Author + information
- Received August 10, 2018
- Revision received October 8, 2018
- Accepted October 15, 2018
- Published online January 7, 2019.
- Uri Landes, MDa,b,
- Zaza Iakobishvili, MDa,b,
- Daniella Vronsky, MDa,
- Oren Zusman, MDa,b,
- Alon Barsheshet, MDa,b,
- Ronen Jaffe, MDc,
- Ayman Jubran, MDc,
- Sung-Han Yoon, MDd,
- Raj R. Makkar, MDd,
- Maurizio Taramasso, MDe,
- Marco Russo, MDe,
- Francesco Maisano, MDe,
- Jan-Malte Sinning, MDf,
- Jasmin Shamekhi, MDf,
- Luigi Biasco, MDg,
- Giovanni Pedrazzini, MDg,
- Marco Moccetti, MDg,
- Azeem Latib, MDh,
- Matteo Pagnesi, MDh,
- Antonio Colombo, MDh,
- Corrado Tamburino, MDi,
- Paolo D' Arrigo, MDi,
- Stephan Windecker, MDj,
- Thomas Pilgrim, MDj,
- Didier Tchetche, MDk,
- Chiara De Biase, MDk,
- Mayra Guerrero, MDl,
- Omer Iftikhar, MDl,
- Johan Bosmans, MDm,
- Edo Bedzra, MDn,
- Danny Dvir, MDn,
- Darren Mylotte, MDo,
- Horst Sievert, MDp,
- Yusuke Watanabe, MDq,
- Lars Søndergaard, MDr,
- Hanna Dagnegård, MDr,
- Pablo Codner, MDa,b,
- Susheel Kodali, MDs,
- Martin Leon, MDs and
- Ran Kornowski, MDa,b,∗ ()
- aCardiology Department, Rabin Medical Center, Petah Tikva, Israel
- bCardiology Department, Carmel Medical Center, Haifa, Israel
- cCardiology Division, Cedars-Sinai Medical Center, Los Angeles, California
- dCardiology Department, University Hospital of Zurich, Zurich, Switzerland
- eCardiology Department, University Hospital Bonn, Bonn, Germany
- fFondazione Cardiocentro Ticino, Lugano, Switzerland
- gSan Raffaele Scientific Institute, Milan, Italy
- hCardiology Department, Ferrarotto Medical Center, Catania, Italy
- iCardiology Department, Bern University Hospital, Bern, Switzerland
- jCardiology Department, Clinique Pasteur, Toulouse, France
- kCardiology Department, Evanston Hospital, Evanston, Illinois
- lCardiology Department, Antwerp University Hospital, Antwerp, Belgium
- mCardiology Department, University of Washington Medical Center, Seattle, Washington
- nCardiology Department, University Hospital and National University of Ireland, Galway, Ireland
- oCardiovascular Center Frankfurt, Frankfurt, Germany
- pTeikyo University School of Medicine, Tokyo, Japan
- qCardiology Department, Rigshospitalet, Copenhagen, Denmark
- rCardiology Division, Columbia University Medical Center, New York, New York
- sSackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
- ↵∗Address for correspondence:
Prof. Ran Kornowski, Cardiology Department, Rabin Medical Center, Derech Ze’ev Jabotinski 39, Petah Tikva 4941492, Israel.
Objectives The authors sought to collect data on contemporary practice and outcome of transcatheter aortic valve replacement (TAVR) in oncology patients with severe aortic stenosis (AS).
Background Oncology patients with severe AS are often denied valve replacement. TAVR may be an emerging treatment option.
Methods A worldwide registry was designed to collect data on patients who undergo TAVR while having active malignancy. Data from 222 cancer patients from 18 TAVR centers were compared versus 2,522 “no-cancer” patients from 5 participating centers. Propensity-score matching was performed to further adjust for bias.
Results Cancer patients’ age was 78.8 ± 7.5 years, STS score 4.9 ± 3.4%, 62% men. Most frequent cancers were gastrointestinal (22%), prostate (16%), breast (15%), hematologic (15%), and lung (11%). At the time of TAVR, 40% had stage 4 cancer. Periprocedural complications were comparable between the groups. Although 30-day mortality was similar, 1-year mortality was higher in cancer patients (15% vs. 9%; p < 0.001); one-half of the deaths were due to neoplasm. Among patients who survived 1 year after the TAVR, one-third were in remission/cured from cancer. Progressive malignancy (stage III to IV) was a strong mortality predictor (hazard ratio: 2.37; 95% confidence interval: 1.74 to 3.23; p < 0.001), whereas stage I to II cancer was not associated with higher mortality compared with no-cancer patients.
Conclusions TAVR in cancer patients is associated with similar short-term but worse long-term prognosis compared with patients without cancer. Among this cohort, mortality is largely driven by cancer, and progressive malignancy is a strong mortality predictor. Importantly, 85% of the patients were alive at 1 year, one-third were in remission/cured from cancer. (Outcomes of Transcatheter Aortic Valve Implantation in Oncology Patients With Severe Aortic Stenosis [TOP-AS]; NCT03181997)
Transcatheter aortic valve replacement (TAVR) is currently an accepted therapy for symptomatic patients with severe aortic stenosis (AS) who have intermediate or greater surgical risk as assessed by the heart team (1). The benefit of TAVR may be limited among patients with short life expectancy due to noncardiac conditions. Thus, patients with an estimated life expectancy <2 years due to malignancies were excluded from major TAVR trials (1–5). As cancer treatment improves, some patients with severe AS and malignancy, including advanced metastatic diseases, may be more threatened by their valvular disease (if left untreated) than by their cancer (6). In addition, TAVR might allow a more aggressive and optimal oncology treatment, whether operative or via pharmacotherapy and/or radiation treatment. The heart and cardio-oncology teams are then compelled to make difficult, individualized choices, integrating life expectancy and quality-of-life assumptions. In our current investigation, we therefore sought to collect data on cancer patients who underwent TAVR due to severe AS. Describing this group’s characteristics, clinical outcome, and cause of death is expected to improve the decision-making process among these patients.
The TOP-AS (TAVR in Oncology Patients with severe AS) registry is an investigator-initiated registry commenced in December 2016, designed to collect data on patients who undergo TAVR (in native aortic valves, with any transcatheter heart valve type) while having active malignancy (all types, excluding nonmelanoma skin cancer). Transcatheter valves implanted in positions other than the aortic valve or valve-in-valve procedures were not included. As a reference group, key data were also collected from 5 centers participating in the TOP-AS registry regarding their same time-frame “all-comers” TAVR patients’ characteristics and outcomes (Online Table 1). Data collection started retrospectively for cases performed before registry initiation and has been continued prospectively thereafter. Data were collected for cases performed between January 2008 and December 2016 using a dedicated case report form. As for the present report, a total of 18 centers from Europe, North America, Japan, and the Middle East contributed their patient-level data (Online Table 1). All inconsistencies were resolved directly by communicating with the local investigators. The inclusion of patients was approved in each center by a local ethics committee.
Baseline demographics, cancer type and time of diagnosis, cancer stage, laboratory data (hemoglobin, platelet count, serum creatinine, and albumin), oncological treatment (including chemotherapy, biological and hormone treatment, chest radiation, and surgery), treatment goal (palliative or curative), oncological outcome (disease stage, progression, regression, remission, cure, recurrence, and associated conditions, i.e., cardiotoxicity) were collected by the coinvestigators at each institution. Cancer diagnosis was taken from the patient medical history and restrictively included patients with active, noncured cancer disease. Patients with suspicious incidental findings that were discovered by computed tomography (CT) evaluation before TAVR were included in our registry only if diagnosis of cancer has been confirmed. Data collection and monitoring regarding TAVR outcomes were assessed according to the Valve Academic Research Consortium 2 (VARC 2) definitions (7).
Principal endpoints included 30-day, 1-year, and last follow-up all-cause mortality and cause of death, changes in New York Heart Association (NYHA) functional class, performance of the transcatheter valve (gradients and presence of paravalvular leakage), periprocedural bleeding, vascular complications, kidney injury, and stroke. For surgical risk evaluation, both the European System for Cardiac Operative Risk Evaluation score (EuroSCORE II) and Society of Thoracic Surgeons (STS) predicted risk of mortality score were used. The cause of death was determined as cardiovascular (CV) or non-CV. All deaths due to non-CV causes were determined as cancer related or non-cancer related as reported and validated by the coinvestigator at each participating institution. Because we expected cancer and no-cancer patients to have different characteristics, we also planned to compare outcomes in propensity-matched cohorts.
Results are presented as mean ± SD for normally distributed continuous variables, as median (interquartile range [IQR]) for non-normally distributed continuous variables, and as number (percentage) for categorical data. The Student’s t-test was used to compare normally distributed continuous variables and the Wilcoxon rank sum test was used for variables not normally distributed. The chi-square and Fisher exact tests were used to compare categorical variables. To account for missing data, we used multiple imputation and pooled the results across imputations. To estimate STS score–adjusted 30-day mortality, we used the predicted probabilities derived from a logistic regression model where the dependent variable was 30-day mortality and the independent variable was each patient’s calculated STS risk score. The Mann-Kendall test was used to compare year trends. Time-to-event curves were drawn using the Kaplan-Meier method, for 1 of the imputed sets. Cox multivariable analysis for mortality in cancer and noncancer patients was used with hazard ratios (HRs) with 95% confidence intervals (CIs). Propensity-score (PS) matching was performed by a 1:4 ratio (cancer vs. no-cancer) using the “nearest neighbor” method, without replacement, restricted by a caliper of 0.3 standard deviations of the distance measure. Covariates for the PS included age, sex, baseline STS score, and serum creatinine level, history of chronic obstructive pulmonary disease, cerebrovascular events, anemia, hypertension, dyslipidemia, diabetes, percutaneous coronary intervention, myocardial infarction, frailty, and baseline NYHA functional class. After PS matching, comparison of the 2 groups was performed using the tests listed in the preceding text, with Cox regression performed with the Huber sandwich estimator to account for matched data. For all regression analyses, variables were chosen according to subject matter without imposing selection, and models were tested for multicollinearity. A 2-sided p < 0.05 was considered statistically significant. We calculated that with power of 0.8, significance level of 0.05, and an effect size of 0.1, we would need a total of 1,090 patients, although given the retrospective nature, we sought to gather as much patients as we could. All analyses were performed using R and relevant packages (8–11).
During the study period, there were 8,497 TAVR procedures conducted at the 18 participating sites. During that time, 222 TAVR procedures were performed in patients with active malignancy and 2,522 in patients with no cancer at the 5 centers` reference group. Thirty-day and 1-year survival status was available in 100% and 98.3% of patients, respectively. The median number of cancer patients per center was 9 (IQR: 5.25 to 19). Over time, there was a rise in the number of overall TAVR cases and in the number of TAVR in cancer patient cases, as well as in the portion of cancer patients among the overall TAVR cohort in the 18 participating centers (from 0.57% in 2008 to 4% in 2016; p = 0.037) (Online Figure 1). The median post-TAVR follow-up duration was 330 (IQR: 118 to 656) days. The median cancer diagnosis to TAVR time interval was 195 (IQR: 64 to 834) days. Of note, 12% of the malignancies were incidentally diagnosed during the CT scan done as part of the pre-TAVR patient assessment.
Data regarding baseline patients’ malignancy characteristics are presented in Table 1. The most prevalent malignancies, in order of frequencies, were gastrointestinal (22%), prostate (16%), breast (15%), hematologic (15%), lung (11%), and urinary tract (5%) cancers. At the time of TAVR, one-half of the patients were at advanced cancer stage (III to IV), 31% had metastasis, 29% were receiving antineoplastic therapy (biological, hormonal, or chemotherapy), and 26% had an indication for downstream cancer-related surgery. The goal of therapy was defined as palliative in 37% and as curative in 55% of the patients (8% undecided).
Compared with patients from the control group, cancer patients undergoing TAVR were more frequently assessed as frail (38% vs. 17%; p < 0.001) and had serum albumin levels lower than 3.5 g/dl more often (18.4% vs. 6.25%; p < 0.001). Yet, cancer patients were younger (78.8 ± 7.5 vs. 81.3 ± 7.1; p < 0.001), had lower STS risk (4.9 ± 3.4% vs. 6.2 ± 4.4%; p < 0.001), and had fewer CV comorbidities other than AS (Table 2).
Transfemoral TAVR access was used in 213 (96%) and 1,194 (87%) of cancer and no-cancer patients, respectively (p < 0.001), and self-expandable valves were used in 137 (62%) and 1,942 (77%) of cancer and no-cancer patients, respectively (p = 0.012). A more detailed examination of the technical and procedural data of cancer patients compared with no-cancer patients is provided in the Appendix (Online Table 2).
In-hospital outcomes and VARC-2 complication rates are provided in Table 3. The STS-adjusted mortality was similar between patients with and without cancer (2.7% vs 3.0%; p = 0.95). Periprocedural complication rates (i.e., stroke, myocardial infarction, kidney injury, and so on) were similar between patients with or without cancer, excluding higher rate of bleeding among cancer patients (any bleeding 14.4% vs. 6.9%; p < 0.001; major bleeding 8.6% vs. 3.1%; p < 0.001), despite a similar rate of major vascular complications (4.1% vs. 2.4%; p = 0.3). Platelet count was similar between patients with or without cancer (217 ± 78 k/μl vs. 227 ± 79 k/μl; p = 0.6), as were the rate of bleeding and platelet count between patients with hematologic (n = 31) versus nonhematologic (n = 191) cancer (data not shown). Data regarding the source of bleeding (i.e., gastrointestinal, genitourinary, and so on) was unattainable for analysis.
Adjusted mortality at 1 year was significantly higher among cancer patients: 33 (14.8%) vs. 237 (9.4%); p = 0.01. Among cancer patients, 75% of deaths (24 cases) were due to non-CV reasons. One-half of the deaths in this group were related to cancer (n = 17).
Using PS matching across imputation, 215 cancer patients were matched to 734 TAVR patients with no cancer. Baseline covariates were evenly distributed across the matched groups (standardized mean difference <0.07 for all). The baseline characteristics of the matched cohorts are summarized in the Appendix (Online Table 3). No significant difference was observed between the groups in terms of in-hospital and 30-day mortality, cerebrovascular events, bleeding events, vascular complications, need for a pacemaker, or acute kidney injury. Compared with the control group, mortality at 1 year was higher among the PS-matched cancer patient group: 32 (14.9%) versus 66 (9.0%); p < 0.001, with lower survival among cancer patients during follow-up (HR: 2.37, 95% CI: 1.74 to 3.23; p < 0.001). Using Cox regression analysis in the matched groups, progressive cancer (stage III to IV) at the day of TAVR was associated with high risk for mortality (HR: 3.21, 95% CI: 2.35 to 4.35; p < 0.001) whereas stage I to II cancer was not associated with higher mortality risk (HR: 1.31, 95% CI: 0.96 to 1.78; p = 0.3) compared with patients with no-cancer. Survival curves for the unmatched and matched groups, in addition to survival with respect to cancer stage is presented in Figure 1.
Prosthetic valve function and symptomatic outcome
Both at 1 month and 1 year, post-procedural transaortic valve gradients were reduced and low in cancer patients, although higher in cancer patients compared with no-cancer patients: mean aortic valve gradient (in mm Hg): 10.71 ± 7.26 versus 8.39 ± 5.44 at 30 days; p < 0.001, and 11.77 ± 5.95 vs. 8.13 ± 5.25 at 1 year; p < 0.001, respectively. Similarly, the proportion of patients with ≤mild aortic regurgitation was lower in cancer patients as compared with patients without cancer (Table 4). One-half of the patients with cancer were free of heart failure symptoms (NYHA functional class I) at 1 month after the TAVR, and this benefit persisted during 1-year follow-up (Figure 2).
Malignancy state at 1 year
During the 1-year follow-up after the TAVR, 17% of the cancer patients had been treated with chemotherapy, 16.5% with biological anticancer therapy, and 21% (80% of patients with an upstream surgery indication) underwent surgery targeted for their oncological disease. Among the patients who survived 1 year after TAVR, the cancer progressed in 29% and regressed in 9%, whereas 21% were in remission and another 14% were cured from their oncological disease.
In our study, we describe the contemporary, international level of practice and outcome of TAVR in oncology patients with severe AS. The central findings are: 1) throughout the years, the portion of cancer patients among TAVR recipients increased; 2) most prevalent malignancies were gastrointestinal, prostate, breast, hematologic, lung, and urinary tract, one-half of the patients were at an advanced (III/IV) cancer stage; 3) compared with patients without cancer, oncology patients were more frequently frail, even though they were younger, had lower STS risk, and had fewer CV comorbidities; 4) TAVR seemed safe in oncology patients, with similar short-term mortality and periprocedural complication rates as in patients without cancer; 5) at 1 year, mortality was higher among cancer patients, with one-half of deaths in this group being cancer related; 6) advanced cancer (stage III to IV) at baseline was the strongest predictor of late mortality, whereas limited (stage I to II) cancer was not independently associated with mortality; and 7) among the 85% patients who survive 1 year after the TAVR, one-third entered remission or were cured of cancer.
The transformative innovation of TAVR has provided a tremendous opportunity to treat many patients with severe AS, but also raised awareness for important futility questions which needs to be further stressed when addressing patients with cancer. One should bear in mind that untreated severe symptomatic AS is a malignant disease by itself in terms of its dreadful prognosis. In patients with cancer, AS creates a greater misfortune, as it may interfere with optimal antineoplastic therapy (i.e., high-risk oncological surgery or oncological drug intolerance). The European Society of Cardiology position paper on cancer treatments and CV toxicity (12) recommend afterload reduction (using angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers) for attenuation of left ventricular dysfunction and heart failure induced by anthracyclines and other antineoplastic therapies. In AS patients, afterload reduction is only possible by aortic valve intervention. Balloon valvuloplasty has repeatedly failed to improve survival in AS patients and is associated with limited efficacy, complications, and high early restenosis rates (13,14). Although surgical aortic valve replacement (AVR) has been shown to improve survival in cancer patients with severe AS compared with conservative management (15), it was associated with increased perioperative mortality and morbidity compared with patients without cancer (16). Additionally, the invasive nature of open heart surgery and cardiopulmonary bypass make surgical AVR less suitable for many “real life” cancer patients (17). TAVR may be an optimal treatment strategy in selected oncology patients with severe AS. At least theoretically, TAVR may address many of the concerns associated with open heart surgery in cancer patients, such as increased risk for bleeding and infections, and the suspension of anticancer therapy during the post-surgical recovery window (16,18). In addition, TAVR may allow a more aggressive and optimal cancer treatment soon after the procedure.
In the present study, the mortality rate 1 year after the TAVR was nearly 2 times higher in cancer patients compared with no-cancer patients. Higher mortality appeared to be driven by progressive cancer stages (III to IV). Therefore, according to the present study, it seems more justified to treat patients in early cancer stages with TAVR, but may be less justified in late cancer stages. On the other hand, the overall 85% cancer patients’ survival rate is still better than conventional post-TAVR survival benchmarks such as the PARTNER high-risk trial (Placement of AoRTic TraNscathetER Valve Trial) and FRANCE 2 (French Aortic National CoreValve and Edwards) registry (with 1-year survival rate of 75% in both) (19,20). Current guidelines state that patients with <12 month’s predicted life expectancy (due to nonvalvular cause) should be excluded from TAVR. In retrospect, predicted post-TAVR life expectancy >12 months was correct (in terms of cancer-related death) 92.3% of times, although 50% of cancer patients’ deaths were cancer related. Therefore, it seems that the heart team in conjunction with the oncologists should make individual multifaceted decisions about the treatment options, while recognizing that advanced cancer before TAVR should not unequivocally exclude patients but is a strong predictor of subsequent mortality.
In addition to survival, a question remains whether TAVR improves symptoms and quality of life in cancer patients, because their symptoms may at times be multifactorial, less specific, and overlapping with paraneoplastic symptoms. Both at 1 month and 1 year, cancer patients experienced a significant and persistent symptomatic benefit with regard to NYHA functional class, albeit less pronounced compared with no-cancer patients. Similarly, post-TAVR echocardiographic outcomes were less optimal in cancer versus no-cancer patients: mean aortic valve gradient (in mm Hg): 10.71 ± 7.26 versus 8.39 ± 5.44; p < 0.001, and ≤mild aortic regurgitation in 89.4% versus 93.8%; p < 0.001, respectively. Although these hemodynamic differences are statistically significant, they are too small to serve as a plausible explanation for the discrepancy in symptoms. Symptoms are subjected to bias and may be influenced by dissimilarities in clinical assessments. Unfortunately, other quality-of-life data (e.g., EQ-5D or Minnesota Living With Heart Failure questionnaire) were not available for analysis. Consequently, whether a correlation exists between cancer state and symptomatic benefit remains unknown. Irrespective of this, one should bear in mind that AS symptoms may indeed be more multifactorial in cancer patients and not merely caused by the stenotic valve.
The data on TAVR in oncological patients are scarce. Mangner et al. (21) reported the outcome of 99 cancer patients treated with TAVR at a single center. Consistent with our findings, 30-day outcome showed TAVR to be as safe and effective as in no-cancer patients, but 1-year mortality was higher in patients with cancer compared with no-cancer patients, and as in our study, was mainly driven by non-CV mortality. By contrast, Watanabe et al. (22) reported similar 1-year mortality rate in cancer and no-cancer Japanese patients that underwent TAVR. Due to the small number of patients (N = 47), it is likely that the study was underpowered to find a mortality difference. Also, geographic and ethnic discrepancies might have contributed to this contrast. Compared with these 2 previous studies, the present study collected a larger number of patients from a designated registry of 18 centers worldwide, thus supports better validation and allows a more comprehensive data analysis (i.e., looking into death cause, malignancy state at 1 year, PS matching, and so on).
Our findings should be interpreted with caution and should be considered as hypothesis generating. First, our registry is built with inherent selection bias because it has collected a miscellany of patients with cancer and severe AS carefully chosen for TAVR by the local heart and oncology teams of experts. The selection bias may also be center specific because results can also be influenced by differences in disease assessment, mode of practice, and documentation patterns at participating institutions. Even though, reference patients’ enrollment was only practical from 5 centers. Second, the study population represents a widely diverse heterogeneous group of cancer patients. Hence, it is difficult to stratify them according to each cancer type. To obtain some estimates about prognosis, we have stratified patients according to cancer stage, and the data indeed showed a major influence on clinical outcomes. A more detailed subgroup analysis of specific cancer types could not yet be adequately performed and will be a target for ongoing prospective data collection of the TOP-AS registry. Third, it would be most interesting to know whether cancer patients with severe AS live longer if they undergo TAVR compared with conservative treatment. However, this issue is beyond the scope of our study. Fourth, important data were unattainable for analysis. Additional anatomic (particularly CT) data could shed more light on the hemodynamic result in the cancer cohort (i.e., more bicuspid aortic valve?) and data regarding the source of bleeding (i.e., gastrointestinal, genitourinary, and so on) could explain better the higher bleeding rate in cancer versus no-cancer patients. Because vascular complication rates were similar and because this difference was no longer present after PS matching, we hypothesized that it could be an accidental finding (because the number of bleeding events was small: 19 major, 32 any). With that, other explanations such as dysfunctional platelet function in cancer patients should also be considered. Finally, despite being the largest registry examining TAVR in cancer patients, the number of cancer patients is relatively low (because TAVR in cancer patients is still infrequent) and imbalance in size and baseline characteristics exist between the study groups. We tried to overcome these differences by multivariate adjustment and PS matching, but other confounders may still influence the results.
Treating severe AS with TAVR in oncology patients appears effective and safe in the short term, but carries a worse 1-year prognosis compared with patients without cancer. Among this cohort of patients, mortality was largely due to cancer, yet 85% of patients were alive at 1 year after the procedure, one-third either in remission or cured of their cancer. The heart team should make individual case decisions while recognizing that advanced cancer is a strong predictor of subsequent mortality.
WHAT IS KNOWN? Oncological patients with severe AS are often denied surgical valve replacement. TAVR may be an emerging treatment option.
WHAT IS NEW? Compared with the propensity-matched control group of no-cancer patients, oncological patients who underwent TAVR exhibited similar 30-day outcome, yet the mortality rate at 1 year was doubled and mostly related to cancer. Progressive cancer was a strong mortality predictor.
WHAT IS NEXT? TAVR in selected oncology patients is associated with worse, yet acceptable, 1-year outcome. The heart team should make individual treatment decisions, recognizing that progressive malignancy is the strongest predictor for death at 1 year. Further study should focus on evaluating TAVR versus palliative care for AS in patients with stage III to IV cancer.
Dr. Guerrero is currently affiliated with the Department of Cardiovascular Medicine, Mayo Clinic Hospital, Rochester, Minnesota. Dr. Makkar has been a consultant to or received research funding from Cordis, Medtronic, Cedars-Sinai Medical Center, Abbott, and Edwards Lifesciences. Dr. Taramasso is a consultant for Abbott, Boston Scientific, 4tech and CoreMedic; and received speaker fees from Edwards Lifesciences. Dr. Sinning has received speaker honoraria and research grants from Medtronic, Edwards Lifesciences, and Boston Scientific. Dr. Latib has served on advisory boards for Medtronic and Abbott Vascular. Dr. Windecker has received institutional research grants from Abbott, Amgen, Boston Scientific, Biotronik, Edwards Lifesciences, Medtronic, St. Jude Medical, and Terumo. Dr. Pilgrim has received institutional research grants from Biotronik, Boston Scientific, and Edwards Lifesciences; and speaker fees from Biotronik and Boston Scientific. Dr. Guerrero has received research funding from Edwards Lifesciences; and has served on the Speakers Bureau for Boston Scientific. Prof. Sievert has received study honoraria, travel expenses, and consulting fees from 4tech Cardio, Abbott, Ablative Solutions, Ancora Heart, Bavaria Medizin Technologie GmbH, Bioventrix, Boston Scientific, Carag, Cardiac Dimensions, Celonova, Comed B.V., Contego, CVRx, Edwards Lifesciences, Endologix, Hemoteq, Lifetech, Maquet Getinge Group, Medtronic, Mitralign, Nuomao Medtech, Occlutech, pfm Medical, Recor, Renal Guard, Rox Medical, Terumo, Vascular Dynamics, and Vivasure Medical. Dr. Watanabe has served as a proctor for Edwards Lifesciences and Medtronic. Dr. Kodali is a consultant for Abbott Vascular, Merrill Lifesciences, Claret Medical; is on the advisory board for Dura Biotech, Thubrikar Aortic Valve Inc., Biotrace Medical; received honoraria from Abbott Vascular, Merrill Lifesciences, and Claret Medical; and received equity from Thubrikar Aortic Valve, Inc., Dura Biotech, BioTrace Medical, and Microinterventional Devices. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- aortic stenosis
- confidence interval
- computed tomography
- hazard ratio
- interquartile range
- New York Heart Association
- propensity score
- Society of Thoracic Surgeons
- transcatheter aortic valve replacement
- Received August 10, 2018.
- Revision received October 8, 2018.
- Accepted October 15, 2018.
- 2019 American College of Cardiology Foundation
- Baumgartner H.,
- Falk V.,
- Bax J.J.,
- et al.
- Reardon M.J.,
- Adams D.H.,
- Kleiman N.S.,
- et al.
- Kappetein A.,
- Head S.,
- Généreux P.,
- et al.
- R Core Team
- ↵Dowle M, Srinivasan A, Gorecki J, et al. data.table: extension of Data.frame. 2018. Available at: https://cran.r-project.org/web/packages/data.table/index.html. Accessed September 30, 2018.
- ↵van Buuren S, Groothuis-Oudshoorn K. mice: multivariate imputation by chained equations in R. J Stat Softw. Available at: https://www.jstatsoft.org/article/view/v045i03. Accessed December 26, 2016.
- Ho D.E.,
- Imai K.,
- King G.,
- Stuart E.A.
- Lieberman E.B.,
- Bashore T.M.,
- Hermiller J.B.,
- et al.
- Bach D.S.,
- Cimino N.,
- Deeb G.M.
- Mangner N.,
- Woitek F.J.,
- Haussig S.,
- et al.
- Watanabe Y.,
- Kozuma K.,
- Hioki H.,
- et al.