cardiovascular Medicine, Cleveland Clinic Abu Dhabi, PO Box 112412, Al Maryah Island, Abu Dhabi, United Arab Emirates.
Background Valve-in-valve (ViV) transcatheter aortic valve replacement (TAVR) has been shown to be feasible, yet the safety and efficacy in relation to native valve (NV) TAVR are not known.
Objectives This study sought to evaluate the safety and effectiveness of ViV TAVR for failed surgical aortic valve replacement (SAVR) by comparing it with the benchmark of NV TAVR.
Methods Patients who underwent ViV-TAVR (n = 1,150) were matched 1:2 (on sex, high or extreme risk, hostile chest or porcelain aorta, 5-m-walk time, and Society of Thoracic Surgeons Predicted Risk of Mortality for reoperation) to patients undergoing NV-TAVR (n = 2,259). Baseline characteristics, procedural data, and in-hospital outcomes were obtained from the Transcatheter Valve Therapy Registry. The 30-day and 1-year outcomes were obtained from linked Medicare administrative claims data.
Results Unadjusted analysis revealed lower 30-day mortality (2.9% vs. 4.8%; p < 0.001), stroke (1.7% vs. 3.0%; p = 0.003), and heart failure hospitalizations (2.4% vs. 4.6%; p < 0.001) in the ViV-TAVR compared with NV-TAVR group. Adjusted analysis revealed lower 30-day mortality (hazard ratio: 0.503; 95% confidence interval: 0.302 to 0.839; p = 0.008), lower 1-year mortality (hazard ratio: 0.653; 95% confidence interval: 0.505 to 0.844; p = 0.001), and hospitalization for heart failure (hazard ratio: 0.685; 95% confidence interval: 0.500 to 0.939; p = 0.019) in the ViV-TAVR group. Patients in the ViV-TAVR group had higher post-TAVR mean gradient (16 vs. 9 mm Hg; p < 0.001), but less moderate or severe aortic regurgitation (3.5% vs. 6.6%; p < 0.001). Post-TAVR gradients were highest in small SAVRs and stenotic SAVRs.
Conclusions Comparison with the benchmark NV-TAVR shows ViV-TAVR to be a safe and effective procedure in patients with failed SAVR who are at high risk for repeat surgery.
Transcatheter aortic valve replacement (TAVR), which was initially introduced for patients with severe aortic stenosis who are at prohibitive risk for surgical aortic valve replacement (SAVR), has evolved to be an important treatment option for high and intermediate surgical risk patients (1–7). Another important group of patients for which transcatheter heart valves (THV) are approved in the United States includes patients with failed SAVR. An early global registry has shown that treating such patients by TAVR is feasible (8). Follow-up data from this international registry, which included patients who are at prohibitively high risk for repeat open-heart surgery, provided insights regarding safety and efficacy of this procedure called valve-in-valve (ViV) TAVR (9). Yet there is no trial comparing ViV-TAVR with redo SAVR, which would be the definitive test for comparative effectiveness. In the absence of such data, we sought to evaluate the safety and efficacy of ViV-TAVR in comparison with standard TAVR performed for native aortic valve stenosis. Cognizant of the fact that a comparative effectiveness study could not be done in inherently different patient populations, our intention was to test the hypothesis of whether ViV-TAVR is comparable in safety and effectiveness in high-risk patients with failed SAVR to native valve (NV)-TAVR, which is a standard treatment modality in high-risk aortic stenosis patients.
Transcatheter valve therapies registry
The Transcatheter Valve Therapy (TVT) Registry was established by the Society of Thoracic Surgeons (STS) and the American College of Cardiology in collaboration with the Centers for Medicare and Medicaid Services (CMS) the Food and Drug Administration and multiple industry groups. Submission of the consecutive TAVR cases to the registry was mandated by the CMS National Coverage Determination of May 2012 (10).
Patients undergoing ViV from November 9, 2011, through June 30, 2016, were matched on sex, inoperable/extreme risk designation, hostile chest or porcelain aorta, 5-m-walk time, and STS Predicted Risk of Mortality (PROM) for reoperation in a 1:2 fashion to patients undergoing NV-TAVR. Patients being considered for ViV-TAVR for failing surgical bioprosthesis have a higher STS-PROM than patients undergoing NV-TAVR with similar comorbidities. This is because the STS-PROM, which is a surgical risk score, treats surgeries requiring repeat sternotomy as higher risk procedures. However, in the present analysis, the STS-PROM score is used as a marker of comorbidity. Therefore, to make sure that the STS-PROM scores of ViV and NV patients represent the same comorbidity burden, the STS-PROM scores of NV patients were calculated as though they were undergoing reoperation. For example, an 80-year-old man with diabetes, hypertension, and coronary artery disease undergoing a first operation would have an STS-PROM of 5%, whereas the same patient undergoing a reoperation would have an STS-PROM of 7%. Calculating the STS-PROM for reoperation in both the ViV-TAVR and NV-TAVR groups allows the STS scores to be directly comparable between the 2 groups and more accurately reflect the risk attributable to comorbidities rather than a second sternotomy.
Pre-hospital, intra-hospital, and in-hospital post-procedural data elements were prospectively collected according to the Valve Academic Research Consortium 1 and 2 and previously published standardized definitions via TVT Registry case report forms, which include patient demographics, comorbidities, functional status, pre- and post-TAVR echocardiographic and catheterization results, procedural and in-hospital outcome data (10–12). Death, stroke, and unplanned surgery or intervention events were adjudicated by board-certified cardiologists at the analysis center. Thirty-day and 1-year mortality, stroke, hospitalization for heart failure, and aortic valve intervention data were obtained from linkage of the TVT Registry to Medicare administrative claims files. Specifically, Medicare inpatient administrative claims files were used for detection of rehospitalization events during the study interval using the following International Classification of Diseases-9th Revision-Clinical Modification codes: for stroke, 433.x1, 434.x1, 997.02, 436, 437.1, 437.9, 430, 431, and 432.x; for heart failure, 398.x, 402.x1, 404.x1, 404.x3, and 428.x; and for aortic valve re-intervention, 35.11, 35.21, 35.22, 35.01, 35.05, 35.06, and 35.09. For rehospitalization, follow-up was censored at the time of death, end of fee-for-service coverage, loss of Part A or B coverage, or end of the follow-up period, whichever occurred first.
Major safety endpoints include in-hospital, 30-day death, and stroke. In-hospital events of major bleeding, major vascular complication, new pacemaker, and other procedural complications were also analyzed as safety endpoints. Efficacy endpoints included site-reported echocardiographic data obtained before hospital discharge and 1-year mortality, freedom from stroke, hospitalization for heart failure, and aortic valve reintervention.
The sizes of surgical prostheses were not available for patients whose data were entered into the first-generation case report form, which did not include data about surgical valves. The initial TVT Registry was designed only for TAVR performed for native aortic valve stenosis. The data about the sizes of surgical prostheses were collected retrospectively until July 2014 when the case report form was updated. Of the 1,150 ViV-TAVR patients, 868 (75.5%) had the manufacturer-reported surgical valve sizes documented. Determination of the failure mode of the surgical bioprostheses (stenosis vs. regurgitation) was performed according to the criteria of the American Society of Echocardiography from the baseline echocardiographic information (13). If concomitant moderate or severe stenosis and regurgitation were present, the failure mode was considered to be combined.
Data were analyzed after the quality check at the National Cardiovascular Data Registry data warehouse and Duke Clinical Research Institute to optimize data completeness and accuracy. The TVT-Registry Steering Committee monitors data quality and completes annual audits. TVT Registry activities have been approved by the Chesapeake Central Institutional Review Board, and the Duke University Institutional Review Board granted a waiver of individual patient-informed consent and authorization for this study.
Baseline patient characteristics were summarized as percentage or median (interquartile range [IQR]), as appropriate, and compared across subgroups using chi-square, Wilcoxon, or Kruskal-Wallis 2-sided tests. For analyses of mortality, the Kaplan-Meier method was used for unadjusted mortality rates to 1 year.
For analyses of stroke, heart failure, and aortic valve reintervention, analysis focused on estimating the cumulative incidence function and the cumulative probability of an endpoint occurring over time in a patient’s lifetime. The cumulative incidence function is the accepted parameter for describing nonfatal events from a patient perspective in a setting of high competing mortality risk. Unlike standard time-to-event methods, which describe the probability of a nonfatal endpoint occurring in a hypothetical death-free environment, the cumulative incidence function models the probability that an endpoint will actually occur, given that death may preclude an event from occurring.
Unadjusted and adjusted hazard ratios (HRs) comparing mortality risks across subgroups were estimated along with 95% approximate confidence intervals (CIs) and Wald-type p values. Within-group comparisons were accomplished using a marginal Cox model where the intracluster (hospital) correlation dependence was accounted for using a robust sandwich covariance matrix estimate. Covariates for adjusted outcomes are listed in Online Table 1. For all analyses, a 2-sided p < 0.05 was considered statistically significant. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, North Carolina).
Factors associated with successful ViV were determined by backward selection logistic regression with p < 0.05 considered significant. For factors associated with 1-year mortality among patients undergoing ViV-TAVR, Cox proportional hazards models were used. Covariates are listed in Online Table 2.
Between November 2011 and June 2016, a total of 71,481 commercial TAVRs (n = 2,936, ViV-TAVR and n = 68,545, NV-TAVR) from 457 centers were captured by the TVT registry and included in this study. Of these patients 44,326 (n = 1,549, ViV-TAVR and n = 68,545, NV-TAVR) were linked with 1-year Medicare administrative claims. Patients who underwent ViV-TAVR were matched 1:2 based on the aforementioned criteria. After the match 1,150 ViV-TAVR patients and 2,259 NV-TAVR patients constituted the final study group. Although there were some differences, baseline characteristics of the 1,150 patients who constituted the study group and 1,786 who were not included in the study were similar to a large extent (Online Table 3). In-hospital outcomes were similar in the included and excluded patient groups (Online Table 4).
Matching resulted in 2 groups with similar characteristics, although there were some differences (Table 1). The ViV-TAVR group more frequently had New York Heart Association functional class III or IV symptoms, moderate or severe mitral regurgitation, moderate or severe tricuspid regurgitation, permanent pacemaker, and lower left ventricular ejection fraction. Two or more previous cardiac surgeries, prior bypass surgery, and nonaortic valve surgery including mitral valve replacement and repair were also more common in the ViV-TAVR group. Patients in the NV-TAVR group were older (84 [IQR: 78 to 88] years vs. 79 [IQR: 74 to 85] years; p < 0.001); had higher rates of diabetes, coronary artery disease, prior percutaneous coronary intervention, and peripheral vascular disease; and more frequently required a nontransfemoral approach.
The failure mode of the SAVR was stenosis in 702 (61.0%) and regurgitation in 140 patients (12.2%). There was combined stenosis and regurgitation in 283 (24.6%) patients. Of the 868 patients whose existing surgical valve sizes were available, 61 (7.0%) had 19 mm, 240 (27.7%) had 21 mm, 285 (32.8%) had 23 mm, 192 (22.1%) had 25 mm, and 88 (10.1%) had ≥27 mm valves.
Balloon-expandable THV was used in 501 patients (n = 41, 20 mm; n = 310, 23 mm; n = 111, 26 mm; n = 39, 29 mm). Self-expanding THV was used in 647 patients (n = 385, 23 mm; n = 196, 26 mm; n = 57, 29 mm; n = 9, 31 mm). The procedure was done under general anesthesia in the ViV-TAVR group less frequently than in the NV-TAVR group (78.7% vs. 83.7%; p < 0.001). Fluoroscopy time was longer (21 [IQR: 15 to 30] min vs. 18 [IQR: 13 to 25] min; p < 0.001) and contrast volume was less (60 [IQR: 30 to 100] ml vs. 105 [IQR: 75 to 150] ml; p < 00.001) in the ViV-TAVR compared with NV-TAVR group.
In-hospital death was similar in the ViV-TAVR and NV-TAVR groups (2.1% vs. 2.7%; p = 0.25). Stroke was less frequent in the ViV-TAVR group (1.2% vs. 2.4%; p = 0.02). Patients were discharged home more frequently after ViV-TAVR. Need for a new permanent pacemaker, vascular complications requiring intervention, and major bleeding were less frequent in the ViV-TAVR group. Although device embolization rates were similar in the 2 groups, need for device capture and retrieval was more common in the ViV-TAVR group. Other procedural adverse outcomes occurred at similar frequencies in the 2 groups (Table 2).
Mortality and stroke rates, as well as frequency of other in-hospital outcomes, were similar in patients presenting with differing surgical prosthesis failure modes. In-hospital mortality was higher (6.6% in 19 mm, 2.1% in 21 mm, 1.1% in 23 mm, and 1.8 in ≥25 mm; p = 0.045) in patients with small surgical valves. There were 61 patients with 19-mm SAVR. These patients were more frequently female, had higher STS scores, and had slower 5-m-walk times. There was no statistically significant difference in mortality based on the THV size used.
30-day and 1-year outcomes
Unadjusted analysis revealed lower 30-day mortality (2.9% vs. 4.8%; p < 0.001), stroke (1.7% vs. 3.0%; p = 0.003), and heart failure hospitalizations (2.4% vs. 4.6%; p < 0.001) in the ViV-TAVR group compared with NV-TAVR. Aortic valve reintervention frequency was similarly low in the 2 groups, although numerically lower in the ViV-TAVR group (Central Illustration).
After adjustment, 30-day mortality (HR: 0.503; 95% CI: 0.302 to 0.839; p = 0.008) and 1-year mortality (HR: 0.655; 95% CI: 0.505 to 0.844; p = 0.001) remained lower in the ViV-TAVR group. Adjusted analysis showed a trend toward a lower 30-day stroke rate (HR: 0.559; 95% CI: 0.302 to 1.036; p = 0.064) in ViV, but at 1 year, there was no difference. Adjusted analysis of the hospitalization for heart failure at 30 days (HR: 0.601; 95% CI: 0.353 to 1.023; p = 0.061) showed a trend toward being lower in the ViV-TAVR group, and it was significantly lower at 1 year (HR: 0.685; 95% CI: 0.500 to 0.939; p = 0.019) (Figure 1). Adjusted analysis of aortic valve reintervention showed no difference between ViV-TAVR and NV-TAVR at 30 days or 1 year (Table 3).
To evaluate the possible impact of the differences in age between the 2 groups we compared the ViV-TAVR and NV-TAVR patients according to age <80 years and ≥80 years. One-year mortality was lower in the ViV–TAVR compared to the NV-TAVR group in younger as well as older patients (Figure 2).
Predictors of mortality
Several patient characteristics were associated with 1-year mortality after ViV-TAVR. These included emergent or salvage TAVR, creatinine >2 mg/dl or dialysis, nonfemoral access, anemia, and low platelet count (Figure 3).
The mean aortic-valve gradient was decreased significantly after the ViV-TAVR (40 mm Hg to 16 mm Hg; p < 0.001) and the NV-TAVR (42 mm Hg to 9 mm Hg; p < 0.001). However post-TAVR mean gradients were higher in the ViV-TAVR group compared with the NV-TAVR group (16 mm Hg vs. 9 mm Hg; p < 0.001). The frequency and severity of aortic regurgitation were lower after ViV-TAVR than after NV-TAVR (Table 4, Figure 4A).
Mean gradients after ViV-TAVR were different in patients with different modes of SAVR failure (17 mm Hg, 12 mm Hg, and 15 mm Hg in the stenosis, regurgitant, and combined groups, respectively) (Figure 4B). Post-procedure mean gradients were higher in patients with smaller surgical prosthesis (Figure 4C) and those in whom smaller THVs were used (Figure 4D). Additionally, gradients differed by THV type, particularly in smaller SAVR sizes (Figure 5).
Among patients who were treated with balloon-expandable valves, 256 received Sapien-XT and 147 received S3. There was no difference in baseline or post-ViV-TAVR mean gradients (39 [IQR: 28 to 49] mm Hg and 38 [IQR: 26 to 48] mm Hg; p = 0.69; 16 mm Hg [IQR: 11 to 23] mm Hg and 17 [IQR: 11 to 26] mm Hg; p = 0.28, respectively). Similarly, there was no difference in mean gradients between 201 patients who were treated with earlier generation CoreValve and 434 with Evolute-R (41 [IQR: 30 to 51] mm Hg and 40 [IQR: 30 to 48] mm Hg; p = 0.30; 15 [IQR: 9 to 22] mm Hg and 16 [IQR: 11 to 262] mm Hg; p = 0.17, respectively).
Our findings show that surgical aortic bioprosthesis failure can be safely treated with ViV-TAVR in higher risk patients with acceptable 1-year outcome based on the analysis of comprehensive data from the U.S. real-world experience using commercially available valves. The results of ViV-TAVR were compared with NV-TAVR to provide a clinical context and benchmark for a variety of outcomes. Procedural, 30-day, and 1-year outcomes are similar or better when compared with NV-TAVR with fewer procedural complications. Post-TAVR mean gradients are higher after ViV-TAVR, particularly in patients with small surgical prosthesis, but post-TAVR moderate or severe aortic regurgitation are less frequent in this group.
The first large systematic report of ViV-TAVR was published in 2012 using the data from the Global-ViV registry (8). The 30-day mortality of 2.9% in our study is lower than the 30-day mortality of 7.6% in that report. The large variation between the mortality rates may be explained not only by differing patient characteristics but also by different technologies during the eras the studies were conducted. The Global Valve in Valve Registry started to enroll patients in December 2010, and initial results were published 1 month before the case entries were started to the TVT Registry in October 2012 (14).
Initial global registry data demonstrated high rate of valve malpositioning and coronary ostial obstruction, which were frequently lethal complications. This report served as an important warning, an opportunity to learn from the early experience of ViV-TAVR, and an impetus to improving patient selection and refining the techniques and technology used in ViV-TAVR. Substantially lower mortality in our study was accompanied by lower coronary complications including an obstruction rate of only 0.6%, which was 7-fold higher in the Global Registry. Device malposition was <1% in the current study compared with 15% in the initial global experience.
Findings in our study are in line with the recent ViV TAVR experience from device-company-sponsored investigational device exemption trials: PARTNER 2 (Placement of Aortic Transcatheter Valves) ViV Registry and the CoreValve U.S. Expanded Use Study reported in 2017 (15,16). In the PARTNER-2 ViV Registry only balloon-expandable THV valves (Edwards Lifesciences Irvine, California) were used. The 365 patients (age 78.9 ± 10.2 years; 64% male; STS score 9.1 ± 4.7%) had 30-day mortality of 2.7% and stroke rate of 2.7%. In the recent CoreValve registry, only self-expanding valves were used. The 277 patients (age 76.7 ± 10.2 years; 63% male; STS score 9.0 ± 6.7%) had 30-day mortality of 2.2% and stroke rate of 0.9%. In both clinical trials, rates of serious periprocedural complications were low. Furthermore, improved outcomes over time were demonstrated within the PARTNER ViV registry, which started enrollment in June 2012. The initial 96 patients had a 2-fold higher 1-year mortality compared with the 269 patients treated subsequently in the continued access phase of the study (19.8%, vs. 9.8%; p ≤ 0.006). Refinement of patient selection and procedural technique guided by the accumulating experience are the most likely reasons for the betterment of the outcomes.
Adjustment of outcomes for patient comorbidities using TVT Registry data and assessment of 1-year outcomes using CMS database linkage were important features of the presently reported study. After adjustment for comorbidities, 30-day and 1-year mortality and 1-year hospitalization for heart failure compared favorably with NV-TAVR. Mortality between 30 days and 1 year of 10.6% in the ViV-TAVR group is lower than the 16.1% in the NV-TAVR group but similar to the 30-day to 1-year mortality rates reported by the Global ViV Registry (10.4%), PARTNER ViV Registry (9.7%), and CoreValve study (12.2%). These findings suggest that symptomatic patients with degenerated surgically placed aortic prostheses who are at high risk for conventional surgery do reasonably well after a successful ViV-TAVR. The differences in the post-procedure outcomes between the ViV and NV-TAVR groups including higher frequency of atrial fibrillation, need for new permanent pacemaker, vascular complications requiring intervention, and major bleeding in the NV-TAVR group might have contributed to the diverging mortality curves between the 2 groups.
Our analysis revealed several patient characteristics including acuity of the procedure, chronic kidney disease, nonfemoral access, anemia and low platelets, which are associated with 1-year mortality after ViV-TAVR. These factors should be taken into account during pre-procedure decision making.
We were unable to determine cause of death in the present analysis using the CMS database linkage. As a result, it is not clear if the mortality after the first 30 days is related to cardiovascular causes or not. However, the fact that during this period <10% of patients in the ViV-TAVR group were admitted to the hospital for heart failure and a very small percentage had stroke or aortic valve intervention suggests most of the mortality may be due to noncardiac causes. These findings are in line with previous analysis showing that noncardiovascular causes dominate the reasons for rehospitalization after TAVR (17).
We found that procedure-related vascular complications and bleeding were higher in the NV-TAVR group. These differences may be caused by the higher frequency of peripheral vascular disease in these patients who are slightly older. The finding of fewer patients requiring pacemaker implantation with ViV-TAVR is in line with previous reports and is probably caused by placement of the TAVR valve within the surgical bioprosthetic valve with limited contact of the THV to the myocardium. The rate of coronary artery occlusion was similarly low in the ViV-TAVR and NV-TAVR groups in the current study and is in marked contrast to the early ViV-TAVR experience reported in the Global Registry (8).
Many procedural outcomes were similar or better in the ViV-TAVR group except the post-TAVR gradients. Mean gradient after ViV-TAVR was 2-fold higher than after NV-TAVR. The mean gradients in ViV-TAVR were related to 3 factors: 1) the size of the surgical prosthesis; 2) mode of its failure; and 3) the size and the type of THV used. However, it is important to note that baseline gradients varied considerably among the groups. Patients with small surgical prosthesis that were stenotic who were treated with small-size THV have the highest gradients at baseline and post-ViV-TAVR. These findings, which are similar to previous reports, should be considered in patient selection and procedural planning.
Caution is warranted particularly in small surgical prostheses. The PARTNER registry excluded patients with 19-mm SAVR. The CoreValve and Global registries did not report specifically on patients with 19-mm SAVR. In our study, in-hospital mortality, procedural complication rate, and residual gradient were higher in the 61 patients with 19-mm SAVR.
Currently, there is no study comparing TAVR and SAVR for failed surgical aortic bioprostheses primarily because there is also no proper surgical benchmark. Patients undergoing ViV-TAVR are often undesirable surgical candidates because of age, frailty, and a high burden of comorbid conditions. Data about reoperation for SAVR are therefore very limited for elderly and high-risk patients. Mortality and stroke rates in a single-center study including 216 patients with a mean age of 59 years who underwent reoperation for isolated aortic valve replacement between 1990 and 2002 were 4.6% and 4.6%, respectively (18). In another study, which enrolled 155 patients with a mean age of 58 between 1992 and 2006, mortality and stroke rates were 4.5% and 5.8%, respectively (19). In a 2006 report of 71 elderly patients, who were ≥80 years old, operative mortality was 15.5% (20).
Because of the aforementioned constraints, we decided to use NV-TAVR as a benchmark for evaluating ViV-TAVR, recognizing the inherent limitations of such a choice. A patient who needs NV-TAVR is by definition different from the one needing ViV-TAVR. These 2 procedures are not interchangeable. They are performed for 2 very different, mutually exclusive disease states. Although identical devices and similar techniques are used, landing zones and associated issues are quite different. Our purpose in comparing the 2 procedures was not to find out which one is superior in safety and efficacy. Rather, we wanted to see if ViV-TAVR is comparable in safety and effectiveness in high-risk patients with the benchmark procedure NV-TAVR. NV-TAVR was chosen as a comparator because its value has been proven in multiple randomized clinical trials, and its extensive use in clinical practice has been documented with real-world registries. Therefore NV-TAVR provides a benchmark and a clinical context to evaluate ViV TAVR.
Although ViV-TAVR seems to compare favorably with NV-TAVR, there are differences between the 2 groups that need to be taken into account. In addition to unmeasured variables, there are significant inequalities, which may have an impact on differences in the clinical outcomes including mortality. In the NV-TAVR group, diabetes, coronary artery disease, and peripheral vascular disease were more common than those who underwent ViV-TAVR. The nontransfemoral approach was more common in NV-TAVR.
Patients who underwent NV-TAVR were also older than those having ViV-TAVR. However, when we stratified the ViV-TAVR and NV-TAVR groups according to age, the results were similar to the main cohort. ViV-TAVR patients younger than 80 years of age had lower in-hospital and 1-year mortality than their NV-TAVR counterparts. The results were similar when comparisons were made among patients 80 years or older. These findings are in concordance with previous reports from NV-TAVR studies, which demonstrated that age is not a predictor of 30-day mortality (15,16).
However, the 2 groups differed in several aspects that raised the risk profile of the ViV-TAVR patients. Patients in this group had higher rates of moderate or severe mitral regurgitation, prior mitral valve surgery, moderate or severe tricuspid regurgitation, and history of 2 or more open heart surgeries.
This is a retrospective analysis of observational data reported by the individual centers. As a result, despite our use of matching and covariate adjustment, it may be subject to unmeasured residual confounding.
Although data quality checks were completed for all submitted data, adjudication was limited to in-hospital mortality, stroke, and unplanned surgery. Baseline and post-procedure echocardiography was reported by the centers without core laboratory assessment. Only Medicare fee-for-service patients were included in the analysis. Exclusion of some patients did not constitute a selection bias. Although there was some difference between those included and excluded from the study, many of those differences indicate that included patients were at higher risk than those who were excluded. Administrative claims data may not have sensitivity and specificity similar to the rigorously physician-adjudicated efficacy data in the pivotal trials. Conversely, TVT registry data, together with the administrative claims data, provide real-world effectiveness information.
In the present study, ViV-TAVR, which was compared with a benchmark, NV-TAVR, is shown to be a safe and effective procedure in high-risk patients with degenerated SAVR. We demonstrated low mortality, acceptable complication rates, and clear improvement in valvular hemodynamics with a low risk of moderate to severe aortic regurgitation. Transvalvular gradients are higher and related to both the mode of SAVR failure and the size of the failed SAVR. Self-expanding TAVR has lower transvalvular gradients that appears most clearly in small SAVR valves. ViV TAVR is increasingly performed in patients at elevated risk for reoperation. These results show that the clinical outcomes in real-world settings are within the bounds of clinical acceptability and justify this treatment being the favored approach in these patients.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: ViV-TAVR can be a safe and effective alternative to cardiac reoperation for high-risk patients with degenerated bioprosthetic aortic valves.
TRANSLATIONAL OUTLOOK: Longer-term follow-up studies are needed to clarify the durability of ViV-TAVR.
Dr. Vemulapalli has received research grants from the American College of Cardiology, Society of Thoracic Surgeons, Abbott Vascular, Patient Centered Outcomes Research Institute, and Boston Scientific; is a consultant for Premiere, Janssen, Zafgen, and Novella; and has served on the speakers bureau for Boston Scientific. Dr. Carroll is an investigator in research trials sponsored by Edwards LifeSciences, Abbott Vascular, and Direct Flow; and an investigator in a research trial sponsored by Medtronic. Dr. Mack is a co-principal investigator for the Partner 3 Trial, Edwards Lifesciences, uncompensated. Dr. Thourani is a consultant for Edwards Lifesciences. Dr. Brennan is a consultant for Edwards LifeSciences. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- Centers for Medicare and Medicaid Services
- hazard ratio
- native valve
- surgical aortic valve replacement
- Society of Thoracic Surgeons Predicted Risk of Mortality
- transcatheter aortic valve replacement
- transcatheter heart valves
- transcatheter valve therapy
- Received December 18, 2017.
- Revision received April 20, 2018.
- Accepted April 23, 2018.
- 2018 American College of Cardiology Foundation
- Reardon M.J.,
- Van Mieghem N.M.,
- Popma J.J.,
- et al.
- Dvir D.,
- Webb J.,
- Brecker S.,
- et al.
- Carroll J.D.,
- Edwards F.H.,
- Marinac-Dabic D.,
- et al.
- Leon M.B.,
- Piazza N.,
- Nikolsky E.,
- et al.
- Kappetein A.P.,
- Head S.J.,
- Genereux P.,
- et al.
- Zoghbi W.A.,
- Chambers J.B.,
- Dumesnil J.G.,
- et al.
- Webb J.G.,
- Mack M.J.,
- White J.M.,
- et al.
- Deeb G.M.,
- Chetcuti S.J.,
- Reardon M.J.,
- et al.
- Nombela-Franco L.,
- del Trigo M.,
- Morrison-Polo G.,
- et al.