Author + information
- Received July 24, 2017
- Revision received September 26, 2017
- Accepted September 28, 2017
- Published online January 1, 2018.
- Alaide Chieffo, MDa,
- Anna Sonia Petronio, MDb,
- Julinda Mehilli, MDc,
- Jaya Chandrasekhar, MBBS, MSd,
- Samantha Sartori, PhDd,
- Thierry Lefèvre, MDe,
- Patrizia Presbitero, MDf,
- Piera Capranzano, MDg,
- Didier Tchetche, MDh,
- Alessandro Iadanza, MDi,
- Gennaro Sardella, MDj,
- Nicolas M. Van Mieghem, MD, PhDk,
- Emanuele Meliga, MDl,
- Nicholas Dumonteil, MDm,
- Chiara Fraccaro, MD, PhDn,
- Daniela Trabattoni, MDo,
- Ghada Mikhail, MDp,
- Samin Sharma, MDq,
- Maria Cruz Ferrer, MDr,
- Christoph Naber, MDs,
- Peter Kievit, MDt,
- Usman Baber, MD, MSd,
- Clayton Snyder, BScd,
- Madhav Sharma, MBBSd,
- Marie Claude Morice, MDe,
- Roxana Mehran, MDd,∗ (, )
- on behalf of the WIN-TAVI Investigators
- aDepartment of Cardiology, San Raffaele Scientific Institute, Milan, Italy
- bDepartment of Cardiology, AOUP Cisanello, University Hospital, Pisa, Italy
- cDepartment of Cardiology, Ludwig-Maximilians-University of Munich, Munich, Germany
- dThe Zena and Michael A. Wiener Cardiovascular Institute, Icahn School of Medicine at Mount Sinai, New York, New York
- eDepartment of Cardiology, Institut Hospitalier Jacques Cartier, Ramsay Générale de Santé, Massy, France
- fDepartment of Cardiology, Istituto Clinico Humanitas, Milan, Italy
- gDepartment of Cardiology, University of Catania, Catania, Italy
- hDepartment of Cardiology, Clinique Pasteur, Toulouse, France
- iDepartment of Cardiology, Azienda Ospedaliera Universitaria Senese, Policlinico Le Scotte, Siena, Italy
- jDepartment of Cardiology, Policlinico “Umberto I,” Sapienza University of Rome, Rome, Italy
- kDepartment of Cardiology, Erasmus Medical Center, Thoraxcenter, Rotterdam, the Netherlands
- lDepartment of Cardiology, Mauriziano Hospital, Turin, Italy
- mDepartment of Cardiology, Rangueil University Hospital, Toulouse, France
- nDepartment of Cardiology, University of Padova, Padova, Italy
- oDepartment of Cardiology, Centro Cardiologico Monzino, Milan, Italy
- pDepartment of Cardiology, Imperial College Healthcare NHS Trust, Hammersmith Hospital, London, United Kingdom
- qDepartment of Cardiology, Mount Sinai Hospital, New York, New York
- rDepartment of Cardiology, Hospital Universitario Miguel Servet, Zaragoza, Spain
- sDepartment of Cardiology, Contilia Heart and Vascular Centre, Elisabeth Krankenhaus, Essen, Germany
- tDepartment of Cardiology, Radboud University Nijmegen Medical Center, Nijmegen, the Netherlands
- ↵∗Address for correspondence:
Dr. Roxana Mehran, Zena and Michael A. Wiener Cardiovascular Institute, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1030, New York, New York 10029-6574.
Objectives This study sought to examine the safety and performance of contemporary transcatheter aortic valve replacement (TAVR) in an exclusive all-women TAVR population, and to further investigate the potential impact of female sex–specific characteristics on composite 1-year clinical outcomes.
Background Women comprise ≥50% patients undergoing TAVR. Several data have shown the noninferiority of TAVR compared with surgical aortic valve replacement for symptomatic significant aortic stenosis, but no study so far has been specifically powered to detect differences by sex.
Methods The WIN-TAVI (Women’s INternational Transcatheter Aortic Valve Implantation) registry is a multinational, prospective, observational registry of women undergoing TAVR for significant aortic stenosis, across 18 sites in Europe and 1 site in the United States, between January 2013 and December 2015. The primary Valve Academic Research Consortium (VARC)-2 efficacy endpoint was a composite of mortality, stroke, myocardial infarction, hospitalization for valve-related symptoms or heart failure or valve-related dysfunction beyond 30 days. Secondary endpoints included composite 1-year death or stroke. Predictors of 1-year outcomes were determined using Cox regression methods.
Results A total of 1,019 intermediate to high-risk women, with mean age 82.5 ± 6.3 years, mean European System for Cardiac Operative Risk Evaluation (EuroSCORE) I 17.8 ± 11.7% and mean Society of Thoracic Surgeons score 8.3 ± 7.4% were enrolled. TAVR was performed via transfemoral access in 90.6% and new-generation devices were used in 42.1%. The primary VARC-2 efficacy composite endpoint occurred in 111 (10.9%) patients beyond 30 days and in 167 (16.5%) patients at 1 year. The incidence of 1-year death or stroke was 13.9% (n = 141). Death occurred in 127 (12.5%) patients and stroke in 22 (2.2%) patients. Prior coronary revascularization (hazard ratio [HR]: 1.72; 95% confidence interval [CI]: 1.17 to 2.52; p = 0.006) and EuroSCORE I (HR: 1.02; 95% CI: 1.00 to 1.04; p = 0.027) were independent predictors of the VARC-2 efficacy endpoint. Similarly, EuroSCORE I (HR: 1.02; 95% CI: 1.00 to 1.04; p = 0.013), baseline atrial fibrillation (HR: 1.58; 95% CI: 1.07 to 2.33; p = 0.022), and prior percutaneous coronary intervention (HR: 1.50; 95% CI: 1.03 to 2.19; p = 0.035) were independent predictors of 1-year death or stroke. After adjustment, no significant association was observed between history of pregnancy or any sex-specific factors and 1-year TAVR outcomes.
Conclusions Intermediate to high-risk women enrolled in this first ever all-women contemporary TAVR registry experienced a 1-year VARC-2 composite efficacy endpoint of 16.5%, with a low incidence of 1-year mortality and stroke. Prior revascularization and EuroSCORE I were independent predictors of the VARC-2 efficacy endpoint, whereas EuroSCORE I, baseline atrial fibrillation, and prior percutaneous coronary intervention were independent predictors of the 1-year death or stroke.
- female-specific characteristics
- first female registry
- transcatheter aortic valve replacement
- 1-year outcomes
Transcatheter aortic valve replacement (TAVR) is now proven to be a suitable alternative treatment to surgical aortic valve replacement (SAVR) for severe aortic stenosis (AS) in both high- or intermediate-risk patients (1–4). In most randomized clinical trials and observational registries evaluating TAVR, women represent ≥50% of the patient population, and despite older age and significantly greater early vascular and bleeding complications, women have consistently shown superior 1-year survival compared with men (5–7). Furthermore, the results of both PARTNER (Placement of AoRTic TraNscathetER valve) trials (PARTNER 1 and PARTNER 2) suggest superiority of TAVR compared with SAVR for women, with lower rates of 2-year death in the PARTNER 1 trial among women undergoing transfemoral TAVR versus SAVR (23.4% vs. 36.9%; p = 0.02) and lower incidence of 2-year death or disabling stroke in the PARTNER 2 trial (16.8% vs. 20.4%; p = 0.05) with transfemoral TAVR versus SAVR attributed to better outcomes in women (3,8).
Women and men undergoing TAVR have distinct baseline characteristics, which can have a differential impact on post-TAVR and long-term events (7). Compared with men, women have smaller aortic valve annuli, lower origins of coronary arteries, smaller peripheral vessels (9), as well as higher prevalence of osteoporosis (10) and frailty (11), and greater risk of bleeding (5,7). Women also have greater prevalence of coexistent valve disease and heart failure (7). The understanding and appreciation of these risk differences provided an urgent impetus to study women undergoing TAVR in better detail, and to investigate whether female sex–specific characteristics contributed to TAVR outcomes. Indeed, female specific factors such as pregnancy and its complications (12) and age of menopause (13,14) have been suggested to modify future risk for cardiovascular disease, but the role of these parameters in TAVR patients has never hitherto been investigated.
The aim of this principal report from the WIN-TAVI (Women’s INternational Transcatheter Aortic Valve Implantation) registry (15) is to present the 1-year primary Valve Academic Research Consortium (VARC)-2 efficacy endpoint in women undergoing TAVR and to explore for the predictors of the VARC-2 efficacy endpoint and other important composite TAVR outcomes. In addition, we aimed to evaluate whether the observed inverse association between previous history of pregnancy and the 30-day VARC-2 safety endpoint (15) extended to 1-year outcomes.
The WIN-TAVI registry (NCT01819181) is an international, multicenter, prospective, observational registry of women undergoing TAVR at 18 European centers and 1 North American center treated with commercially available and approved TAVR devices for the treatment of severe symptomatic AS (15). Participating centers were selected based on review of site survey responses to determine the total number of TAVR procedures performed at each center (minimum of 50) and the planned number of TAVR to be performed in the following year. All sites had institutional approval from the local ethics committee and the study was conducted according to the principles of the Declaration of Helsinki, International Organization for Standardization Guidelines, and Good Clinical Practice Guidelines.
All patients who met the eligibility criteria and provided written informed consent were enrolled in the study. The study was conducted without any external funding and was driven by the scientific interest of the investigators. The protocol and study endpoints were designed by the executive committee and principal investigators of the study (see the Online Appendix).
The eligibility criteria were indicated in detail in the primary paper (15). In brief, women with: 1) severe AS determined by echocardiography defined as mean gradient >40 mm Hg or peak jet velocity >4.0 m/s and an aortic valve area ≤0.8 cm2 or aortic valve area index ≤0.5 cm2/m2; and 2) symptoms of angina, congestive heart failure, New York Heart Association functional class ≥II, or syncope were considered for enrollment if no exclusions applied.
Additional inclusion criteria were based on high European System for Cardiac Operative Risk Evaluation (EuroSCORE) or presence of other comorbidities (e.g., chronic obstructive lung disease, need for home oxygen, porcelain aorta, previous thoracic radiotherapy, Childs-Pugh class B and C liver disease) leading to heart team decision for TAVR rather than SAVR. The chief exclusion criteria were untreated clinically significant (>70% obstruction) proximal vessel coronary artery disease (CAD) amenable to revascularization, hemodynamic instability (e.g., requiring inotropic support), active endocarditis or sepsis within 6 months before the study procedure, or use of an investigational device without Conformité Européenne mark.
TAVR procedure and clinical follow-up
Pre-screening included evaluation of medical history and standard of care diagnostic imaging (transthoracic or transesophageal echocardiogram or multidetector computed tomography measurements). We specifically collected information on female sex–related factors including menstrual history, history of pregnancy, pregnancy complications such as gestational hypertension or diabetes, ectopic pregnancy, history of breastfeeding, use of hormone replacement therapy, osteoporosis and use of osteoporosis treatment, gynecological or breast cancer, and gynecological surgery.
Procedural decisions regarding TAVR access, TAVR device type, use of pre- and post-dilation and interventional therapies was at the discretion of the treating physicians. Patient follow-up was conducted by phone contact or clinic visit at 1 month, 6 months, and 12 months after TAVR to record clinical status and occurrence of adverse events. Total planned follow-up is to 24 months. Per standard of care at the participating sites patients underwent neurological evaluation after TAVR, only if clinically indicated.
The clinical and data coordinating center for the study was at the Icahn School of Medicine at Mount Sinai (New York, New York), which was responsible for the monitoring of electronic data entry for accuracy of data, database and data management, and statistical analyses. All events were adjudicated by an independent clinical events committee using source documents provided by the sites. The study was endorsed by the Society for Cardiovascular Angiography and Interventions Women in Innovations initiative.
Study endpoints and definitions
The primary study endpoint for the 1-year analysis was the VARC-2 efficacy endpoint beyond 30 days—a composite of all-cause mortality, all stroke, myocardial infarction, hospitalizations for valve-related symptoms or worsening congestive heart failure or valve-related dysfunction (clinical presentation with valve thrombosis or endocarditis) (16).
Endpoints were adjudicated using the standardized VARC-2 criteria (16).
Because echocardiographic examination at 1-year clinical follow-up was not mandatory by study protocol but rather left to local practice, prosthetic valve dysfunction in the patients not undergoing echocardiographic follow-up was evaluated as clinical presentation leading to diagnosis of valve thrombosis or endocarditis.
History of pregnancy was defined as any history of pregnancy and not pregnancy resulting in a live birth. Frailty was subjective as judged by the heart team but the use of objective frailty scales was recommended.
Old-generation devices comprised Edwards SAPIEN XT (Edwards Lifesciences, Irvine, California) and Medtronic CoreValve (Medtronic, Minneapolis, Minnesota). All other prosthesis types are considered new-generation devices.
Categorical data are presented as frequencies and percentages and were compared using the chi-square or Fisher exact test. Continuous variables are presented as mean ± SD or medians and interquartile range and were compared using the Student t test or Wilcoxon signed rank test. Time-to-event curves were constructed using the Kaplan-Meier method and outcomes were compared using the log-rank test. Using Cox regression methods, we generated a multivariable model for predictors of the 1-year composite outcomes. Variables with univariate associations at a significance level of p < 0.20 were considered as candidate covariates to generate a multivariable model using stepwise selection (p for entry <0.10, p for exit <0.20). Country was entered as a random effect. All analyses were performed using Stata version 14.0 (StataCorp, College Station, Texas) and p values <0.05 were considered significant.
From January 2013 to December 2015, 1,019 women were enrolled across 18 centers in Europe and 1 center in North America. One-year follow-up was achieved in 98.8% of eligible patients (Figure 1). The baseline and procedural characteristics have previously been presented in detail (15). The study population included all women with a mean age of 82.5 ± 6.3 years, with mean EuroSCORE I 17.8 ± 11.7% and mean Society of Thoracic Surgeons score 8.3 ± 7.4%. Prevalence of baseline comorbidities included 26.1% patients with diabetes, 30.8% with chronic kidney disease, 22.9% with prior percutaneous coronary intervention (PCI), 20.0% with prior atrial fibrillation (AF), 7.5% with prior stroke, and 9.2% with peripheral arterial disease. The mean left ventricular ejection fraction (LVEF) was 55.7 ± 10.7%. A total of 738 (72.4%) patients had a history of pregnancy; patients without history of pregnancy were more likely to be considered frail on surgical assessment (70.0% vs. 61.3%; p = 0.01) and were more often current smokers (5.4% vs. 2.5%; p = 0.02) or had left main disease ≥50% (8.7% vs. 4.6%; p = 0.06).
The TAVR procedure was conducted under local anesthesia or conscious sedation in 64.2% of patients. TAVR was mainly performed via transfemoral access (90.6%). In 32% of patients the sheath size used was 16-F or smaller. New-generation devices were used in 42.1%.
1-year primary and secondary study endpoints
The clinical outcomes are shown in Table 1 and Figure 2. The primary VARC-2 composite efficacy endpoint beyond 30 days occurred in 111 (10.9%) patients. During 1-year follow-up the VARC-2 efficacy composite occurred in 167 (16.5%) patients. The baseline characteristics in patients with and without 1-year VARC-2 efficacy endpoint are shown in Online Table 1.
One-year death or stroke occurred in 141 (13.9%) patients. Death was reported in 127 (12.5%) patients, of which cardiac death occurred in 108 (10.6%) patients. One-year stroke occurred in 22 (2.2%) patients. Four patients experiencing stroke had baseline AF and 1 patient had new AF post-TAVR. Among patients experiencing stroke, where discharge medications were recorded (n = 18), 8 patients were discharged on an anticoagulant agent (6 on dual therapy with 1 antiplatelet and 1 anticoagulant agent, and 2 on an anticoagulant agent alone), 7 patients were discharged on dual-antiplatelet therapy (DAPT), and 3 patients were discharged on single antiplatelet therapy. There were 4 hemorrhagic strokes, 2 occurred in patients on dual therapy and 2 occurred in patients where discharge medications were not recorded.
One-year major vascular complications were observed in 83 (8.2%) patients, VARC life-threatening bleeding in 46 (4.5%) patients, and Bleeding Academic Research Consortium 3 or 5 bleeding in 131 (12.9%) patients. TAV-in-TAV implantation occurred in 18 (1.8%) patients, valve thrombosis in 3 (0.3%) patients, and endocarditis in 2 (0.2%) patients. Among patients undergoing echocardiography in follow-up, moderate to severe paravalvular aortic insufficiency (AI) was reported in 9.7% of the patients at 1 year. Any arrhythmia or conduction disturbance was observed in 232 (22.8%) patients after TAVR and new AF in 37 (3.6%) patients, whereas new permanent pacemaker implantation occurred in 129 (12.7%) patients. The incidence of hospitalizations for heart failure or valve-related symptoms during 1-year follow-up occurred in 3.2% of patients. Figure 3 shows the change in New York Heart Association functional class status over time, among survivors.
At 6 months, echocardiography was available in 275 patients; mean aortic valve gradient was 10.3 ± 5.5 mm Hg, mean peak gradient was 19.2 ± 9.6 mm Hg, mean aortic valve area was 1.79 ± 1.68 cm2, and LVEF was 57.3 ± 8.8%. At 1 year, echocardiography was available in 373 patients; mean aortic valve gradient was 10.8 ± 5.2 mm Hg, mean peak gradient was 20.2 ± 12.6 mm Hg, mean aortic valve area was 1.50 ± 0.40 cm2, and LVEF was 57.9 ± 8.7%. The incidence of aortic valve mean gradient >20 mm Hg at 6- or 12-month echocardiography was 4.4%, and the incidence of aortic valve area ≤0.9 cm2 was 6.0%. This included the 3 patients with valve thrombosis. No patients had severe paravalvular AI. Figure 4 shows the proportion of patients with different grades of paravalvular AI over time.
Figure 5 shows the frequency of patients on different antithrombotic therapies at discharge and at 1-year follow-up. Although 55.9% patients were discharged on DAPT, only 11.5% remained on DAPT at 1 year. The use of oral anticoagulant agent was similar at discharge (27.1%) and at 1 year (28.5%).
Predictors of the 1-year outcomes
On multivariable stepwise Cox regression (Tables 2, 3, 4, and 5⇓⇓⇓⇓), the independent predictors of the 1-year primary VARC-2 efficacy endpoint were prior PCI or coronary bypass surgery (hazard ratio [HR]: 1.72; 95% confidence interval [CI]: 1.17 to 2.52; p = 0.006) and EuroSCORE I (per unit increase, HR: 1.02; 95% CI: 1.00 to 1.04; p = 0.027), whereas a borderline association was observed with DAPT use (HR: 0.70; 95% CI: 0.49 to 1.01; p = 0.059). Independent predictors of 1-year death or stroke were EuroSCORE I (per unit increase, HR: 1.02; 95% CI: 1.00 to 1.04; p = 0.013), baseline AF (HR: 1.58; 95% CI: 1.07 to 2.33; p = 0.022), and prior PCI (HR: 1.50; 95% CI: 1.03 to 2.19; p = 0.035).
Effect of pregnancy on 1-year outcomes
The 1-year clinical outcomes in patients with and without history of pregnancy are shown in Table 6. Women with versus without a history of pregnancy had tendency for lower unadjusted 1-year risk of death or stroke (Figure 6) (12.8% vs. 17.3%; unadjusted HR: 0.71; 95% CI: 0.50 to 1.00; p = 0.05), but adjusted risk was attenuated and no longer significant between the 2 groups (HR: 0.73; 95% CI: 0.50 to 1.06; p = 0.097). No meaningfully significant associations were observed between female sex–specific factors and 1-year TAVR outcomes (Online Table 2).
The WIN-TAVI registry is the first study to examine the safety and performance of contemporary TAVR in women and to explore the possible influence of female sex–specific characteristics. This unique collaboration was undertaken without external funding support and was made possible by the partnership of participating research teams. The main findings of this principal report are as follows: 1) the VARC-2 efficacy endpoint occurred in 16.5% women at 1 year and in 10.9% women beyond 30 days; 2) the composite of 1-year death or stroke was 13.9%, and all-cause mortality occurred in 12.5% and stroke in 2.2%; 3) EuroSCORE I and prior PCI or coronary artery bypass grafting were independent predictors of the 1-year VARC-2 composite efficacy endpoint whereas a borderline association was observed with DAPT at discharge. The independent predictors of 1-year death or stroke were EuroSCORE I, baseline AF, and prior PCI; and 4) remote history of pregnancy demonstrated a univariate but not multivariate protective association with 1-year death or stroke, and was not associated with the VARC-2 composite efficacy endpoint. No other female sex–specific characteristic showed any association with 1-year composite clinical TAVR outcomes.
Despite greater procedural complications, women demonstrate more favorable 1-year survival following TAVR compared with men, as a function of different baseline risks; lower prevalence of CAD, diabetes, and AF; and higher baseline LVEF (5–8). There is also some evidence that women have faster regression of LV mass after SAVR (17) and significant improvement of LVEF after TAVR compared with men (18), which may account for superior survival. Moreover, women are also observed to have better survival after TAVR versus SAVR, whereas men perform equally well with both operative strategies (3,8,19). Given this background, the WIN-TAVI registry is the first TAVR study to specifically examine women undergoing TAVR, where we have previously demonstrated that the independent predictors of the 30-day VARC-2 safety endpoint in women included age, LVEF <30%, device generation, device size >26 mm, and history of pregnancy (15).
In the current report we show that 1-year incidence of the VARC-2 efficacy endpoint was almost one-half the previously reported rates in prior TAVR studies with primarily older-generation devices. Notably, however, few studies routinely present the VARC-2 endpoint despite the recommendation for such standardization in reporting (16). Although the PRAGMATIC (Pooled Rotterdam-Milan-Toulouse in Collaboration) registry showed 1-year efficacy rates of 25.0% to 32.0% depending on valve type (20), a report by Genereux et al. (21) showed that only 2 of 16 TAVR studies (12.5%) reported the VARC-2 efficacy endpoint, which ranged from 70.2% to 72.0%. Our rate of 1-year death (12.5%) correlated well with other recent studies (22,23) and was in keeping with the results from PARTNER 2 trial (12.3%) in an intermediate-risk population (3). Comparatively, our stroke rate was almost one-half (2.2%) that reported in PARTNER 2 (5.0%) and SURTAVI (Surgical Replacement and Transcatheter Aortic Valve Implantation) (5.4%) trials (3,4). The lower stroke rate in our study might be due to the use of new-generation devices, low incidence of new AF (3.6%), and expertise of the operators, but is largely attributable to baseline patient characteristics. Nevertheless, women have been shown to be at higher risk of stroke than men (7,24), potentially as a consequence of smaller aortic valve area and higher baseline gradients (25). At the same time, women are considered more likely to be treated with smaller balloon expandable valves, which have been linked with greater cerebral embolization due to plaque fracture from valvuloplasty (26,27).
With respect to prosthetic valve dysfunction, there were only 3 (0.3%) clinical cases of valve thrombosis and 2 (0.2%) cases of endocarditis reported in the registry. In addition, moderate AI occurred in 9.7% at 1 year, with no patient experiencing severe AI. Despite the very high rate of clinical follow-up (98.8%), we cannot exclude underreporting of prosthetic valve dysfunction in our registry secondary to lack of systematic echocardiographic follow-up. Most elderly TAVR patients tend to follow-up with local general practitioners or referring centers, rather than returning to the tertiary TAVR hospital. Notwithstanding, systematic echocardiographic follow-up is important for documentation of prosthesis function, to inform recommendations for best practice, especially in light of recent reports on leaflet thickening (28). In our subset of patients with echocardiography available at 6 months or 1 year, only 4.4% patients had mean gradient >20 mm Hg and 6.0% patients had area ≤0.9 cm2. Certainly, the low incidence of clinical valve related dysfunction in this study supports durability of contemporary TAVR valves, noted from the PARTNER trial, with preserved valve hemodynamics up to 5 years (29).
With respect to the low rates of heart failure rehospitalization, although we cannot entirely exclude under reporting from sites, only 3.5% women had LVEF <30% at baseline, which might have influenced readmissions with heart failure. Conversely, in a recent study by Kolte et al. (30), the rate of baseline heart failure was very high (72.6%). Nevertheless, among patients experiencing 30-day hospitalization for any reason in that study, 22.5% were rehospitalized for heart failure, which amounted to 4.0% of the total study population.
Recent studies have highlighted the predicament in the optimal prescription of antithrombotic therapy post-TAVR, with occurrence of both embolic episodes and silent cerebral microbleeds (26,31,32). Further, valve leaflet thickening occurring in 13.6% of TAVR patients (28) is amenable to oral anticoagulant therapy, but appears to recur after cessation. One other multicenter registry also observed that valve hemodynamic deterioration was associated with lack of anticoagulation (33). Although 55.9% of our study population was discharged on DAPT, only 11.5% remained on DAPT at 1 year, whereas under one-third of patients were on anticoagulant agents both at discharge and at 1 year. Although this may account for the lower incidence of major bleeding beyond 30 days, conversely we only observed 3 clinical cases of valve thrombosis in the study.
Predictors of 1-year outcomes
EuroSCORE I is a well-known predictor of procedural mortality with cardiac surgery (34), but its association with outcomes in women undergoing TAVR has not been previously shown. On the other hand, in a meta-analysis O’Connor et al. (5) observed several individual variables that constitute the EuroSCORE to be correlated to greater long-term death. Similarly, coronary revascularization is a well-established risk factor for future cardiovascular outcomes, but interestingly only forms a part of risk calculators for cardiac surgery (34,35), not TAVR (36). Nevertheless, prior PCI serves as a surrogate for CAD and may also be associated with lower LVEF, an important contributor of poorer survival in men compared with women undergoing TAVR. Our findings also support prior association between AF and worse TAVR outcomes (25,37), emphasizing the need to appropriately identify and risk stratify women with AF undergoing TAVR for treatment with oral anticoagulation. Of note, although DAPT use tended to be protective for the efficacy endpoint, it was not protective for 1-year death or stroke.
Interestingly, new generation TAVR devices and access site were only significant predictors of the 30-day safety endpoint but not 1-year outcomes (15). Although transfemoral access is linked with greater procedural vascular complications (38), it is correlated with better longer-term outcomes than nontransfemoral access, especially in women (3,5,8), highlighting the importance of appropriate access site selection. In addition, we must acknowledge the advantages of new-generation devices with smaller sheath sizes for lower vascular event rates and superior valve performance with reduction in paravalvular AI or pacemaker requirement (39). With respect to female sex–specific characteristics, although history of pregnancy was inversely associated with the 30-day safety endpoint, we observed only a weak univariate association with 1-year death or stroke. Given the baseline differences in patients with and without history of pregnancy, pregnancy may well be a surrogate for other unmeasured confounders (15). Although we found no differences in residence location (home vs. nursing home) at 1 year among patients with or without history of pregnancy, systematic review of social support factors is needed to better understand a possible link. In addition, we did not find any associations between other female specific factors collected in our registry and VARC-2 outcomes at 1 year. Nevertheless, female sex–specific variables are important to understand associations with cardiovascular disease progression and outcomes, as well as to avoid disparities in treatment, as has been highlighted by other authors (40).
First, this was an observational study without a randomized control arm (men) to provide definitive conclusions with respect to sex differences. However, the main aim of the study was to provide real-world data in women who comprise 50% or more of patients undergoing TAVR; and to examine the effect of female sex–specific characteristics. Patients in this registry had a comparable prevalence of cardiovascular risk factors to multiple other registries and therefore accurately reflect real world practice. The true incidence of stroke may be underestimated due to lack of systematic neurological evaluation after TAVR. In addition, despite the very high rate of clinical follow-up at 1 year in this registry, lack of mandatory echocardiography and electrocardiography at follow-up and the fact that most of the follow-up visits were done with local general practitioners rather than referring centers might have led to underdiagnosis of prosthesis dysfunction and under-reporting of new-onset AF. Information on female sex–specific characteristics are subject to recall bias.
Intermediate to high-risk women enrolled in this first ever all-women TAVR registry experienced a 1-year VARC-2 composite efficacy endpoint of 16.5%, with a low incidence of 1-year mortality and stroke. EuroSCORE I, baseline AF, and prior PCI were independent predictors of the 1-year death or stroke, and EuroSCORE I and prior PCI or CABG were independent predictors of the VARC-2 efficacy endpoint. After adjustment, no associations were observed between history of pregnancy and 1-year TAVR outcomes.
WHAT IS KNOWN? Women comprise ≥50% patients undergoing TAVR and have superior 1-year survival compared with men. Significant advancements have taken place with respect to experience and skills of TAVR teams, as well as availability of more optimized device iterations.
WHAT IS NEW? The WIN-TAVI registry is the first ever all-women single-arm study to evaluate the 1-year safety and performance of TAVR in women and to further explore the influence of other female sex–specific characteristics that have never hitherto collected in TAVR studies. Intermediate to high-risk women enrolled in this registry experienced a 1-year VARC-2 composite efficacy endpoint of 16.5% and 1-year death or stroke rate of 13.9%.
WHAT IS NEXT? Randomized assessment of TAVR versus SAVR in all-comer women with severe AS is warranted to determine the optimal management strategy.
The authors would like to thank the Society for Cardiovascular Angiography and Interventions for supporting the launch of this study.
Dr. Petronio has served as a consultant for Boston Scientific, Medtronic, and Abbott Vascular. Dr. Mehilli has received lecture fees from Edwards Lifesciences, Abbott Vascular, Biotronik, Lilly/Daiichi-Sankyo, Terumo, and Bristol-Myers Squibb; and institutional research grant support from Abbott Vascular and Edwards Lifesciences. Dr. Lefèvre has served as a proctor for Edwards Lifesciences. Dr. Sardella has received proctor fees for Edwards Lifesciences; and speaker fees from Direct Flow. Dr. Van Mieghem has received research grant support from Boston Scientific, Edwards Lifesciences, Medtronic, St. Jude Medical, Abbott Vascular, and Claret Medical. Dr. Dumonteil has received proctor fees from Edwards Lifesciences, Medtronic, Boston Scientific, and Abbott Vascular. Dr. Mikhail is the director of the Imperial Valve and Cardiovascular Course. Dr. S. Sharma has served on the Speakers Bureau for Boston Scientific, Abbott Vascular, Cardiovascular Systems Inc., and TriReme. Dr. Naber has received speaker fees from Edwards Lifesciences, Direct Flow Medical, Medtronic, and Claret; is a minor shareholder with Claret; and has served as an advisor for Direct Flow Medical. Dr. Mehran has received institutional research grant support from Eli Lilly/Daiichi-Sankyo Inc., AstraZeneca, The Medicines Company, Bristol-Myers Squibb, OrbusNeich, Beth Israel Deaconess, and Bayer; has served as a consultant for Boston Scientific, Cardiovascular Systems Inc., Medscape, and Shanghai BraccoSine Pharmaceutical; has received institutional advisory board funding from Bristol-Myers Squibb; has received institutional funding from Claret Medical; owns equity in Claret Medical and Elixir Medical; has served on the executive committee for Janssen Pharmaceuticals and Osprey Medical; has served on the data safety monitoring board for Watermark Research Partners; and has a spouse who has served as a consultant for Abiomed and the Medicines Company. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- aortic insufficiency
- aortic stenosis
- coronary artery disease
- confidence interval
- dual-antiplatelet therapy
- European System for Cardiac Operative Risk Evaluation
- hazard ratio
- left ventricular ejection fraction
- percutaneous coronary intervention
- surgical aortic valve replacement
- transcatheter aortic valve replacement
- Valve Academic Research Consortium
- Received July 24, 2017.
- Revision received September 26, 2017.
- Accepted September 28, 2017.
- Reardon M.J.,
- Van Mieghem N.M.,
- Popma J.J.,
- et al.
- O'Connor S.A.,
- Morice M.C.,
- Gilard M.,
- et al.
- Chandrasekhar J.,
- Dangas G.,
- Yu J.,
- et al.
- Williams M.,
- Kodali S.K.,
- Hahn R.T.,
- et al.
- Hamdan A.,
- Barbash I.,
- Schwammenthal E.,
- et al.
- Humphries K.H.,
- Toggweiler S.,
- Rodes-Cabau J.,
- et al.
- Park K.,
- Wei J.,
- Minissian M.,
- Bairey Merz C.N.,
- Pepine C.J.
- Chieffo A.,
- Petronio A.S.,
- Mehilli J.,
- et al.
- Petrov G.,
- Regitz-Zagrosek V.,
- Lehmkuhl E.,
- et al.
- Piazza N.,
- Kalesan B.,
- van Mieghem N.,
- et al.
- Chieffo A.,
- Buchanan G.L.,
- Van Mieghem N.M.,
- et al.
- Genereux P.,
- Head S.J.,
- Van Mieghem N.M.,
- et al.
- Levi A.,
- Landes U.,
- Assali A.R.,
- et al.
- Herrmann H.C.,
- Thourani V.H.,
- Kodali S.K.,
- et al.
- Kapadia S.,
- Agarwal S.,
- Miller D.C.,
- et al.
- Van Mieghem N.M.,
- El Faquir N.,
- Rahhab Z.,
- et al.
- Chakravarty T.,
- Sondergaard L.,
- Friedman J.,
- et al.
- Kolte D.,
- Khera S.,
- Sardar M.R.,
- et al.
- Van Belle E.,
- Hengstenberg C.,
- Lefevre T.,
- et al.
- ↵Van Belle E. Silent cerebral microbleeds during TAVR: insight from a prospective study. Presented at: ACC 2017; March 19, 2017; Washington, DC.
- Del Trigo M.,
- Munoz-Garcia A.J.,
- Wijeysundera H.C.,
- et al.
- ↵Online STS Adult Cardiac Surgery Risk Calculator. 2007. Available at: http://riskcalc.sts.org/stswebriskcalc. Accessed April 2017.
- Edwards F.H.,
- Cohen D.J.,
- O’Brien S.M.,
- et al.
- Tarantini G.,
- Mojoli M.,
- Urena M.,
- Vahanian A.
- Chandrasekhar J.,
- Hibbert B.,
- Ruel M.,
- et al.
- Chamandi C.,
- Puri R.,
- Rodriguez-Gabella T.,
- Rodes-Cabau J.