Author + information
- Received August 17, 2015
- Revision received December 21, 2015
- Accepted January 5, 2016
- Published online March 22, 2016.
- Renato D. Lopes, MD, PhDa,b,∗ (, )
- Sergio Leonardi, MD, MHSc,
- Benjamin Neely, MSa,
- Megan L. Neely, PhDa,
- E. Magnus Ohman, MDa,b,
- Diego Ardissino, MDd,
- Christian W. Hamm, MD, PhDe,
- Shaun G. Goodman, MD, MScf,g,
- Deepak L. Bhatt, MD, MPHh,
- Harvey D. White, MB, ChB, DSci,
- Dorairaj Prabhakaran, MD, DM, MScj,
- Felipe Martinez, MDk,
- Jose C. Nicolau, MD, PhDl,
- Kenneth J. Winters, MDm,
- Keith A.A. Fox, MB, ChBn,
- Paul W. Armstrong, MDg and
- Matthew T. Roe, MD, MHSa,b
- aDuke Clinical Research Institute, Duke University School of Medicine, Durham, North Carolina
- bDivision of Cardiology, Department of Medicine, Duke University School of Medicine, Durham, North Carolina
- cFondazione IRCCS Policlinico San Matteo, Pavia, Italy
- dOspedale Maggiore di Parma, Parma, Italy
- eKerckhoff Heart and Thorax Centre, Bad Nauheim, Germany
- fDivision of Cardiology, Department of Medicine, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada
- gCanadian VIGOUR Centre, Division of Cardiology, Department of Medicine, University of Alberta, Edmonton, Alberta, Canada
- hBrigham and Women’s Hospital Heart & Vascular Center and Harvard Medical School, Boston, Massachusetts
- iGreen Lane Cardiovascular Service, Auckland City Hospital, Auckland, New Zealand
- jCentre for Chronic Disease Control and Public Health Foundation of India, New Delhi, India
- kCordoba National University, Cordoba, Argentina
- lHeart Institute—InCor, University of São Paulo Medical School, São Paulo, Brazil
- mEli Lilly and Company, Indianapolis, Indiana
- nCentre for Cardiovascular Science, University of Edinburgh, Edinburgh, Scotland, United Kingdom
- ↵∗Reprint requests and correspondence:
Dr. Renato D. Lopes, Division of Cardiology, Department of Medicine, Duke University School of Medicine, Duke Clinical Research Institute, Box 3850, 2400 Pratt Street, Terrace Level, Room 0311, Durham, North Carolina 27705.
Background Patients with acute coronary syndrome (ACS), especially those receiving medical management without revascularization, are at high risk for spontaneous myocardial infarction (MI), but its frequency and predictors are unknown.
Objectives This study sought to characterize spontaneous MI events in a randomized population during 30 months of follow-up and develop a prediction model for spontaneous MI to assign risk of spontaneous MI events in ACS populations.
Methods We analyzed data from the randomized TRILOGY ACS (TaRgeted platelet Inhibition to cLarify the Optimal strateGy to medically manage Acute Coronary Syndromes) trial of aspirin plus prasugrel or clopidogrel following ACS. The trial included 9,326 patients with non–ST-segment elevation myocardial infarction (NSTEMI)/unstable angina (UA) who were managed medically without planned revascularization. Our study population included 9,294 patients. A multivariable Cox proportional hazards model was developed to determine predictors of time to first spontaneous MI event through 30 months. After model validation, we developed a calculator for model implementation.
Results Among 9,294 patients, 695 spontaneous MI events occurred over a median of 17 months, representing 94% of adjudicated MI events (n = 737). The Kaplan-Meier event rate of spontaneous MI through 30 months was 10.7%. The strongest predictors of spontaneous MI were older age, NSTEMI versus UA as index event, diabetes mellitus, no pre-randomization angiography, and higher baseline creatinine values. The model exhibited good predictive capabilities (c-index = 0.732) and had good calibration, especially for patients with low-to-moderate risk of spontaneous MI.
Conclusions Spontaneous MI following a medically managed UA/NSTEMI event is common. Baseline characteristics can be used to predict subsequent risk of spontaneous MI in this population. These findings provide insight into the long-term natural history of medically managed UA/NSTEMI patients and could be used to optimize risk stratification and treatment of these patients. (A Comparison of Prasugrel and Clopidogrel in Acute Coronary Syndrome Subjects [TRILOGY ACS]; NCT00699998)
Although the use of coronary revascularization has increased over time in patients with non–ST-segment elevation acute coronary syndromes (NSTE-ACS), approximately 30% to 40% of NSTE-ACS patients are still managed medically. However, these patients have not been the focus of contemporary randomized clinical trials (1–3). Compared with patients who received percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG), medically managed NSTE-ACS patients are at increased risk for recurrent ischemic events, specifically myocardial infarction (MI). Spontaneous MI is the most common long-term complication following NSTE-ACS (4). Unlike procedure-related MI, a condition whose optimal definition and clinical implications are still debated, there is broad consensus regarding the clinical relevance of spontaneous MI, and its association with mortality has been extensively verified (5,6). Additionally, a study observed that spontaneous MI is perceived by patients as being equally important as death or stroke (7). Spontaneous MIs are particularly relevant to NSTE-ACS patients who are managed medically following the index acute coronary syndrome (ACS) event, but the frequency and predictors of such events in this setting have not been delineated (8,9).
The TRILOGY ACS (TaRgeted platelet Inhibition to cLarify the Optimal strateGy to medically manage Acute Coronary Syndromes) trial was designed to test if more potent P2Y12 inhibition with prasugrel versus clopidogrel would reduce recurrent cardiovascular events in NSTE-ACS patients managed medically without planned revascularization. We sought to comprehensively characterize patients with spontaneous MI events during long-term follow-up through 30 months to develop a prediction model for time to first spontaneous MI event.
The design and results of the TRILOGY ACS trial have been described (10,11). Briefly, TRILOGY ACS was a phase 3, randomized, double-blind, double-dummy trial that enrolled 9,326 patients with NSTE-ACS at 966 sites worldwide. Patients with unstable angina (UA) or non–ST-segment elevation myocardial infarction (NSTEMI) and at least 1 of 4 high-risk features (age ≥60 years, diabetes mellitus, previous MI, and/or previous PCI/CABG) were eligible for enrollment if they planned to undergo medical management without coronary revascularization within 10 days following the index ACS event. Participants were randomly assigned to receive either clopidogrel or prasugrel. In thienopyridine-naïve patients, prasugrel was given as a 30-mg loading dose, followed by a maintenance dose of 10 mg/day (5 mg/day for patients age ≥75 years or <60 kg body weight); clopidogrel was given as a 300-mg loading dose followed by a 75-mg/day maintenance dose. In addition to randomized treatment, all patients received background aspirin therapy, with a dose of ≤100 mg/day strongly recommended. For patients enrolled after initial dosing with clopidogrel, the loading doses were omitted.
The primary efficacy endpoint was a composite of cardiovascular death, MI, or stroke at 30 months in patients age <75 years in the intention-to-treat population (n = 7,243). Key safety outcomes included moderate to severe/life-threatening bleeding unrelated to CABG as defined by GUSTO (Global Use of Strategies to Open Occluded Coronary Arteries) and major bleeding unrelated to CABG as defined by TIMI (Thrombolysis In Myocardial Infarction) criteria.
Definition of MI
An independent clinical events committee that was blinded to randomized treatment allocation adjudicated all endpoint events, including MIs. All adjudicated MI events were prospectively classified according to the first Universal Definition of MI as spontaneous or procedure-related (12). Of note, at the time these classifications were being planned for TRILOGY ACS, the Third Universal Definition of MI (13) had not yet been published. For this reason, we did not adjudicate MI events as type 1 to 4 MI, but only as spontaneous or procedure-related (PCI- or CABG-related). In TRILOGY ACS, a spontaneous MI event could occur in 1 of 2 circumstances: 1) a cardiac troponin (T or I) or a creatine kinase-myoglobin value greater than the institutional upper limit of normal associated with evidence of ischemia (ischemic symptoms for ≥20 min and/or an ST-segment deviation ≥1 mm in ≥1 electrocardiogram leads); or 2) in the absence of cardiac markers, or if these were not stable or falling before the new suspected event, the presence of an elevation or re-elevation of the ST-segment associated with ischemic symptoms for ≥20 min and/or hemodynamic decompensation.
Baseline clinical characteristics were compared among subjects who did and did not experience spontaneous MI during follow-up using hazard ratios (HRs), 95% confidence intervals (CIs), and p values from a Cox proportional hazards model. Continuous variables are presented as medians with 25th and 75th percentiles; categorical variables are presented as counts (proportions).
The endpoint of interest for the present analysis was the time to first spontaneous MI. A multivariable Cox proportional hazards model was used to determine predictors of spontaneous MI from baseline candidate variables known at the time of randomization. A total of 26 candidate predictors of spontaneous MI were selected a priori on the basis of clinical knowledge: race, weight, age, sex, admission diagnosis (UA or NSTEMI), Killip class on presentation, family history of coronary artery disease, hypertension, hyperlipidemia, diabetes mellitus, current/recent smoker, prior MI, prior PCI, prior CABG, history of peripheral artery disease, history of atrial fibrillation, history of heart failure, baseline systolic blood pressure, baseline heart rate, coronary angiography pre-randomization, baseline hemoglobin, baseline creatinine, and concomitant therapies at baseline (including beta-blockers, angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, statins, and/or proton-pump inhibitors). For each variable, model assumptions were tested and appropriate transformations applied when violations occurred.
We used Fast False Selection Rate (14) for variable selection. This method allowed us to control the expected proportion of uninformative variables that selected in the model at a pre-specified level (5% for the present study). A total of 17 of the 26 variables originally considered were included in the prediction model. Creatinine level, as a continuous variable, was entered into the model with a linear spline (2 degrees of freedom with a knot point of 0.85 mg/dl) on the basis of an observed nonlinear association in both adjusted and unadjusted models. Sensitivity analyses comparing the cumulative incidence curve and the typical 1-Kaplan-Meier curve were similar, indicating that the competing risk of death was low in the population studied; therefore, the model provided the net survival probabilities (ignoring the competing risk of death).
After development, we assessed the predictive accuracy of our model using Harrell’s c statistic (15). To assess model calibration, we used the cumulative incidence estimates and created a calibration plot using methods described by Steyerberg (16) in the overall population as well as the following key subgroups: admission diagnosis (UA vs. NSTEMI), pre-randomization angiography (no vs. yes), and age group (<75 years vs. ≥75 years). Finally, we applied internal bootstrapping validation and used optimism-corrected performance estimates to identify any potential over-fitting during the development of individual model components and summarized the results.
For clinical application, we developed an Excel spreadsheet (Microsoft, Redmond, Washington) in which predictors can be entered. The predicted risk of spontaneous MI with 95% CIs is automatically calculated at a specified time for up to 900 days (available online) (Online Appendix 1). We also developed a web site (17) to help assess patient risk. All of the code used to build the web site is open-source and available for sharing (18). For those interested in incorporating the model with other technologies (e.g., mobile application), we have also made the model available in Predictive Model Markup Language. Predictive Model Markup Language is an XML-based format that can be used to share predictive models across a wide range of programming languages (Online Appendix 2). All analyses presented herein were pre-specified in the main trial and performed by independent statisticians at the Duke Clinical Research Institute using SAS version 9.3 (SAS Institute, Cary, North Carolina) and R version 3.1.0 (R Foundation for Statistical Computing, Vienna, Austria).
From the overall TRILOGY ACS population of 9,326 patients, 9,294 patients were used for this analysis because it was unclear whether 32 reported MI events were spontaneous or procedure-related. From this population, a total of 983 adjudicated MI events were ascertained over 30 months. Of these, 737 (75%) were first MI events and 246 (25%) were second or recurrent MI events. Of the 737 first MI events, 695 (94%) were classified as spontaneous MIs and used as endpoints in the prediction model. Of these spontaneous MI events, 57 (8.2%) were STEMI, 603 (86.8%) were NSTEMI, 22 (3.2%) were Q-wave MI, and 638 (91.8%) were non–Q-wave MI. Data on spontaneous MI subtype was missing in 35 patients (5.0%).
Figure 1 shows Kaplan-Meier event rates for first spontaneous MI. The frequency of spontaneous MI was 1.4% at 30 days, 3.0% at 90 days, 6.2% at 365 days, and 10.7% through 30 months. Baseline characteristics of study participants with and without a first spontaneous MI are provided in Table 1, where univariable statistical analysis indicated that many baseline factors were associated with spontaneous MI. In general, patients who developed subsequent spontaneous MI were older, had more comorbidities, presented more commonly with NSTEMI (vs. UA), and received more medications at discharge, suggesting a perceived higher risk of future cardiovascular events. The final multivariable prediction model for a first spontaneous MI event included 17 variables (Table 2). The model’s ability to distinguish patients who had an event from those who did not was good, with a Harrell’s c Index of 0.732 (standard error = 0.011), meaning that the probability of concordance between predicted and observed responses is 73.2%.
The overall calibration plot (Figure 2), which shows how well predicted probabilities agree with actual observed risk, indicates an excellent calibration for low and intermediate predicted probabilities of spontaneous MI, whereas high predicted probabilities of spontaneous MI were less well calibrated. Similar results were observed in key subgroups (Online Figures 1 to 3 in Online Appendix 3). Finally, the low optimism estimate (0.005) indicates that this model should perform similarly for other medically managed UA/NSTEMI populations. Two practical examples (1 low-risk patient and 1 high-risk patient) demonstrating the application of the Excel spreadsheet calculator are illustrated in Figure 3. For any patient, predictor information can be entered into the calculator to produce: 1) a point estimate and CI of his/her risk of a spontaneous MI for any time point between 0 to 1,200 days after a UA/NSTEMI event; and 2) a figure that allow clinicians to visualize how the patient’s estimated risk changes over time, with error bars showing the variability associated with these estimates.
In a large, contemporary NSTE-ACS population managed medically without coronary revascularization, the overall incidence of spontaneous MI at 30 months was close to 11%. We comprehensively assessed many potential predictors of spontaneous MI, derived a prediction model, and also developed a simplified, pragmatic risk calculator for model implementation that can facilitate individualized risk prediction during follow-up through 30 months.
Our results add to a growing body of evidence indicating that spontaneous MI is a common and clinically important event following ACS (5–7). In a recently reported registry of 8,582 patients undergoing successful implantation of drug-eluting stents, the incidence of nonperiprocedural MI at 2 years was 3.3% in the overall population and 4.1% in the 4,205 patients presenting with ACS (19), much lower than the incidence we observed in our analysis. Also, only 30% of these MIs were classified as type 1 MI according to the universal MI definition, whereas most MIs that occurred during the follow-up interval (46%) were stent-related (stent thrombosis or restenosis; i.e., type 4b or 4c), with the highest risk of subsequent death observed after nonprocedural MIs due to stent thrombosis. These findings indicate that the risk, predictors, and vulnerability to spontaneous MI differ markedly in patients undergoing coronary revascularization as compared with patients managed without revascularization, and further underscores the importance of predicting and potentially preventing spontaneous MI events in medically managed patients.
In a systematic review of large cardiovascular clinical trials, MI was the most common endpoint, representing 45.3% of all first events, whereas spontaneous MI has been shown to be the most common type of MI in recent cardiovascular trials with long-term follow-up (4,5,20). Furthermore, the ominous prognosis carried by spontaneous MI has been demonstrated by a number of recent studies that have consistently observed a higher risk of subsequent mortality following a spontaneous versus procedural MI event (5,6). These observations highlight the importance of being able to identify patients at risk for a spontaneous MI event with an accurate risk prediction tool.
Although clinical trials and registries have traditionally focused on NSTE-ACS patients managed with coronary revascularization (21,22), in clinical practice up to one-third of these patients are managed medically without revascularization (1–3). The TRILOGY ACS dataset, therefore, provides a unique opportunity to model predictors of spontaneous MI in this understudied group of patients, who are typically under-treated with evidence-based medical therapies and who are especially vulnerable to spontaneous MI events, given that revascularization procedures were not performed during the index hospitalization for ACS (23,24). The variables with strongest explanatory power for predicting spontaneous MI were: 1) age; 2) an admission diagnosis of NSTEMI (vs. unstable angina) for the index event (further confirming and expanding the role of troponin in risk assessment for medically managed NSTE-ACS patients and/or perhaps indicating limitations in the diagnosis of UA with increasingly sensitive troponin assays that may be producing false positive results in patients without significant coronary artery disease); 3) diabetes mellitus; and 4) the lack of pre-randomization angiography, which likely reflects the burden of comorbidities in patients for whom revascularization was not an option (25). These unique findings provide important observations about the natural history of spontaneous MI events following ACS.
Clinical perspectives and risk calculator
This model may provide relevant advantages for clinicians and investigators, with important implications for clinical practice and clinical research (Central Illustration). MI is a common and important complication of NSTE-ACS (7). To maximize the usefulness of this model and facilitate its dissemination, we developed an Excel-based risk calculator (Online Appendix 1). This calculator (Figure 3), primarily intended for practicing clinicians, provides real-time and individualized time-varying risk estimates (with error estimates) on the basis of the 17 variables used in the prediction model. By providing time-dependent risk estimates over an extended period of follow-up after medical management for an NSTE-ACS event, this tool may have important clinical applications, informing clinicians and patients of the risk of spontaneous MI. This enhanced awareness of MI risk could reinforce the importance of provision of, and adherence to, evidence-based secondary prevention medications and lifestyle changes. Spontaneous MI also may be the most common and sensitive metric to assess the potential benefit of new post-ACS therapies. The risk prediction tool may also help inform the design of future ACS clinical trials by providing a validated mechanism by which patient populations that are at increased risk for MI during long-term follow-up could be better identified and, therefore, improve the planning and design of future post-ACS clinical trials.
First, we did not capture most in-hospital events, given the time lag typically necessary to confirm that patients were to be medically managed, with a median time from onset of ACS to randomization of 4 to 5 days. Second, type 1 versus type 2 spontaneous MI events were not separately distinguished by the adjudication process, because the trial events adjudication charter and plans were finalized before publication of the Third Universal Definition of MI in 2012 that developed the definitions of these separate types of MI events (13). Finally, while the lack of pre-randomization angiography was a significant predictor of spontaneous MI events, this variable should be interpreted within the context of the trial design. Angiography was not required for trial enrollment (and many enrolling sites did not have angiographic capabilities), but if performed, documentation of moderate coronary disease (at least 1 lesion with stenosis >30%) was mandated for patients without prior PCI or CABG. Thus, patients who did not undergo angiography may have been inherently different in terms of future ischemic risk compared with those who underwent angiography.
The incidence of spontaneous MI following medical management for NSTE-ACS was approximately 11% through 30 months of follow-up. Key baseline characteristics can be used to predict the subsequent risk of spontaneous MI in this population and could potentially optimize risk stratification of patients at highest risk for this prognostically important complication. Our findings provide unique insights into the long-term natural history of medically managed UA/NSTEMI patients and may help inform clinical practice and the design of clinical trials evaluating novel ACS therapies for this under-studied population.
COMPETENCY IN MEDICAL KNOWLEDGE: A total of 1 of 10 patients with NSTE-ACS managed without revascularization experience re-infarction in the following 2.5 years. The risk can be calculated on the basis of several clinical variables, including age, index MI versus unstable angina, diabetes, serum creatinine level, and initial angiographic assessment.
TRANSLATIONAL OUTLOOK: Future clinical trials should evaluate the utility of this type of predictive model to select patients with NSTE-ACS for more aggressive medical therapy or elective revascularization to prevent later spontaneous MI.
The authors thank Karen Pieper, MS, for expert coordination and management of the statistical analytic team, and Jonathan McCall, MS, for expert editorial assistance. Ms. Pieper and Mr. McCall are employees of the Duke Clinical Research Institute, Durham, North Carolina; neither received any compensation for their work on this manuscript other than their usual salaries.
For supplemental figures, the Excel calculator, and a pmml file, please see the online version of this article.
The TRILOGY ACS study was supported by Daiichi-Sankyo Incorporated and Eli Lilly and Company. The study sponsors had no role in the conception and design of this study or in creating the first draft of the manuscript. An employee of Eli Lilly (Dr. Winters) participated as an author during subsequent drafts of the manuscript. All data analyses were performed independently by the Duke Clinical Research Institute. Dr. Lopes has received research grants and consulting fees from Bristol-Myers Squibb, Merck, Portola, and GlaxoSmithKline; and has had consultancies for Bayer, Boehringer Ingelheim, and Pfizer. Dr. Leonardi has received consulting fees from Eli Lilly and Daiichi-Sankyo. Dr. Ohman has received grant support and travel expenses from Daiichi-Sankyo and Eli Lilly; has received consulting fees from Abiomed, AstraZeneca, Biotie, Boehringer Ingelheim, Bristol-Myers Squibb, Daiichi-Sankyo, Eli Lilly, Gilead Sciences, Janssen Pharmaceuticals, Liposcience, Merck, Pozen, Hoffmann-La Roche, Sanofi, Stealth Peptides, The Medicines Company, Medscape, and Web MD; has received grant support from Gilead Sciences and Janssen Pharmaceuticals; and has received lecture fees from Gilead Sciences, Boehringer Ingelheim, and The Medicines Company. Dr. Ardissino has received consulting fees/honoraria from AstraZeneca, Boehringer Ingelheim, Johnson & Johnson, Bayer, Eli Lilly, GlaxoSmithKline, Boston Scientific, Bristol-Myers Squibb, Pfizer, Menarini, Novartis, and Daiichi-Sankyo; and has received grants/travel expenses from AstraZeneca, Bayer, GlaxoSmithKline, Eli Lilly, Pfizer, and Novartis. Dr. Hamm has received consulting fees/honoraria from Brahms, Daiichi-Sankyo, Abbott, Boehringer Ingelheim, AstraZeneca, GlaxoSmithKline, Medtronic, Bayer, and Sanofi. Dr. Goodman has received consulting fees from Sanofi, Eli Lilly, AstraZeneca, Bayer, and Bristol-Myers Squibb; and has received research grants from Johnson & Johnson, AstraZeneca, Bristol-Myers Squibb, Sanofi, Eli Lilly, Boehringer Ingelheim, Bayer, Merck, Daiichi-Sankyo, Servier, and Pfizer. Dr. Bhatt has served on the advisory board of Cardax, Elsevier Practice Update Cardiology, Medscape Cardiology, and Regado Biosciences; has served on the board of directors of Boston VA Research Institute, Society of Cardiovascular Patient Care; is chair of the American Heart Association Get With The Guidelines Steering Committee; is on data monitoring committees for Duke Clinical Research Institute, Harvard Clinical Research Institute, Mayo Clinic, and the Population Health Research Institute; has received honoraria from the American College of Cardiology (Senior Associate Editor, Clinical Trials and News, ACC.org), Belvoir Publications (Editor-in-Chief, Harvard Heart Letter), Duke Clinical Research Institute (clinical trial steering committees, including TRILOGY ACS), Harvard Clinical Research Institute (clinical trial steering committee), HMP Communications (Editor-in-Chief, Journal of Invasive Cardiology), Journal of the American College of Cardiology (Associate Editor; Section Editor, Pharmacology), Population Health Research Institute (clinical trial steering committee), Slack Publications (Chief Medical Editor, Cardiology Today’s Intervention), and WebMD (CME steering committees); has served as Deputy Editor of Clinical Cardiology; has received research funding from Amarin, AstraZeneca, Bristol-Myers Squibb, Eisai, Ethicon, Forest Laboratories, Ischemix, Medtronic, Pfizer, Roche, Sanofi, and The Medicines Company; and has performed unfunded research for FlowCo, PLx Pharma, and Takeda. Dr. White has received grant support from Sanofi, Eli Lilly and Company, the National Institutes of Health (NIH), Merck Sharp & Dohme, AstraZeneca, GlaxoSmithKline, Daiichi-Sankyo Pharma Development, George Institute, Omthera Pharmaceuticals, Pfizer New Zealand, Intarcia Therapeutics Inc., Elsai Inc., DalGen Products and Services; and participates in advisory boards for AstraZeneca. Dr. Prabhakaran has received research grants from Eli Lilly and Medtronic; and has received honoraria from Eli Lilly. Dr. Martinez has received consulting fees/honoraria from Eli Lilly and Daiichi-Sankyo. Dr. Nicolau has received consulting fees from AstraZeneca, Sanofi, and Bayer; has received research grants from Sanofi, GlaxoSmithKline, Bayer, and Novartis; and has received honoraria from Sanofi, Daiichi-Sankyo, AstraZeneca, Bayer, and Bristol-Myers Squibb. Dr. Winters is an employee and minor stockholder of Eli Lilly. Dr. Fox has received research grants from Lilly, Bayer, Johnson & Johnson, and AstraZeneca; has received speakers bureau payments from Bayer, Johnson & Johnson, AstraZeneca, and Sanofi, and has received consulting/other payments from Lilly, Bayer, Johnson & Johnson, AstraZeneca, Sanofi, Boehringer Ingelheim, and Eli Lilly. Dr. Armstrong has received consulting fees from Eli Lilly, Hoffmann-La Roche, Merck, Axio Research, and Orexigen; has received grant support from Boehringer Ingelheim, Hoffmann-La Roche, Sanofi, Scios, Ortho Biotech, Johnson & Johnson, Janssen Pharmaceuticals, GlaxoSmithKline, Amylin Pharmaceuticals, and Merck; and has received payment for developing educational presentations from AstraZeneca and Eli Lilly. Dr. Roe receives research funding from Eli Lilly and Company, Sanofi, Daiichi-Sankyo, Janssen Pharmaceuticals, Ferring Pharmaceuticals, American College of Cardiology, American Heart Association, Familial Hypercholesterolemia Foundation; and consulting or honoraria from Pri-Med, AstraZeneca, Boehringer Ingelheim, Merck, Amgen, Myokardia, Eli Lilly, and Elsevier Publishers. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute coronary syndrome
- myocardial infarction
- non-ST-segment elevation acute coronary syndrome
- non–ST-segment elevation myocardial infarction
- unstable angina
- Received August 17, 2015.
- Revision received December 21, 2015.
- Accepted January 5, 2016.
- American College of Cardiology Foundation
- Masoudi F.A.,
- Ponirakis A.,
- Yeh R.W.,
- et al.
- Chan M.Y.,
- Mahaffey K.W.,
- Sun L.J.,
- et al.
- Leonardi S.,
- Armstrong P.W.,
- Schulte P.J.,
- Ohman E.M.,
- Newby L.K.
- Damman P.,
- Wallentin L.,
- Fox K.A.,
- et al.
- Stolker J.M.,
- Spertus J.A.,
- Cohen D.J.,
- et al.
- Tian J.,
- Dauerman H.,
- Toma C.,
- et al.
- Chin C.T.,
- Roe M.T.,
- Fox K.A.,
- et al.
- Thygesen K.,
- Alpert J.S.,
- White H.D.,
- for the Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of Myocardial Infarction
- Thygesen K.,
- Alpert J.S.,
- Jaffe A.S.,
- et al.
- Steyerberg E.
- ↵TRILOGY Risk Models. Available at: http://trilogyacs-bleedingmodel-01.oit.duke.edu/. Accessed February 11, 2016.
- ↵Neely B. General dockerized shiny application to provide a visualization and interface to a Cox proportional hazards model. GitHub Repository. 2015. Available at: https://github.com/benneely/dockerized-shinyApp-coxModel-pmml. Accessed October 10, 2015.
- Dohi T.,
- Maehara A.,
- Witzenbichler B.,
- et al.
- Leonardi S.,
- Tricoci P.,
- White H.D.,
- et al.
- Fox K.A.,
- Dabbous O.H.,
- Goldberg R.J.,
- et al.