Author + information
- Received June 1, 2017
- Revision received July 25, 2017
- Accepted July 26, 2017
- Published online September 18, 2017.
- Satya Shreenivas, MDa,
- Dean J. Kereiakes, MDa,∗ (, )
- Stephen G. Ellis, MDb,
- Runlin Gao, MDc,
- Takeshi Kimura, MDd,
- Yoshinobu Onuma, MD, PhDe,
- Karine Piard-Ruster, MD, PhDf,
- Yunlong Zhang, MDf,
- Kai Koo, PhDf,
- Minh-Thien Vu, MSf,
- Patrick W. Serruys, MD, PhDg and
- Gregg W. Stone, MDh
- aChrist Hospital, Heart and Vascular Center, Lindner Research Center, Cincinnati, Ohio
- bCleveland Clinic, Cleveland, Ohio
- cFu Wai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences, Beijing, China
- dKyoto University Hospital, Kyoto, Japan
- eThoraxcenter, Erasmus Medical Center, Rotterdam, the Netherlands
- fAbbott Vascular, Santa Clara, California
- gInternational Centre for Cardiovascular Health, Imperial College, London, United Kingdom
- hNew York Presbyterian Hospital, Columbia University Medical Center and the Cardiovascular Research Foundation, New York, New York
- ↵∗Address for correspondence:
Dr. Dean J. Kereiakes, The Christ Hospital Heart and Vascular Center, The Lindner Research Center, 2123 Auburn Avenue, Suite 424, Cincinnati, Ohio 45219.
Objectives Because females are under-represented in coronary trials, this study sought to assess the relative safety and efficacy of Absorb bioresorbable vascular scaffold (BVS) (Abbott Vascular, Santa Clara, California) and the Xience everolimus-eluting stent in females compared with males.
Background The Absorb everolimus-eluting BVS provides drug delivery and mechanical support similar to a metallic drug-eluting stent, followed by resorption and restoration of more normal vascular structure with the potential to improve late clinical outcomes.
Methods The ABSORB II, ABSORB III, ABSORB Japan, and ABSORB China trials were pooled. Baseline clinical, angiography, procedural variables, and 2-year outcomes were analyzed by sex and device.
Results Among 3,384 randomized patients, 932 (27.5%) were female. Females were older, more often had diabetes and hypertension, but had less everolimus-eluting stent, 3-vessel disease, and smoking compared with males (all p≤0.001). The 2-year rates of target lesion failure with BVS versus everolimus-eluting stent in females were 8.9% versus 6.2% (study-level adjusted hazard ratio: 1.47; 95% confidence interval [CI]: 0.88 to 2.46) and 8.9% versus 6.4% in males (HR: 1.40; 95% CI: 1.02 to 1.92; pinteraction = 0.85). There were no significant interactions between sex and device type for any of the components of target lesion failure.
Conclusions The relative treatment effects of BVS and everolimus-eluting stent for the 2-year rates of target lesion failure and other cardiovascular outcomes were consistent in females and males.
Progressive improvements in clinical outcomes and angiographic measures of restenosis have accompanied the evolution in percutaneous coronary intervention (PCI) from balloon angioplasty to bare-metal stents and to drug-eluting stents (1,2). Nevertheless, a persistent 2% to 3% per year annual incidence of device-related events (target lesion failure [TLF], the composite of cardiac death, target vessel myocardial infarction [TVMI], or ischemia-driven target lesion revascularization [ID-TLR]) has been observed beyond 1 year following PCI regardless of stent type (bare-metal stent or first- or second-generation drug-eluting stent) (3,4). The pathogenesis of these events may be the common presence of a metallic frame that distorts and constrains the vessel, thus limiting vasomotion and adaptive vascular remodeling. The permanent metal frame may serve as a nidus for chronic inflammation, neoatherosclerosis, thrombosis, and strut fracture. With the intent of improving very late clinical outcomes, fully bioresorbable vascular scaffolds (BVS) have been developed that completely resorb several years after PCI, restoring more normal vascular structure and function (5). In randomized trials, the Absorb BVS (Abbott Vascular, Santa Clara, California), the most widely used such device, was demonstrated to be statistically noninferior for 1-year TLF when compared with the Xience cobalt chromium everolimus-eluting stent (EES) (Abbott Vascular) in the ABSORB III pivotal trial for U.S. regulatory approval (6). Long-term outcomes for BVS are presently being collected from this and other randomized trials.
Although coronary heart disease remains the leading cause of mortality in females (7), women are underrepresented in trials of coronary stenting, including ABSORB III (29.5% females), so that analysis from any single trial lacks power to provide accurate estimates of device efficacy or safety in this subgroup. Furthermore, females enrolled into clinical trials are often distinguished by significant differences in baseline clinical and angiographic variables than their male counterparts, including older age, more frequent comorbid risk factors, and smaller coronary arteries (8). Meta-analysis of multiple randomized controlled trials accrues power to inform more definitive assessments of device efficacy and safety. In this context, we performed an individual patient data pooled analysis of 4 randomized trials from the ABSORB clinical trial program in which baseline clinical, angiographic, and procedural variables as well as 2-year clinical outcomes were analyzed by sex and device.
For the present study, we pooled the individual patient data from the ABSORB II (9), ABSORB III (6), ABSORB China (10), and ABSORB Japan (11) trials into a common database following sponsor (Abbott Vascular) and principal investigator approval. These 4 randomized trials were chosen based on their availability of individual patient-level data through 2-year follow-up. In each trial patients with coronary artery disease undergoing PCI were randomized to treatment with the Absorb BVS versus the Xience EES. Major characteristics of the 4 trials have previously been described and are summarized in Table 1. The clinical definitions across the trials were similar as were trial inclusion/exclusion criteria. Allowable reference vessel diameter (RVD) was ≥2.5 mm to ≤3.75 mm and lesion length was limited to ≤24 mm in ABSORB Japan, ABSORB China, and ABSORB III. ABSORB II allowed an RVD of ≥2.25 mm to ≤3.8 mm with lesion lengths up to 48 mm and mandated use of quantitative coronary angiography (QCA), whereas the other 3 trials allowed visual angiographic assessment for lesion eligibility. All trials were approved by the institutional review boards of the participating centers and all patients signed informed consent before trial enrollment.
Study objectives and endpoint definitions
The major objectives of the current study were to characterize the demographics of females versus males receiving BVS compared with EES, and to evaluate the sex-specific relative safety and efficacy of BVS compared with EES. The primary endpoint of the present analysis was the 2-year rate of TLF; secondary endpoints included TVMI, ID-TLR, and scaffold/stent thrombosis (ST) (Academic Research Consortium definite or probable definition) (12). For each trial, the pre-specified clinical endpoints were adjudicated by an independent clinical events committee. An independent core laboratory performed QCA as described previously (13). If ST was suspected, the angiographic core laboratory for each trial assessed whether thrombus was present.
Patient-level data were aggregated into a single dataset and analysis was performed in the intention-to-treat population to examine sex-specific differences. Baseline clinical, demographic, and procedural characteristics were reported as mean ± SD for continuous variables and as proportions for categorical variables. Categorical variables were analyzed using the chi-square test or Fisher exact test. Continuous variables were compared with the Student t test. Time-to-first event curves were developed using Cox proportional hazard models in which study indicators were included as fixed effects. Adjusted survival rates were estimated using SAS procedure PHREG version 9.3 (SAS Institute Inc., Cary, North Carolina). The difference in event rates between 2 groups at a given time point was obtained by using the Z-test. Independent predictors of TLF and ST were determined by multivariable Cox proportional hazards regression, with the number of variables for each model chosen according to their historical association with the event in previous studies, with study indicators, sex, and treatment arm forced into each model. Formal interaction testing was performed to examine whether the relative rates of selected endpoints according to randomized treatment varied by sex. All p values were 2-sided, and a p value < 0.05 was considered statistically significant.
A total of 3,389 patients were enrolled in the 4 trials at 301 centers in Asia, Europe, and North America (Table 1). Five patients who withdrew consent and for whom sex information was not available were excluded from the analysis. Of the 3,384 analyzable patients, 932 (27.5%) were females, of which 593 (63.6%) were randomized to BVS. Among 2,452 males, 1,568 (63.9%) were randomized to BVS. Females were older than males, and more likely to be hypertensive and to have diabetes mellitus, particularly insulin-treated. Males were more likely than females to be active tobacco users, have had prior coronary intervention, and have 3 or more diseased major epicardial coronary arteries (Table 2). DAPT adherence (defined as no more than 1 day without DAPT therapy) during 2-year follow-up was 54.0% in males and 51.8% in females (p = 0.25).
Table 3 depicts the baseline and post-procedure angiographic findings according to sex. By QCA, females had a smaller baseline RVD compared with males and were more likely to have calcified lesions, whereas males had lesions that were longer and more often involved a bifurcation. Regarding procedural variables, there were no differences in the frequencies of pre-dilatation or post-dilatation between sexes, although procedure duration was slightly shorter and the requirement for bailout or unplanned stent or scaffold deployment was more frequent among females.
Clinical outcomes by sex and device
Two-year outcomes are shown in Figures 1 and 2 and Table 4. The 2-year rates of TLF with BVS versus EES in females were 8.9% versus 6.2% (study-level adjusted hazard ratio [HR]: 1.47; 95% confidence interval [CI]: 0.88 to 2.46) and 8.9% vs. 6.4% in males (HR: 1.40; 95% CI: 1.02 to 1.92; pinteraction = 0.85). There were no significant interactions between sex and device type for any of the components of TLF. A borderline interaction was present between sex and device for the 2-year rate of ST BVS versus EES, 1.3% versus 1.2% in females (HR: 1.07; 95% CI: 0.32 to 3.55) and 2.0% vs. 0.4% in males (HR: 5.54; 95% CI: 1.69 to 18.14; pinteraction = 0.06).
The 1- and 2-year TLF and ST outcomes stratified by sex and device randomization for patients having at least 1 target lesion with baseline RVD <2.25 mm versus those who had the baseline RVD of all target lesions >2.25 mm are shown in Table 5. Although no statistically significant interactions between sex and device type for TLF were observed, a borderline interaction (pinteraction = 0.046) was present for the occurrence of 2-year ST according to vessel size.
By multivariable analysis in the entire study population, device type but not sex was an independent predictor of TLF after adjusting for baseline differences in comorbidities and angiographic measures (Central Illustration, panel A). The only independent predictors of ST were device type, diabetes, and pre-procedure RVD (Central Illustration, panel B). Multivariable correlates of 2-year TLF and ST by sex are shown in Figure 3. Device type was an independent predictor of TLF in males but not females, although the adjusted HRs were directionally similar (1.39 vs. 1.55, respectively). Device type was also an independent predictor of ST in males but not females, although the adjusted HR was higher in males (5.67 vs. 1.13, respectively).
The principal observations from the present analysis of 4 randomized controlled trials comparing Absorb BVS with the Xience EES in 3,384 patients undergoing PCI are: 1) males and females with coronary artery disease enrolled in these trials had different baseline clinical and angiographic characteristics; 2) nonetheless, consistency of device treatment effects were observed between the sexes for most 2-year outcomes; and 3) by multivariable analysis, sex was not predictive of TLF, the components of TLF, or ST at 2 years.
As in prior studies (14–18), we found that females undergoing PCI had a different risk profile than males. Females were more likely to be older, have diabetes and hypertension, and have smaller target vessels. Males were more likely to be tobacco users, have 3-vessel disease, and to have had prior coronary interventions. These differences in comorbidities contribute to the worse outcomes observed in many studies in females after PCI, although most prior studies were done in the pre-stent era, included small numbers of female patients, and were retrospective (19–21). In the present analysis, sex was not an independent predictor of the 2-year rate of either TLF or ST. Nonetheless, even with 3,384 patients, the present analysis remains underpowered to exclude an effect of sex on outcomes.
In males, the 2-year rates of TLF, TVMI, and ST were higher after treatment with BVS compared with EES, with a trend toward greater ID-TLR. In females, the rates of TLF, TVMI, and ID-TLR were also numerically higher following treatment with BVS compared with EES, although the differences did not reach statistical significance. In both males and females, the increased risk of TLF following BVS was present at 1 year, with continued divergence of the cumulative event curves between years 1 and 2. There were no significant interactions between device treatment and sex for any outcome measure at 2 years for all randomized patients. Acknowledging the fact that subgroup analysis is inherently underpowered (even with 3,384 patients), the most appropriate and conservative interpretation is that there are no significant differences in the relative treatment effects of BVS versus EES on 2-year outcomes in females and males undergoing PCI.
Use of BVS compared with EES was an independent predictor of 2-year ST among males but not in females, with a borderline nonsignificant interaction present (pinteraction = 0.06). Of note, no patient treated with EES had ST after year 1, and only 1 patient treated with BVS (a female) experienced ST between Year 1 and 2. The increase in very late ST with BVS compared with EES was primarily driven by males treated with BVS, all of whom had all target lesion RVDs 2.25 mm or greater. In this regard, like in most prior reports, females had smaller RVDs than males in our study, a finding that has been associated with increased post-PCI complications (22). Thinner struts of newer drug-eluting stent have been associated with improved clinical outcomes in females (14). Absorb BVS has greater strut width and thickness compared with EES, and smaller vessel size in females remains a theoretical concern for BVS. However, in the present study, the interaction between sex and device for 2-year ST was confined to patients who had the RVD of all target lesions >2.25 mm (pinteraction = 0.046). In contrast, although the 2-year ST rates were generally higher in patients with treated target lesions RVD <2.25 mm, there was no interaction between device type and sex in this high-risk subgroup. However, all subgroup analyses are inherently underpowered, and statistical adjustments were not made for multiple comparisons. These findings may thus be caused by chance, especially because they represent subgroups within subgroups (sex stratified by baseline target RVD). Larger studies are required to determine if there is a varying propensity for ST after BVS between men and women, and if so, whether this relationship is affected by target lesion RVD. Nevertheless, event rates were consistently higher for both sexes when Absorb BVS was implanted in small vessels (RVD <2.25 mm), underscoring the importance of staying within the indicated ranges of vessel size (2.5 mm to 3.75 mm) in both men and women.
Although the inclusion and exclusion criteria were similar between the trials, small differences in lesion length and vessel size between studies were present. In addition, the different geographic regions enrolling in each study may yield heterogeneous patient populations and treatments. None of the 4 trials stratified randomization by sex so comparisons of results between males and females should be done with caution (although our findings were adjusted by study). Although this report represents the largest analysis to date of Absorb BVS compared with EES in females, it still lacks power to detect differences in low-frequency events, such as ST. Subgroup analyses including the small vessel cohort (QCA RVD <2.25 mm) are further underpowered to examine clinical event rates. The 4 component trials enrolled only patients with stable coronary artery disease or stabilized acute coronary syndromes, and extrapolation of these findings to patients with unstable syndromes and more complex lesion morphology should be avoided. Finally, longer-term follow-up is required to determine whether differences in outcomes between males and females emerge over time.
In the present individual patient data pooled analysis of 4 randomized trials comparing Absorb BVS with Xience EES, despite the varying cardiovascular risk profile of females, the treatment effect of Absorb BVS was consistent in both females and males with respect to 2-year clinical outcomes.
WHAT IS KNOWN? The Absorb BVS provides drug delivery and mechanical support similar to a metallic drug-eluting stent, followed by resorption and the potential to restore more normal vascular structure and function, with improvement in very late clinical outcomes. Females are underrepresented in coronary stent trials, and the relative safety and efficacy of Absorb BVS in females versus males is not known.
WHAT IS NEW? Despite differing baseline clinical and angiographic characteristics, the relative treatment effects of Absorb BVS and everolimus-eluting stents on 2-year target lesion failure and other cardiovascular outcomes were consistent in males and females.
WHAT IS NEXT? Although this patient-level pooled analysis of 4 prospective trials is the largest to date to compare Absorb BVS and everolimus-eluting stents in females, device randomization was not stratified by sex, and our study is underpowered to detect differences in low-frequency events, such as device thrombosis. Larger prospective analyses in more complex patient cohorts and with longer-term follow-up are warranted.
The ABSORB trials were sponsored by Abbott Vascular, Santa Clara, California. Dr. Kereiakes has served as a consultant to Boston Scientific, Abbott Vascular, Svelte Medical Systems, Inc., Micell Technologies, Inc., and Sino Medical Sciences Technology, Inc.; and received research funding from Boston Scientific and Abbott Vascular. Dr. Ellis has served as a consultant to Boston Scientific, Abbott Vascular, and Medtronic; and has received research funding from Boston Scientific and Abbott Vascular. Dr. Gao has a research grant from Abbott Vascular. Dr. Kimura is an advisory board member for Abbott Vascular; and has received a research grant from Abbott Vascular. Dr. Onuma is an advisory board member for Abbott Vascular. Dr. Stone has served as the global chairman for the Absorb clinical trial program (uncompensated); and as a consultant for Reva. Columbia University, his employer, also receives royalties from Abbott Vascular for sale of the MitraClip. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- bioresorbable vascular scaffold
- confidence interval
- everolimus-eluting stent(s)
- hazard ratio
- ischemia-driven target lesion revascularization
- percutaneous coronary intervention
- quantitative coronary angiography
- reference vessel diameter
- scaffold/stent thrombosis
- target lesion failure
- target vessel myocardial infarction
- Received June 1, 2017.
- Revision received July 25, 2017.
- Accepted July 26, 2017.
- 2017 American College of Cardiology Foundation
- Kirtane A.J.,
- Gupta A.,
- Iyengar S.,
- et al.
- Palmerini T.,
- Benedetto U.,
- Biondi-Zoccai G.,
- et al.
- Joner M.,
- Finn A.V.,
- Farb A.,
- et al.
- Giustino G.,
- Baber U.,
- Aquino M.
- Serruys P.W.,
- Chevalier B.,
- Dudek D.,
- et al.
- Gao R.,
- Yang Y.,
- Han Y.,
- et al.
- Cutlip D.E.,
- Windecker S.,
- Mehran R.,
- et al.
- Lansky A.J.,
- Dangas G.,
- Mehran R.,
- et al.
- Seth A.,
- Serruys P.W.,
- Lansky A.,
- et al.
- Lansky A.J.,
- Costa R.A.,
- Mooney M.,
- et al.
- Solinas E.,
- Nikolsky E.,
- Lansky A.J.,
- et al.
- Morice M.C.
- Bell M.R.,
- Grill D.E.,
- Garratt K.N.,
- et al.