Screening for eligibility data were not available to report. Although the number of patients with available data on clinical follow-up at 30 months is reported in each group, the efficacy end points were analyzed with the last available follow-up information in the intention-to-treat population, which included all patients who underwent randomization. GUSTO indicates Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries.
aParticipants may have experienced more than 1 event.
bSite terminated participation; participant was not recognized to be eligible by site.
cPatients moved, were incarcerated, or were prematurely exited from the study.
eTable 1. Study-specific inclusion and exclusion criteria
eTable 2. Baseline characteristics of all randomized patients
eTable 3. GUSTO and BARC Definitions
eFigure. Enrollment, randomization, and follow-up among randomized patients
eAppendix. DAPT Investigators by Country
Customize your JAMA Network experience by selecting one or more topics from the list below.
Kereiakes DJ, Yeh RW, Massaro JM, et al. Antiplatelet Therapy Duration Following Bare Metal or Drug-Eluting Coronary Stents: The Dual Antiplatelet Therapy Randomized Clinical Trial. JAMA. 2015;313(11):1113–1121. doi:10.1001/jama.2015.1671
Despite antirestenotic efficacy of coronary drug-eluting stents (DES) compared with bare metal stents (BMS), the relative risk of stent thrombosis and adverse cardiovascular events is unclear. Although dual antiplatelet therapy (DAPT) beyond 1 year provides ischemic event protection after DES, ischemic event risk is perceived to be less after BMS, and the appropriate duration of DAPT after BMS is unknown.
To compare (1) rates of stent thrombosis and major adverse cardiac and cerebrovascular events (MACCE; composite of death, myocardial infarction, or stroke) after 30 vs 12 months of thienopyridine in patients treated with BMS taking aspirin and (2) treatment duration effect within the combined cohorts of randomized patients treated with DES or BMS as prespecified secondary analyses.
Design, Setting, and Participants
International, multicenter, randomized, double-blinded, placebo-controlled trial comparing extended (30-months) thienopyridine vs placebo in patients taking aspirin who completed 12 months of DAPT without bleeding or ischemic events after receiving stents. The study was initiated in August 2009 with the last follow-up visit in May 2014.
Continued thienopyridine or placebo at months 12 through 30 after stent placement, in 11 648 randomized patients treated with aspirin, of whom 1687 received BMS and 9961 DES.
Main Outcomes and Measures
Stent thrombosis, MACCE, and moderate or severe bleeding.
Among 1687 patients treated with BMS who were randomized to continued thienopyridine vs placebo, rates of stent thrombosis were 0.5% vs 1.11% (n = 4 vs 9; hazard ratio [HR], 0.49; 95% CI, 0.15-1.64; P = .24), rates of MACCE were 4.04% vs 4.69% (n = 33 vs 38; HR, 0.92; 95% CI, 0.57-1.47; P = .72), and rates of moderate/severe bleeding were 2.03% vs 0.90% (n = 16 vs 7; P = .07), respectively. Among all 11 648 randomized patients (both BMS and DES), stent thrombosis rates were 0.41% vs 1.32% (n = 23 vs 74; HR, 0.31; 95% CI, 0.19-0.50; P < .001), rates of MACCE were 4.29% vs 5.74% (n = 244 vs 323; HR, 0.73; 95% CI, 0.62-0.87; P < .001), and rates of moderate/severe bleeding were 2.45% vs 1.47% (n = 135 vs 80; P < .001).
Conclusions and Relevance
Among patients undergoing coronary stent placement with BMS and who tolerated 12 months of thienopyridine, continuing thienopyridine for an additional 18 months compared with placebo did not result in statistically significant differences in rates of stent thrombosis, MACCE, or moderate or severe bleeding. However, the BMS subset may have been underpowered to identify such differences, and further trials are suggested.
clinicaltrials.gov Identifier: NCT00977938
Quiz Ref IDCurrent clinical practice guidelines recommend a minimum of only 1 month of dual antiplatelet therapy (DAPT) after bare metal stent (BMS) placement following elective percutaneous coronary intervention (PCI), compared with 6 to 12 months for drug-eluting stents (DES),1,2 and patients with acute coronary syndromes benefit from 12 months of therapy whether or not PCI with stent placement is performed.3 Although randomized trial results4 showed a reduction in stent thrombosis and non–stent-related myocardial infarction (MI) with thienopyridine therapy beyond 12 months after DES placement (among patients tolerating DAPT to 12 months), few trials have assessed optimal duration of DAPT after BMS.5 Because BMS remain a commonly used alternative treatment strategy to DES, particularly for patients who present with acute coronary syndromes or in whom DAPT has perceived increased bleeding risk,6,7 we compared (1) rates of stent thrombosis or major adverse cardiovascular and cerebrovascular events (MACCE) among randomized patients treated with BMS and (2) treatment duration effect among all randomized patients in the Dual Antiplatelet Therapy (DAPT) Study.
We compared the randomized treatment effect of continuing to receive thienopyridine vs receiving placebo beyond 12 months with regard to stent thrombosis, MACCE, and bleeding after randomization until the completion of study drug treatment at 30 months among patients treated with BMS as well as the combined cohort of patients treated with BMS or DES. As a prespecified analysis, we assessed the consistency of treatment duration effect between patients treated with BMS or DES.
Quiz Ref IDThe DAPT Study design has previously been described.8 This double-blind, international, randomized clinical trial compared the risks and benefits of continued thienopyridine (clopidogrel or prasugrel) vs placebo, when given in addition to aspirin for the prevention of stent thrombosis or MACCE after coronary stent placement with either DES or BMS in patients who tolerated DAPT to 12 months. The results comparing randomized treatments in the DES-treated cohort have been reported separately.4
All institutions received approval from their institutional review boards, and each patient provided written informed consent for study participation.
In brief, patients who were candidates for DAPT and who received treatment with either DES or BMS were recruited. Stent treatment was performed according to site standards of care using only US Food and Drug Administration–approved DES and BMS devices. Types of DES included Cypher sirolimus-eluting stent (Cordis), Endeavor zotarolimus-eluting stent (Medtronic), TAXUS paclitaxel-eluting stent (Boston Scientific), and Xience/Promus everolimus-eluting stents (Abbott Vascular or Boston Scientific). All patients older than 18 years who met all enrollment inclusion and none of the exclusion criteria (eTable 1 in the Supplement) and signed the consent were enrolled into the trial within 3 days of the index procedure, and all received open-label aspirin plus thienopyridine for the first 12 months. As permitted by regulatory authorities, race and ethnicity data were collected via patient self-report. Race categories for this study were prespecified as American Indian or Alaska Native, Asian, black or African American, Native Hawaiian or other Pacific Islander, white, and other. Ethnicity was collected as Hispanic or Latino and not Hispanic or Latino.
At 12 months, patients who were alive and free from MI, stroke, repeat coronary revascularization, stent thrombosis, and moderate or severe bleeding and who demonstrated adherence with thienopyridine treatment were then eligible for randomization (Figure) to continue receiving thienopyridine or to receive placebo, and all continued aspirin. A computer-generated randomization schedule stratified patients according to the type of stent they had received (DES vs BMS), hospital site, thienopyridine type, and presence or absence of at least 1 prespecified clinical- or lesion-related risk factor for stent thrombosis (eTable 2 in the Supplement). Postrandomization study procedures and follow-up were the same for all patients regardless of whether they had BMS or DES.
The co-primary effectiveness end points were cumulative incidence of definite or probable stent thrombosis according to the Academic Research Consortium classification9 and incidence of MACCE at 12 to 30 months. For randomized comparison of DAPT duration among patients treated with BMS, the primary safety end point was moderate or severe bleeding (Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Arteries [GUSTO] classification)10 at 12 to 30 months. Finally, clinically actionable bleeding not related to coronary artery bypass graft procedures was also evaluated according to the Bleeding Academic Research Consortium definitions (BARC type 2, 3, or 5).11 These events were adjudicated by an independent clinical events committee blinded to treatment assignment and administered by the Harvard Clinical Research Institute (HCRI). An unblinded independent central data monitoring committee oversaw the safety of all patients.
Among patients treated with BMS and randomized to continued thienopyridine vs placebo, the cumulative incidence of stent thrombosis and of MACCE are presented according to intention-to-treat. Treatments were compared using a log-rank test stratified by geographic region (North America, Europe, and Australia/New Zealand), thienopyridine type, and presence or absence of stent thrombosis risk factors (Table 1).8 For each end point, the stratified hazard ratio (HR) and its 2-sided 95% CI comparing continued thienopyridine vs placebo are presented. Patients not experiencing the co-primary end points at 12 to 30 months after the index procedure were censored at the time of last known contact or 30 months, whichever was earlier.
The analysis of the BMS cohort comparing randomized treatment groups was a prespecified secondary analysis of the DAPT Study that was not powered to compare treatment groups within this cohort (the powered DES-treated cohort has been previously presented4) but was performed to assess consistency of the randomized treatment effect among patients treated with BMS vs DES from the DAPT Study. As a prespecified analysis, stent type × randomized treatment interaction was assessed using Cox proportional hazards regression for ischemic events and logistic regression for bleeding events; stratified HR for ischemic events and nonstratified risk difference for bleeding events, their 95% CI, and P values for interaction are presented. All other analyses presented were prespecified.
All statistical analyses were conducted at HCRI with SAS version 9.2 (SAS Institute). All P values are 2-sided and considered significant at the .05 level.
Enrollment in the DAPT Study was conducted between August 2009 and July 2011, with the last follow-up visit conducted in May 2014. Of 2816 enrolled patients treated with BMS, 583 (20.7%) were not eligible for randomization (mainly due to clinical events requiring continuation of DAPT, such as MI or repeat revascularization procedures) after 12 months of follow-up, 546 (19.4%) were eligible but not randomized, and 1687 (59.9%) were randomized (Figure). Of 25 682 total enrolled patients, 5844 (22.8%) were not eligible for randomization after 12 months of follow-up, 8190 (31.9%) were eligible but not randomized, and 11 648 (45.4%) were randomized, with median follow-up of 990 days (interquartile range, 981-990) (eFigure in the Supplement). The most common reason eligible patients were not randomized was withdrawal of consent.
Baseline characteristics of randomized patients treated with BMS were similar between the groups (Table 1). While the same inclusion and exclusion criteria were applied to all enrolled patients, DES- and BMS-treated patients differed according to clinical and procedural characteristics (eTable 2 in the Supplement). Patients treated with DES were more likely to have a history of diabetes mellitus (30.6% vs 21.2%, P < .001), hypertension, and previous PCI and to have longer lesions, with smaller reference vessel diameter, while patients treated with BMS were more likely to present with ST-elevation MI (STEMI, 37.6% vs 10.5%, P < .001) or non-STEMI (20.9% vs 15.5%, P < .001) and were more likely to have thrombus noted in the treated lesion. The baseline characteristics of the randomized patients treated with DES have been previously published.4 Baseline characteristics of all randomized patients were similar between the randomly assigned treatment groups (eTable 2 in the Supplement). Predefined risk factors for stent thrombosis were present in 54% of patients in each randomly assigned treatment group.
Among randomized patients treated with BMS, the cumulative incidence of stent thrombosis and MACCE were 0.5% vs 1.1% (HR, 0.49; 95% CI, 0.15-1.64; log-rank P = .24) and 4.0% vs 4.7% (HR, 0.92; 95% CI, 0.57-1.47; log-rank P = .72), respectively, for continued thienopyridine vs placebo at 12 to 30 months after the index procedure (Table 2). Moderate or severe GUSTO bleeding events occurred in 2.03% vs 0.90% among patients treated with BMS randomized to continued thienopyridine vs placebo (P = .07); BARC type 2, 3, or 5 bleeding events occurred in 4.56% vs 1.80%, respectively (P = .002). Severe bleeding was uncommon, fatal bleeding events (BARC type 5) were rare, and rates were not different between treatment groups (Table 2).
The results comparing continued thienopyridine vs placebo in the cohort treated with DES have been reported previously and demonstrated significant reductions in study co-primary end points of stent thrombosis (0.4% vs 1.4%, respectively; HR, 0.29; 95% CI, 0.17-0.48) and MACCE (4.3% vs 5.9%, respectively; HR, 0.71; 95% CI, 0.59-0.85) (driven by a reduction in both stent-related and non–stent-related MI) (Table 3). An increase in moderate/severe bleeding events was observed (2.5% vs 1.6%, respectively; P = .001), and a difference in all-cause mortality rate that was not statistically significant was seen (2.0% vs 1.5%; P = .052), yet mortality was infrequently related to bleeding (0.15% vs 0.09% with fatal bleeding, P = .38, and 0.22% vs 0.06% with bleeding-related mortality within the full 33-month follow-up, P = .057).4
The prespecified analysis of the effect of continued thienopyridine found nonsignificant interactions between randomized BMS- and DES-treated patients for both stent thrombosis (HR, 0.49 vs 0.29; interaction P = .42) and MACCE (HR, 0.92 vs 0.71; interaction P = .32) (Table 3).
Among all randomized patients, the co-primary effectiveness end points of stent thrombosis (0.41% vs 1.32%; HR, 0.31; 95% CI, 0.19 to 0.50; P < .001) and MACCE (4.29% vs 5.74%; HR, 0.73; 95% CI, 0.62 to 0.87; P < .001) were reduced by continued thienopyridine vs placebo, respectively (Table 4). The reduction in stent thrombosis was largely explained by a reduction in definite stent thrombosis, and the reduction in MACCE was largely explained by a 48% relative reduction (1.83% absolute) in MI. Significant reductions were observed in MI related to stent thrombosis (0.38% vs 1.28%, HR, 0.29; 95% CI, 0.18 to 0.48; P < .001) as well as MI not related to stent thrombosis (1.84% vs 2.75%, HR, 0.65; 95% CI, 0.50 to 0.84; P < .001). In contrast, there was an increased incidence of severe/moderate bleeding events (2.45% vs 1.47%, risk difference, 0.98; 95% CI, 0.46 to 1.50; P < .001) largely explained by the relative increase in moderate bleeding (1.65% vs 0.96%, risk difference, 0.70; 95% CI, 0.27 to 1.12; P = .001). Similarly, BARC type 2, 3, or 5 bleeding events were significantly increased in the continued thienopyridine treatment group (5.44% vs 2.78%; HR, 2.65; 95% CI, 1.91 to 3.40; P < .001), yet fatal bleeding events (BARC type 5) were rare (0.13% vs 0.09%; P = .58) (Table 4).
Quiz Ref IDAmong patients undergoing coronary stent placement with BMS and who tolerated 12 months of thienopyridine, continuing thienopyridine for an additional 18 months compared with placebo did not result in statistically significant differences in rates of stent thrombosis, MACCE, or moderate or severe bleeding.Quiz Ref ID However, limitations in sample size and power make definitive conclusions regarding DAPT treatment duration effects within BMS difficult. While fewer patients treated with BMS were enrolled and randomized because of the prevailing use of DES in clinical practice, among patients eligible for continued DAPT, a prespecified analysis found nonsignificant interactions for the effect of continued thienopyridine therapy on stent thrombosis among BMS- and DES-treated patients who were randomized in the DAPT Study.4 As this comparison of treatment interaction was not adequately powered for definitive interpretations, true differences in treatment effect size may not have been detected, and any interpretation that continued thienopyridine therapy beyond 1 year (among patients who tolerated DAPT for 1 year without major bleeding) may prevent ischemic events independent of stent type (DES or BMS) should be considered hypothesis-generating.
Indeed, over late-term follow-up (≥5 years), patients treated with BMS accrue cardiac events related to the target lesion at a rate of 2% or more per year13 and beyond the target lesion at a rate of 5% or more per year.14,15 Late atherothrombotic events after BMS may be due to lack of healing or uncovered stent struts, neoatherosclerosis,16 restenosis,17 or disease progression outside the stent, in other regions or vessels. The largest portion of MI prevented by extended-duration thienopyridine therapy in this study did not involve the stented coronary segments for either DES or BMS. While bleeding events were similarly increased with continued thienopyridine therapy beyond 1 year among both BMS- and DES-treated patients, these events were infrequently severe and rarely fatal (BARC type 5).11 The numeric increase in mortality associated with continued thienopyridine therapy (2.0% vs 1.5%, P = .052) that was observed in the cohort treated with DES was not evident among randomized patients treated with BMS (1.0% vs 1.2%, P = .83).
Quiz Ref IDThe lack of apparent treatment interaction between DES and BMS supports the combined analysis of treatment effects of continued duration of therapy independent of stent type. Among the combined BMS and DES cohort, the reductions in stent thrombosis and MACCE were 69% and 27%, respectively, in patients continuing thienopyridine therapy together with aspirin. Fifty percent of the MIs prevented by continued DAPT were not related to stent thrombosis. These ischemic event benefits were balanced by a 67% relative increase in moderate or severe bleeding.
The major limitation of the BMS randomized comparison of DAPT duration is small sample size and lack of power, which limits the interpretability of the findings. However, an adequately powered randomized BMS cohort would require approximately 8000 additional patients, which was practically not feasible. An adequate number of patients treated with BMS were enrolled to allow a powered comparison of stent thrombosis and MACCE rates with patients treated with DES,8 the results of which have been presented separately.18 In this context, the design of the BMS randomized comparison was to evaluate for consistency or heterogeneity compared with the DES treatment effect in an exploratory fashion, rather than to be powered for a separate, independent analysis. Nonetheless, the BMS cohort sample size exceeds that of prior randomized BMS cohorts evaluating duration of antiplatelet therapy5 and is similar in size to many prior randomized trials of DAPT duration in DES.5,19-22 Although similar inclusion criteria were required of BMS- and DES-treated patients, there were systematic differences between BMS- and DES-treated patients, with a higher frequency of MI presentation before the index PCI procedure for patients treated with BMS and a higher prevalence of restenosis risk factors for patients treated with DES. Nevertheless, each cohort was balanced across randomized treatment groups as expected according to the stratified randomization.
Among patients undergoing coronary stent placement with BMS who tolerated 12 months of thienopyridine and aspirin therapy without major bleeding, continuing thienopyridine therapy in addition to aspirin beyond 12 months did not result in statistically significant differences in rates of stent thrombosis, MACCE, or moderate or severe bleeding. However, the BMS subset may have been underpowered to determine such differences.
Corresponding Author: Laura Mauri, MD, MSc, Division of Cardiovascular Medicine, Department of Medicine, Brigham and Women’s Hospital, 75 Francis St, Boston, MA 02115 (email@example.com).
Author Contributions: Drs Kereiakes, Massaro, and Mauri had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Kereiakes, Yeh, Massaro, Cutlip, Mauri.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Kereiakes, Yeh, Mauri.
Critical revision of the manuscript for important intellectual content: Kereiakes, Yeh, Massaro, Driscoll-Shempp, Cutlip, Steg, Gershlick, Darius, Meredith, Ormiston, Tanguay, Windecker, Garratt, Kandzari, Lee, Simon, Iancu, Trebacz, Mauri.
Statistical analysis: Massaro, Mauri.
Obtained funding: Mauri.
Administrative, technical, or material support: Driscoll-Shempp, Cutlip, Mauri.
Study supervision: Driscoll-Shempp, Mauri.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Yeh reported having received personal fees from Abbott Vascular, personal fees and nonfinancial support from Harvard Clinical Research Institute, and personal fees from Gilead Sciences. Dr Massaro reported having received personal fees from Harvard Clinical Research Institute. Dr Cutlip reported having received support from Medtronic, Boston Scientific, Cordis, and Abbott Vascular and grants from the National Heart, Lung, and Blood Institute. Dr Steg reported having received personal fees from Amarin, Bayer, Boehringer-Ingelheim, Bristol-Myers Squibb, Daiichi-Sankyo, GlaxoSmithKline, Lilly, Merck-Sharpe-Dohme, Novartis, Otsuka, Pfizer, Roche, Medtronic, Vivus, The Medicines Company, and Orexigen and grants and personal fees from sanofi-aventis and Servier. Dr Gershlick reported having received personal fees from Medtronic and Abbott and grants from The Medicines Company. Dr Darius reported having received grants from Harvard Clinical Research Institute. Dr Meredith reported having received support from Boston Scientific and Medtronic. Dr Tanguay reported having received personal fees and/or other support from Abbott Vascular, AstraZeneca, Bayer, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, Roche, sanofi-aventis, Servier, Ikaria, and Merck. Dr Windecker reported having received grants from St Jude Medical, Biotronik, The Medicines Company, Abbott, Medtronic, and Edwards Lifesciences and personal fees from AstraZeneca, Eli Lilly, Abbott, Biosensors, Biotronik, and Bayer. Dr Garratt reported having received grants from Boston Scientific, Abbott Vascular, and CeloNova and personal fees or other support from Boston Scientific, The Medicines Company, Abbott Vascular, Infarct Reduction Technologies, Guided Delivery Systems, and Daiichi-Sankyo/Lilly. Dr Kandzari reported having received grants from Medtronic CardioVascular, Abbott Vascular, Boston Scientific, and Biotronik and personal fees from Micell Technologies, Medtronic CardioVascular, and Boston Scientific. Dr Lee reported having received grants from Boston Scientific and personal fees from Boston Scientific and Medtronic. Dr Simon reported having received other support from Cordis/Johnson & Johnson and personal fees from Cordis/Johnson & Johnson and Medtronic Vascular. Dr Mauri reported having received grants from Abbott, Boston Scientific, Medtronic, Cordis, Eli Lilly, Daiichi Sankyo, Bristol-Myers Squibb, sanofi-aventis, and Biotronik and personal fees from Medtronic, St Jude, and Biotronik. No other disclosures were reported.
Funding/Support: The study was sponsored by the Harvard Clinical Research Institute (HCRI) and funded by Abbott, Boston Scientific, Cordis, Medtronic, Bristol-Myers Squibb/Sanofi Pharmaceuticals Partnership, Eli Lilly, and Daiichi Sankyo and the US Department of Health and Human Services (1RO1FD003870-01).
Role of the Funder/Sponsor: The funding manufacturers and US Food and Drug Administration had input on the study design and conduct of the study. The HCRI oversaw the collection, management, and analysis of the data; the study authors were responsible for interpretation of the data, preparation of the manuscript, and the decision to submit the manuscript for publication.
Previous Presentation: Results were presented in an abstract and in a presentation at the American Heart Association Scientific Sessions; Chicago, Illinois; November 18, 2014.
Additional Contributions: We thank the other investigators, the staff, and the participants of the DAPT Study for their valuable contributions. We wish to acknowledge Joanna Suomi, MSc, for assistance editing and formatting the manuscript and Wen-Hua Hsieh, PhD, for assistance with statistical analysis. Both are employed by Harvard Clinical Research Institute and were compensated for their contributions.
Correction: This article was corrected online April 28, 2015, for errors in describing the prespecified analyses and July 5, 2016, to add an institution to an author’s affiliation.