Author + information
- Received December 16, 2015
- Revision received February 9, 2016
- Accepted February 24, 2016
- Published online May 10, 2016.
- Pascal Vranckx, MD, PhDa,
- Harvey D. White, DScb,
- Zhen Huang, MSc,
- Kenneth W. Mahaffey, MDd,
- Paul W. Armstrong, MDe,
- Frans Van de Werf, MDf,
- David J. Moliterno, MDg,
- Lars Wallentin, MD, PhDh,
- Claes Held, MD, PhDh,
- Philip E. Aylward, MDi,
- Jan H. Cornel, MD, PhDj,
- Christoph Bode, MDk,
- Kurt Huber, MDl,
- José C. Nicolau, MD, PhDm,
- Witold Ruzyllo, MDn,
- Robert A. Harrington, MDd and
- Pierluigi Tricoci, MD, MHS, PhDc,∗ ()
- aHartcentrum Hasselt, Hasselt, Belgium
- bGreen Lane Cardiovascular Service, Auckland City Hospital, Auckland, New Zealand
- cDuke Clinical Research Institute, Durham, North Carolina
- dStanford University, Stanford, California
- eDivision of Cardiology, University of Alberta, Edmonton, Canada
- fDepartment of Cardiology, University of Leuven, Leuven, Belgium
- gGill Heart Institute and Division of Cardiovascular Medicine, University of Kentucky, Lexington, Kentucky
- hDepartment of Medical Sciences, Cardiology, Uppsala Clinical Research Center, Uppsala University, Uppsala, Sweden
- iSAHMRI, Flinders University and Medical Centre, Adelaide, Australia
- jDepartment of Cardiology, Medisch Centrum Alkmaar, Alkmaar, the Netherlands
- kInternal Medicine and Cardiology, Universitätsklinikum, Freiburg, Germany
- l3rd Department of Medicine, Cardiology and Intensive Care Medicine, Wilhelminen Hospital, Vienna, Austria
- mHeart Institute (InCor), University of São Paulo Medical School, São Paulo, Brazil
- nDepartment of Coronary Artery Disease and Cardiac Catheterization Laboratory, Institute of Cardiology, Warsaw, Poland
- ↵∗Reprint requests and correspondence:
Dr. Pierluigi Tricoci, Duke Clinical Research Institute, Box 3850, 2400 Pratt Street, Durham, North Carolina 27705.
Background The Bleeding Academic Research Consortium (BARC) scale has been proposed to standardize bleeding endpoint definitions and reporting in cardiovascular trials. Validation in large cohorts of patients is needed.
Objectives This study sought to investigate the relationship between BARC-classified bleeding and mortality and compared its prognostic value against 2 validated bleeding scales: TIMI (Thrombolysis In Myocardial Infarction) and GUSTO (Global Use of Strategies to Open Occluded Arteries).
Methods We analyzed bleeding in 12,944 patients with acute coronary syndromes without ST-segment elevation, with or without early invasive strategy. The main outcome measure was all-cause death.
Results During follow-up (median: 502 days), noncoronary artery bypass graft (CABG) bleeding occurred in 1,998 (15.4%) patients according to BARC (grades 2, 3, or 5), 484 (3.7%) patients according to TIMI minor/major, and 514 (4.0%) patients according to GUSTO moderate/severe criteria. CABG-related bleeding (BARC 4) occurred in 155 (1.2%) patients. Patients with BARC (2, 3, or 4) bleeding had a significant increase in risk of death versus patients without bleeding (BARC 0 or 1); the hazard was highest in the 30 days after bleeding (hazard ratio: 7.35; 95% confidence interval: 5.59 to 9.68; p < 0.0001) and remained significant up to 1 year. The hazard of mortality increased progressively with non-CABG BARC grades. BARC 4 bleeds were significantly associated with mortality within 30 days (hazard ratio: 10.05; 95% confidence interval: 5.41 to 18.69; p < 0.0001), but not thereafter. Inclusion of BARC (2, 3, or 4) bleeding in the 1-year mortality model with baseline characteristics improved it to an extent comparable to TIMI minor/major and GUSTO moderate/severe bleeding.
Conclusions In patients with acute coronary syndromes without ST-segment elevation, bleeding assessed with the BARC scale was significantly associated with risk of subsequent death up to 1 year after the event and risk of mortality increased gradually with higher BARC grades. Our results support adoption of the BARC bleeding scale in ACS clinical trials. (Trial to Assess the Effects of Vorapaxar [SCH 530348; MK-5348] in Preventing Heart Attack and Stroke in Participants With Acute Coronary Syndrome [TRACER] [Study P04736]; NCT00527943)
Antithrombotic and revascularization strategies to reduce the risk of recurrent ischemic events in patients with non–ST-segment elevation acute coronary syndromes (NSTE-ACS) are associated with increased risk of bleeding. Therefore, bleeding is a central safety outcome in cardiovascular clinical trials, especially for antithrombotic strategies and invasive procedures (1,2).
Many definitions of bleeding exist and are used inconsistently in clinical trials and registries, making it difficult to compare results across trials. Additionally, the prognostic impact of bleeding may depend upon definitions used. In 2011, the Bleeding Academic Research Consortium (BARC) proposed standardized bleeding definitions in patients receiving antithrombotic therapy by implementing a hierarchical approach (3). In patients with ACS following a percutaneous coronary intervention (PCI), this consensus classification has been shown to be independently associated with 1-year mortality, but it has not been validated in a broad population with ACS, including patients treated with other treatment strategies (4,5). Furthermore, the prognostic value of certain BARC categories has not been established, such as bleeding requiring medical or laboratory evaluation but not meeting specific thresholds of clinical or laboratory severity and coronary artery bypass graft (CABG)-related bleeding.
The TRACER (Thrombin Receptor Antagonist for Clinical Event Reduction in Acute Coronary Syndrome) trial was a contemporary trial testing the efficacy and safety of the protease-activated receptor-1 antagonist vorapaxar against placebo in patients with NSTE-ACS in addition to standard of care, where treatment decisions were made by the treating physician (6,7). The large TRACER dataset provides opportunities to: 1) examine the association of BARC-defined bleeding with mortality over a long follow-up in a broad group of NSTE-ACS patients in whom different types of treatment strategies were adopted (PCI, CABG, medical treatment); and 2) compare the BARC definition with other established bleeding definitions in predicting mortality.
The trial’s primary results and design have been previously reported (6,7). In brief, TRACER was a multicenter, global, randomized, double-blind, event-driven trial with a minimum follow-up of 1 year. The study compared placebo with vorapaxar administered as a 40 mg loading dose followed by a daily 2.5 mg maintenance dose in patients with NSTE-ACS at high risk for recurrent ischemic events (6,7). Concomitant treatment was according to the applicable practice guidelines (e.g., North American, European) (8,9). Subjects who discontinued study treatment for any reason were followed for occurrence of clinical events.
The TRACER trial complied with the Declaration of Helsinki and was approved by the appropriate national and institutional regulatory authorities and ethics committees. All patients provided written informed consent.
Endpoint definitions and follow-up
The main outcome measure for this analysis was all-cause mortality. All suspected bleeding events in the TRACER trial were systematically identified via an integrated assessment of investigator-reported events along with central review of relevant data in the electronic case report form (eCRF). The eCRF contained information on localization, imaging tests, hemoglobin levels, hematocrit levels, and blood transfusions. For patients who underwent CABG surgery during the index hospitalization, the eCRF contained data on the occurrence of bleeding and its characteristics. All suspected bleeding events were prospectively assessed and adjudicated according to the Thrombolysis In Myocardial Infarction (TIMI) classification and the GUSTO (Global Use of Strategies to Open Occluded Coronary Arteries) classification (10,11). Protocol-defined bleeding and efficacy endpoints included in the primary outcome measure were adjudicated by a central clinical events committee. The BARC consensus document was published while TRACER was being conducted. BARC bleeding definition criteria were retrospectively derived using an algorithm on the basis of the adjudicated data points collected during the clinical events committee review. Bleeding definitions are provided in Online Table 1 (3). BARC grade 5 is fatal bleeding and was not included in the analyses of the relationship with all-cause mortality. Subjects with BARC grade 5 were analyzed according to the bleeding events before the fatal bleeding event.
The TRACER protocol required that all enrolled patients return for study visits at 30 days; at 4, 8, and 12 months; and then every 6 months thereafter. Patients who prematurely discontinued treatment were followed by telephone at the same intervals. During follow-up visits, patients underwent a complete clinical evaluation. The TRACER trial was part of an investigational new drug application; therefore, quality of data collection was assessed according to current regulatory standards (12).
The analysis population included all randomized subjects in the TRACER trial. Bleeding events post-randomization to last follow-up were classified as BARC bleeding categories on the basis of the algorithm described in Online Table 2. The first event of each BARC bleeding category for a subject was included in the analyses described later.
Baseline characteristics were reported for subjects with and without BARC 2, 3, and 4 bleeding. All p values were calculated according to multivariate Cox modeling that included randomized treatment and baseline covariates. Cumulative event rates of mortality post-bleeding were estimated by the Kaplan-Meier method. Mortality rates at 30 days, 1 year, and 2 years post-bleeding were estimated among subjects with bleeding events according to the BARC, TIMI, and GUSTO classifications. Event curves were created to describe the incidence and timing to death post-bleeding.
To evaluate the association between various bleeding classifications and mortality, hazard ratios (HRs) of mortality and 95% confidence intervals (CIs) were estimated by Cox regression models, in which bleeding was included as a time-varying covariate in addition to the baseline characteristics (as listed in Table 1) and randomized treatment. The time-varying aspect of bleeding was 2-fold: the effect of bleeding and the status of bleeding for a subject, both of which might change over time. As a post-randomization risk factor, bleeding occurred in different subjects at different times. The bleeding status of all subjects was examined at each time point since randomization, where everyone was first classified as “no bleeding.” A subject remained in this category until a bleeding event had occurred and was then reclassified as “bleeding.” Prior knowledge suggested that the risk for mortality following a bleeding event differed in the short and long term; therefore, a piecewise hazard function was built representing periods within day 30, day 31 to 1 year, and 1 year to 2 years post-bleeding. Wald chi-square tests were carried out to examine whether the HRs were consistent across these periods. In subjects who had more than 1 bleed during follow-up, the subsequent bleeds were included in the analyses and analyzed in the same way as the initial one. Goodness-of-fit tests of -2 log likelihood were used to evaluate the improved model performance comparing models with and without bleeding. These analyses were carried out for bleeding events according to BARC, TIMI, and GUSTO classifications.
All statistical tests were 2-sided with a significance level of 0.05; p values are not adjusted for multiple comparisons. SAS software, version 9.2 (SAS Institute, Cary, North Carolina) was used for statistical analyses.
A total of 12,944 patients were enrolled at 818 sites in 37 countries. Approximately 9 of 10 patients (n = 11,399; 88%) underwent coronary angiography; 7,075 (54%) patients received a coronary stent; and 1,312 (10%) patients were treated with CABG. The cumulative incidence of a BARC bleed of grade 2 or higher from randomization to last follow-up was 15.3% (n = 1,975); 3.7% (n = 484) of patients had a TIMI major or minor bleed and 4.0% (n = 514) of patients had a bleed according to GUSTO severe or moderate criteria. The cumulative incidence of bleeding according to BARC criteria is shown in Online Figure 1.
Baseline demographic characteristics for patients with (BARC 2, 3, or 4) or without (BARC 0 or 1) bleeding are reported in Table 1. Patients with actionable bleeding were older, more frequently from North America, more likely to be current smokers, and have reduced renal function, higher Killip class, and a history of prior stroke. Use of glycoprotein IIb/IIIa inhibitors at baseline was also more common in patients with BARC actionable bleeding.
Bleeding events and mortality
In the 12,944 patients analyzed, there were 652 (5.0%) deaths at 2 years, with 196 (1.5%) deaths occurring within the first 30 days after randomization. The cumulative mortality rates post-bleeding by BARC, TIMI, and GUSTO categories are shown in the Central Illustration. The crude mortality rates were comparable among patients with BARC grade 3, GUSTO moderate or severe, and TIMI minor or major bleeding (Table 2).
The association between BARC bleeding and mortality was significant in the 30 days following bleeding and at 1 year post-bleeding (Table 3). In the interval beyond 1 year from bleeding, the hazard decreased and was no longer significant. The hazard of mortality increased progressively with higher non-CABG BARC grades (p for trend <0.0001 within 30 days and between 30 days and 1 year) (Table 2). BARC grade 4 (bleeding following CABG), as a separate grade, contained a more than 10-fold increase in hazard up to 30 days but not beyond. The inclusion of BARC bleeding in a multivariable model predicting mortality significantly improved the goodness of fit (Table 4).
TIMI major or minor bleeding and GUSTO moderate or severe bleeding were also significantly associated with mortality (Tables 3 and 4). Similar to BARC bleeding, the relationship with mortality was significant up to 1 year after bleeding, but not thereafter. In comparing the 3 scales, a similar degree of increase in the hazard was observed with BARC grades 3 or 4, TIMI major or minor, and GUSTO moderate or severe bleeding.
The timing of a BARC 2, 3, or 4 bleed (during the index stay, including bleeding events related to instrumentation or trauma; within 7 days of discharge; and beyond 7 days after discharge) did not affect the relationship with mortality (Online Figure 2). Finally, transfusion was not independently associated with mortality and did not confer additional prognostic value on top of BARC criteria (Online Table 3).
Comparison of bleeding definitions
In the distribution of bleeding using BARC, TIMI major/minor, and GUSTO moderate/severe criteria (Table 5), only one-half of BARC 3 bleeds were concomitantly captured in both the TIMI minor or major and GUSTO moderate or severe categories. Of the remaining BARC 3 bleeds, approximately 30% were captured as GUSTO moderate or severe only (but not TIMI major or minor) and approximately 20% were captured as TIMI major or minor only (but not GUSTO moderate or severe).
There were 159 bleeds that were not classified by BARC criteria. Those were CABG-related bleeds that were classified as GUSTO moderate or severe but did not meet the BARC CABG definition. In fact, transfusion and hemodynamic instability, key elements of the GUSTO criteria, are not part of the BARC 4 criteria.
To understand the additive predictive value of bleeding grades on mortality, we first created a multivariable model to predict 1-year mortality using baseline characteristics and then added various bleeding classifications to assess to what degree the inclusion of bleeding scales in the mortality model would improve the prediction capability of the model. The inclusion of BARC 2, 3, or 4 and of BARC 3 or 4 criteria to the baseline model significantly increased the predictive value of the model (Table 4), to a similar degree as TIMI minor or major bleeding and GUSTO moderate or severe bleeding.
Bleeds are not equal and do not carry the same hazard for mortality over time; therefore, accurate classification of severity is necessary. The BARC consensus document proposed a hierarchical grading system to classify bleeding events in cardiovascular investigations (3), but validation (especially on its prognostic implication) is required. We studied the impact of bleeding classified with the BARC scale on mortality in a cohort of patients with NSTE-ACS treated according to current clinical practice and enrolled in the TRACER trial. We concluded that bleeds classified as BARC grades 2, 3, or 4 were independently associated with an increased risk of mortality. Blood transfusion for bleeding was not an independent predictor for mortality after full adjustment for all available variables and bleeding. There was a gradual increase in the risk of mortality up to 1 year following a bleeding event with an increasing BARC-graded severity of bleeding.
The strengths of our analysis relied on the extended follow-up, the well-described and contemporary NSTE-ACS patient population, and the prospective collection and blinded adjudication of bleeding events. TRACER patients were largely included with positive biomarkers and then treated with a variety of strategies according to modern standards of care and physicians’ choices. More than 90% of patients in TRACER had positive troponin, 90% underwent diagnostic angiography, and nearly 60% underwent PCI. One of 10 patients had CABG during the index hospitalization.
The TIMI and GUSTO criteria were the prespecified bleeding scales used for the main safety analysis and reporting in the TRACER trial. The TIMI definition integrates mainly laboratory-based data, whereas the GUSTO definition is largely clinically based (10,11). For this analysis, all bleeding events were reclassified according to the BARC hierarchical bleeding scale (grade 2, 3, or 4) using data prospectively collected as part of the clinical event committee adjudication for the TIMI and GUSTO scales (5). The hazard of death associated with BARC grade 3 criteria was similar in magnitude compared with TIMI (major or minor) and GUSTO (moderate or severe) bleeding criteria. However, our analysis demonstrated that BARC classification is potentially able to capture a larger proportion of clinically significant bleeding than either the GUSTO or TIMI scales. We showed that TIMI major or minor criteria only capture about 70% of bleeds meeting BARC 3 criteria, whereas GUSTO moderate or severe criteria capture approximately 80% of BARC 3 bleeding. The similar prognostic significance of the 3 scales, but with a more inclusive set of criteria in the BARC scale, may make BARC a more desirable bleeding scale given the possibility to capture more bleeding.
Not surprisingly, BARC 3C (intracranial) bleeding events were associated with the worst prognosis. Blood transfusion for bleeding was not an independent predictor for mortality after full adjustment for baseline variables, treatment regimen, and BARC-graded bleeding.
The increase in mortality was significant up to 1 year after a BARC actionable bleeding, whereas the relationship between bleeding and mortality became weaker and nonsignificant beyond 1 year from the bleed. The reason for this finding is not entirely clear, but it is possible that patients at a higher risk of mortality have a fatal outcome relatively early after a bleed so those who survive beyond 1 year represent a lower risk group. On the other hand, it may also be that the impact of bleeding on the risk of death is increased only for a certain period after the event (13).
The vast majority of bleeding events occurred early after randomization. However, in the current analysis we could not demonstrate a differential hazard for mortality by BARC-graded bleeding occurring during the index stay, including the majority of instrumented and traumatic bleeds, and after discharge.
Interestingly, in TRACER, even BARC 1 bleeds—recently shown to be associated with decreased short- and long-term quality of life (14)—carried an increased hazard of death up to 30 days.
Finally, this analysis provides a first validation of BARC 4 grade (CABG-related) bleeding in a large cohort of patients with ACS (1,312 patients underwent CABG during the index hospitalization). We observed that BARC 4–defined CABG-related bleeding carried a high 30-day HR, similar to BARC 3b grade, but leveled off thereafter.
Our findings complement the results from TRITON-TIMI 38 (Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition with Prasugrel–Thrombolysis In Myocardial Infarction 38) and PRODIGY (Prolonging Dual Antiplatelet Treatment After Grading Stent-Induced Intimal Hyperplasia) (5,15). In both trials, which were confined to patients presenting with ACS undergoing an early invasive treatment strategy, the impact of bleeding on mortality was restricted to more serious bleeding events. Focusing on the time dependency of this relationship, the associated HR was similar for bleeding events occurring during the index hospitalization—clustering the instrumented bleeds—or postdischarge. This finding points to the detrimental nature of a major bleed per se, regardless of the timing of the index procedure or randomization. The decline in the hazard for mortality late after a first serious non-CABG bleeding event is biologically plausible and may explain the attenuation in the hazard beyond 1 year in our analysis.
In TRACER, we could not detect any additional risk for 30-day or 1-year mortality with blood transfusion for bleeding beyond the BARC serious (grade 3 or 4) events. Use of transfusion is highly variable, and local policy may differ across centers and providers, which may explain this finding. Information on policy and guidance adopted in the use of blood transfusion was not collected.
The most relevant limitation of the present analysis is the different timing of the adjudication of bleeding. More specifically, bleeding according to BARC criteria was derived retrospectively, although it was on the basis of prospectively collected data by a central adjudication committee.
Although TRACER did not have significant exclusion criteria on the basis of the presumed risk of bleeding, patients with active bleeding, hemorrhagic diathesis, or history of intracranial hemorrhage at any time were excluded. Patients with concomitant or foreseeable need for oral anticoagulation were not included in the trial. Even after adjustment, the possibility of residual confounding remains for the relationship between bleeding and mortality.
For the outcome analysis, models were adjusted for baseline variables but not for post-randomization factors, such as recurrent myocardial infarction or reinterventions during follow-up.
In this validation of the BARC hierarchical classification in patients with a recent NSTE-ACS, regardless of the initial treatment strategy, BARC bleeding grades independently and incrementally predicted 1-year mortality. BARC grade 3 had a hazard of death similar to that provided by TIMI and GUSTO scales, but could capture more bleeding than TIMI major or minor or GUSTO moderate or severe criteria only. These data add to the clinical validity of the BARC classification that, by integrating elements of both GUSTO and TIMI scales, may be helpful to standardize bleeding endpoint definitions in clinical investigations and may thus be used in addition to or in substitution of these 2 scales.
COMPETENCY IN PRACTICE-BASED LEARNING: In patients with non–ST-segment elevation acute coronary syndromes, bleeding is a composite endpoint associated with a risk of mortality over the course of the subsequent year. The Bleeding Academic Research Consortium criteria capture most overt bleeding events and therefore form a meaningful tool for assessment of the safety of antithrombotic therapy.
TRANSLATIONAL OUTLOOK: Further studies are needed to develop methods for integrated assessment of both safety and efficacy to reflect the net clinical benefit of antithrombotic interventions in randomized trials and clinical practice.
For supplemental tables and figures, please see the online version of this article.
The TRACER trial was supported by Merck & Co., Inc. Dr. White has received research grants from Sanofi, Eli Lilly, the National Institutes of Health, Merck Sharpe & Dohme, AstraZeneca, GlaxoSmithKline, Daiichi-Sankyo Pharma Development, George Institute, Omthera Pharmaceuticals, Pfizer New Zealand, Intarcia Therapeutics Inc., Elsai Inc., and DalGen Products and Services; and participates in an advisory board at AstraZeneca. Dr. Mahaffey has received research grants from Amgen, Daiichi, Johnson & Johnson, Medtronic, Merck, St. Jude, and Tenax; has provided consulting or other services for the American College Cardiology, AstraZeneca, BAROnova, Bayer, Bio2 Medical, Boehringer Ingelheim, Bristol-Myers Squibb, Cubist, Eli Lilly, Elsevier, Epson, Forest, GlaxoSmithKline, Johnson & Johnson, Medtronic, Merck, Mt. Sinai, Myokardia, Omthera, Portola, Purdue Pharma, Springer Publishing, The Medicines Company, Vindico, and WebMD; and has equity in BioPrint Fitness. Dr. Van de Werf has received a research grant, honoraria for lectures, and advisory board membership from Merck & Co. Dr. Moliterno has served on a Data and Safety Monitoring Board for Janssen Pharmaceuticals; and has received research grants from Merck and AstraZeneca. Dr. Wallentin has received research grants from AstraZeneca, Merck & Co., Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, and GlaxoSmithKline; lecture fees from AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, GlaxoSmithKline, and Merck & Co.; honoraria from Boehringer Ingelheim, AstraZeneca, Bristol-Myers Squibb/Pfizer, GlaxoSmithKline, and Merck & Co.; been a consultant/served on an advisory board from Merck & Co., Regado Biosciences, Evolva, Portola, C.S.L. Behring, Athera Biotechnologies, Boehringer Ingelheim, AstraZeneca, GlaxoSmithKline, and Bristol-Myers Squibb/Pfizer; and received travel support from Bristol-Myers Squibb/Pfizer. Dr. Held has received research grants from AstraZeneca, GlaxoSmithKline, Pfizer/Bristol-Myers Squibb, Roche, and Schering-Plough (now Merck); has served on an advisory board for AstraZeneca; and has served as a consultant for Bayer. Dr. Aylward has received a research grant from Merck & Co., AstraZeneca, Sanofi, and GlaxoSmithKline; and has received honoraria/served on a speaker bureau and advisory board for AstraZeneca, Eli Lilly, Boehringer Ingelheim, Bayer/Johnson & Johnson, Servier, and Bristol-Myers Squibb. Dr. Cornel has received consulting fees/honoraria from AstraZeneca, MSD, Eli Lilly, and Bristol-Myers Squibb/Pfizer; and travel support from Bayer and AstraZeneca. Dr. Bode has received research grants from AstraZeneca, Bayer, Boehringer Ingelheim, Merck, and Sanofi; and speakers' and consulting honoraria from Bayer, Bristol-Myers Squibb, and Daiichi-Sankyo. Dr. Huber has received lecture fees from AstraZeneca, Daiichi-Sankyo, Eli Lilly, Sanofi, and The Medicines Company; and a research grant from AstraZeneca. Dr. Nicolau has received a research grant and consultant/advisory fees from Merck; and research grants and/or personal fees (consultancy, lectures, travel support) from Sanofi, AstraZeneca, Daiichi-Sankyo, Bayer/Johnson & Johnson, and Roche, outside the submitted work. Dr. Harrington has received research grants from Merck & Co., Astra, Sanofi, Bristol-Myers Squibb, The Medicines Company, Portola Pharma, and Regado; has consulted for Merck and The Medicines Company; and has served on advisory boards for Gilead and WebMD. Dr. Tricoci has a consultant agreement and has received a research grant from Merck & Co. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- Bleeding Academic Research Consortium
- coronary artery bypass graft
- confidence interval
- electronic case report form
- Global Use of Strategies to Open Occluded Coronary Arteries
- hazard ratio
- non–ST-segment elevation acute coronary syndrome
- percutaneous coronary intervention
- Thrombolysis In Myocardial Infarction
- Received December 16, 2015.
- Revision received February 9, 2016.
- Accepted February 24, 2016.
- American College of Cardiology Foundation
- Merlini P.A.,
- Bauer K.A.,
- Oltrona L.,
- et al.
- Mehran R.,
- Rao S.V.,
- Bhatt D.L.,
- et al.
- Ndrepepa G.,
- Schuster T.,
- Hadamitzky M.,
- et al.
- Vranckx P.,
- Leonardi S.,
- Tebaldi M.,
- et al.
- Anderson J.L.,
- Adams C.D.,
- Antman E.M.,
- et al.
- Bassand J.P.,
- Hamm C.W.,
- Ardissino D.,
- et al.
- Rao A.K.,
- Pratt C.,
- Berke A.,
- et al.
- ↵U.S. Food and Drug Administration. Guidance for Industry: Oversight of Clinical Investigations—A Risk-Based Approach to Monitoring. August 2013. Available at: http://www.fda.gov/downloads/Drugs/.../Guidances/UCM269919.pdf. Accessed March 4, 2016.
- Amin A.P.,
- Bachuwar A.,
- Reid K.J.,
- et al.
- Hochholzer W.,
- Wiviott S.D.,
- Antman E.M.,
- et al.