Differential occurrence, profile, and impact of first recurrent cardiovascular events after an acute coronary syndrome

Clinical Investigations

Differential occurrence, profile, and impact of first recurrent cardiovascular events after an acute coronary syndrome Connie N. Hess, MD, MHS, a Robert M. Clare, MS, b Megan L. Neely, PhD, b Pierluigi Tricoci, MD, b,c Kenneth W. Mahaffey, MD, d Stefan K. James, MD, PhD, e,f John H. Alexander, MD, MHS, b,c Claes Held, MD, e,f Renato D. Lopes, MD, PhD, b,c Keith A. A. Fox, MB, ChB, g Harvey D. White, MB, ChB, DSc, h Lars Wallentin, MD, PhD, e,f Paul W. Armstrong, MD, i Robert A. Harrington, MD, d Erik Magnus Ohman, MD, b,c and Matthew T. Roe, MD, MHS b,c Aurora, Denver, CO; Durham, NC; Stanford, CA; Uppsala, Sweden; Scotland, United Kingdom; Auckland, New Zealand; and Alberta, Canada

Objective Acute coronary syndrome (ACS) trials typically use a composite primary outcome (myocardial infarction [MI], stroke, or cardiovascular death), but differential patient characteristics, timing, and consequences associated with individual component end points as first events have not been well studied. We compared patient characteristics and prognostic significance associated with first cardiovascular events in the post-ACS setting for initially stabilized patients. Methods We combined patient-level data from 4 trials of post-ACS antithrombotic therapies (PLATO, APPRAISE-2, TRACER, and TRILOGY ACS) to characterize the timing of and characteristics associated with first cardiovascular events (MI, stroke, or cardiovascular death). Landmark analysis at 7 days after index ACS presentation was used to focus on spontaneous, postdischarge events that were not confounded by in-hospital procedural complications. Using a competing risk framework, we tested for differential associations between prespecified covariates and the occurrence of nonfatal stroke vs MI as the first event, and we examined subsequent events after the first nonfatal event. Results Among 46,694 patients with a median follow-up of 358 (25th, 75th percentiles 262, 486) days, a first ischemic event occurred in 4,307 patients (9.2%) as follows: MI in 5.8% (n = 2,690), stroke in 1.0% (n = 477), and cardiovascular death in 2.4% (n = 1,140). Older age, prior stroke/transient ischemic attack, prior atrial fibrillation, and higher diastolic blood pressure were associated with a significantly greater risk of stroke vs MI, whereas prior percutaneous coronary intervention was associated with a greater risk of MI vs stroke. Second events occurred in 32% of those with a first nonfatal stroke at a median of 13 (3, 59) days after the first event and in 32% of those with a first nonfatal MI at a median of 35 (5, 137) days after the first event. The most common second event was a recurrent MI among those with MI as the first event and cardiovascular death among those with stroke as the first event. Conclusions Approximately 9% of patients experienced a first cardiovascular event in the post-ACS setting during a median follow-up of 1 year. Although the profile and prognostic implications of stroke vs MI as the first nonfatal event differ substantially, approximately one-third of these patients experienced a second event, typically soon after the first event. These findings have implications for improving post-ACS care and influencing the design of future cardiovascular trials. (Am Heart J 2017;0:1-10.) Cardiovascular (CV) disease is the leading cause of death globally, with approximately 7.4 million deaths from ischemic heart disease and 6.7 million stroke-related

deaths per year. 1 Although ischemic heart disease and stroke share similar risk factors and treatments, 2-6 previous CV outcomes trials evaluating parenteral anti-

From the aDivision of Cardiology, Department of Medicine, University of Colorado

Vladimir Dzavik, MD served as guest editor for this article.

School of Medicine, Aurora, CO, b Duke Clinical Research Institute, Durham, NC, c Division of Cardiology, Department of Medicine, Duke University School of Medicine, Durham, NC, d Department of Medicine, Stanford University, Stanford, CA, e Department of Medical Sciences, Cardiology, Uppsala University, Uppsala, Sweden, f Uppsala Clinical Research Center, Uppsala University, Uppsala, Sweden, g British Heart Foundation Centre for Cardiovascular Sciences, University of Edinburgh, Edinburgh, Scotland, United Kingdom, h Auckland City Hospital, Green

Trial registration: ClinicalTrials.gov: PLATO, NCT00391872; APPRAISE-2, NCT00831441; TRACER, NCT00527943; TRILOGY ACS, NCT00699998. Submitted June 17, 2016; accepted January 17, 2017. Reprint requests: Matthew T. Roe, MD, MHS, Duke Clinical Research Institute, 2400 Pratt St, Rm 0311 Terrace Level, Durham, NC 27705. E-mail: [email protected] 0002-8703

Lane Cardiovascular Service, Auckland, New Zealand, and i Division of Cardiology, University of Alberta, Edmonton, Alberta, Canada.

© 2017 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ahj.2017.01.016

American Heart Journal Month Year

2 Hess et al

thrombotic therapies for acute coronary syndromes (ACSs) excluded stroke as an outcome, focusing instead on a common short-term (typically 30 days) composite end point of myocardial infarction (MI) and all-cause death. 7-10 The composition of CV outcomes trial end points shifted when clopidogrel was found to significantly reduce a composite of CV death, MI, or stroke compared with aspirin for patients with stable coronary artery, cerebrovascular, and peripheral vascular disease in the Clopidogrel versus Aspirin in Patients at Risk of Ischemic Events (CAPRIE) trial. 11 Thereafter, stroke was included in the primary composite end point for the first time in a large ACS CV outcomes trial, Clopidogrel in Unstable Angina to Prevent Recurrent Events (CURE). 12 Subsequent trials evaluating the long-term use of oral antithrombotic therapies post-ACS have used a similar primary composite end point of CV death, MI, or stroke and have collectively shown composite event rates of approximately 10% through 12 months. [13-18] Although commonly incorporated into the standard primary composite end point for post-ACS trials, stroke as a unique CV event has not been well studied. Specifically, the timing, patient risk profile, and prognostic significance of post-ACS stroke compared with other CV events remain poorly characterized. We therefore analyzed combined patient-level data from the following trials evaluating post-ACS antithrombotic therapies: Platelet Inhibition and Patient Outcomes (PLATO), 13 Apixaban for Prevention of Acute Ischemic Events 2 (APPRAISE-2), 14 Thrombin Receptor Antagonist for Clinical Event Reduction in Acute Coronary Syndrome (TRACER), 15 and Targeted Platelet Inhibition to Clarify the Optimal Strategy to Medically Manage Acute Coronary Syndromes (TRILOGY ACS). 16 Our objectives were to (1) describe patient characteristics according to first post-ACS CV event (stroke, MI, or CV death), (2) examine the differential timing and trajectory of first CV events, and (3) compare patient characteristics and downstream outcomes by type of first nonfatal CV event (stroke vs MI).

(Supplementary Table III) because of an ischemic event (MI, stroke, CV death; n = 1,176) or non-CV death (n = 12) occurring within 7 days of index ACS presentation or study discontinuation during this time frame without complete ascertainment of nonfatal end points during subsequent trial follow-up (n = 351). We chose a 7-day landmark time point for this analysis because of the different enrollment periods across the trials, with APPRAISE-2 and TRILOGY ACS having longer enrollment windows, and to ensure a common starting time for end point ascertainment. We designed this analysis to study the secondary prevention phase of post-ACS treatment when most CV events would be expected to be spontaneous (ie, unrelated to in-hospital ACS treatments, including medications and revascularization procedures) and therefore more likely to be comparable across the 4 trials. Accordingly, we included only patients without early (within 7 days) in-hospital events. A total of 65 patients were excluded due to inability to determine the timing of their first recurrent CV event within the merged dataset. Our final analysis population consisted of 46,694 stabilized post-ACS patients without a CV event within the first 7 days after ACS presentation. Baseline characteristics according to source trial are shown in Supplementary Table IV.

Outcomes and definitions

Methods

Primary outcomes included MI, stroke (including ischemic and hemorrhagic strokes, as per stroke end points in the 4 trials studied), and CV mortality. 13-16 All-cause mortality and stroke subtypes (ischemic and hemorrhagic) were secondary outcomes. End points were independently adjudicated and classified by clinical events committees for each trial according to similar prespecified end point definitions. 13-16 APPRAISE-2, PLATO, and TRACER classified MI end points based on the 2007 universal definition of MI as types 1-5 19; types 1-3 MI events were considered spontaneous for this analysis. TRILOGY ACS classified MI end points as spontaneous or procedural based on an adaptation of the universal definition of MI and did not classify MI end points as types 1-5. [18]

Data sources and study population Data were merged from 4 CV outcomes trials of post-ACS antithrombotic therapy for which we had access to patient-level data: PLATO, APPRAISE-2, TRACER, and TRILOGY ACS. Trial characteristics and major inclusion/exclusion criteria are shown in Supplementary Tables I and II, respectively. 13-16 In APPRAISE-2 and TRILOGY ACS, patients were randomized close to discharge from index hospitalization, whereas patients in PLATO and TRACER were randomized within 24 hours of index presentation. The combined 4-trial population was 48,286 patients. A total of 1,527 patients were excluded from this analysis

Statistical analysis We examined patient characteristics at randomization stratified by first event type (MI, stroke, CV death, or none). Categorical variables were presented as counts and frequencies, and continuous variables were presented as median values with interquartile ranges (IQRs). Continuous characteristics were compared using the Kruskal-Wallis test, and categorical characteristics were compared using the Pearson χ 2 test; otherwise, a Fisher exact test was used. Outcomes were assessed using landmark analyses beginning 7 days after hospitalization for the index ACS. Median times to first event and IQRs

American Heart Journal Volume 0, Number 0

were calculated. We used a competing risk framework where patients could have only one first recurrent event and constructed cumulative incidence curves representing cumulative cause-specific hazard for MI, stroke, and CV death as first events. Patients who died before MI or stroke were censored at time of death. We assessed factors associated with MI vs stroke as the first nonfatal recurrent event. In our competing risk framework, we fit a model for MI and stroke simultaneously to test for differential associations between each covariate and each end point. 20 The “differential association P value” obtained from this model allowed testing of whether the hazard ratio of a covariate significantly differs in direction and/or magnitude between end points. We considered for inclusion in this model the following baseline covariates, which were chosen a priori based on clinical judgment and consensus among participating investigators: age, race (white vs black, white vs Asian, black vs Asian), weight, sex, hypertension, hyperlipidemia, diabetes, current/recent tobacco use, prior MI, prior percutaneous coronary intervention (PCI), prior coronary artery bypass graft (CABG), prior peripheral artery disease, prior atrial fibrillation, prior heart failure, prior stroke or transient ischemic attack (TIA), type of index ACS event (ST-segment elevation MI [STEMI] vs non–ST-segment elevation ACS), randomization heart rate, randomization systolic blood pressure, randomization diastolic blood pressure, electrocardiogram changes on index presentation for patients with non–ST-segment elevation ACS (transient ST elevation, ST depression, or none), Killip class, randomization hemoglobin, and randomization creatinine. The candidate variables Killip class, hemoglobin, and creatinine were excluded from modeling due to unacceptable degrees of missing data. A sensitivity analysis including trial as a baseline covariate was performed. Finally, we examined subsequent events among patients with a first recurrent nonfatal event (MI or stroke). The relative incidence of a second recurrent event (MI, stroke, CV death, or non-CV death) and median first-to-second-event times were calculated according to first event being MI vs stroke. No imputation was performed for missing data, and a P value of b.05 was considered statistically significant. All analyses were performed by statisticians at DCRI using SAS software version 9.2 (SAS Institute, Inc., Cary, NC) and an independent copy of the combined database.

Results Baseline characteristics Among 46,694 patients followed up for a median (IQR) of 358 (262, 486) days, a first ischemic event occurred in 4,307 patients (9.2%). Among the analysis population, 5.8% (n = 2,690) had MI as the first event, 1.0% (n = 477)

Hess et al 3

had stroke as the first event, and 2.4% (n = 1,140) had CV death as the first event during the trial follow-up periods. Of the 2,690 patients with a first MI event, 81.4% (n = 2,190) were spontaneous, 14.3% (n = 384) were procedural, and 4.3% (n = 116) were not classified due to missing data for MI type. Of the 477 patients with a first stroke event, 79.3% (n = 378) were ischemic, 15.3% (n = 73) were hemorrhagic, and 5.5% (n = 26) were not classified due to missing data for stroke type. Baseline characteristics were examined according to first event type (Table I). Compared with patients without a first recurrent CV event, those with events were older, were more often female, had a higher burden of comorbidities, and less often presented with STEMI. Among patients with a first recurrent event, those with initial CV death were older, more frequently had prior CHF, had higher presentation Killip class, and had lower baseline hemoglobin. Patients with a first MI event more often presented with non–ST-segment elevation ACS, more likely had prior PCI or CABG, and more frequently underwent PCI during the index hospitalization before the 7-day landmark time point. Patients with a first stroke event more likely had prior atrial fibrillation and in-hospital CABG than did patients with an initial MI or CV death.

Timing of first recurrent event The median times (IQRs) to first MI, stroke, and CV death events were 96 (27, 224) days, 114 (26, 263) days, and 112 (28, 265) days, respectively. Figure 1 shows the cumulative incidence curves for each as the first recurrent event. The trajectory for MI incidence was steepest during the first 3 months and then appeared linear thereafter. The trajectories for stroke and CV death incidence were mostly linear throughout follow-up. Factors associated with first nonfatal CV events Factors associated with first nonfatal MI vs stroke were analyzed to determine whether associations were significantly different between end points (Figure 2). Older age, prior stroke/TIA, and prior atrial fibrillation were associated with significantly greater risk of stroke vs MI. Increased diastolic blood pressure was associated with greater stroke risk but decreased MI risk, whereas prior PCI was associated with decreased stroke risk but greater MI risk. Certain factors, including diabetes mellitus, smoking, prior MI, and peripheral artery disease, were associated with greater risk of both stroke and MI, but these associations were not different between the 2 end points (Pdifferential N .05 for all). A sensitivity analysis adjusted for trial showed similar results. Subsequent events after a first nonfatal CV event Table II describes second recurrent events for the 3,167 patients with an initial nonfatal recurrent stroke or MI. Of these 3,167 patients, 13.8% (n = 437) had MI, 2.0% (n =

American Heart Journal Month Year

4 Hess et al

Table I. Baseline characteristics according to type of first CV event MI (n = 2690)

Stroke (n = 477)

CV death (n = 1140)

None (n = 42,387)

P

Demographics Age, y Race White Black Asian Other Female sex Weight (kg)

2318/2688 69/2688 229/2688 72/2688 896/2690 2684, 79

(86.2) (2.6) (8.5) (2.7) (33.3) (68, 90)

394/476 (82.8) 13/476 (2.7) 52/476 (10.9) 17/476 (3.6) 167/477 (35.0) 475, 77 (67, 89)

888/1140 28/1140 183/1140 41/1140 386/1140 1137, 74

(77.9) (2.5) (16.1) (3.6) (33.9) (63, 85)

35,871/42,356 (84.7) 749/42,356 (1.8) 4796/42,356 (11.3) 940/42,356 (2.2) 12,981/42,387 (30.6) 42,302, 79 (69, 90)

b.0001 b.0001

Medical history Hypertension Diabetes mellitus Current tobacco use Hyperlipidemia Prior MI Prior CHF Prior atrial fibrillation Prior PAD Prior PCI Prior CABG Prior stroke or TIA

2220/2690 1253/2689 649/2689 1811/2665 1271/2681 569/2684 176/2668 427/2673 857/2688 560/2688 249/2683

(82.5) (46.6) (24.1) (68.0) (47.4) (21.2) (6.6) (16.0) (31.9) (20.8) (9.3)

390/477 (81.8) 201/477 (42.1) 114/477 (23.9) 282/470 (60.0) 177/475 (37.3) 89/476 (18.7) 48/472 (10.2) 61/476 (12.8) 91/476 (19.1) 65/476 (13.7) 72/477 (15.1)

937/1139 533/1139 239/1140 596/1117 475/1138 321/1138 100/1123 171/1128 238/1139 173/1139 85/1138

(82.3) (46.8) (21.0) (53.4) (41.7) (28.2) (8.9) (15.2) (20.9) (15.2) (7.5)

30,061/42,364 13,328/42,367 12,208/42,379 22,673/41,986 11,955/42,318 4809/42,334 1745/42,213 3304/42,248 8231/42,354 3758/42,368 2398/42,357

(71.0) (31.5) (28.8) (54.0) (28.3) (11.4) (4.1) (7.8) (19.4) (8.9) (5.7)

b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001

(86.0) (14.0)

381/471 (80.9) 90/471 (19.1)

897/1121 (80.0) 224/1121 (20.0)

32,416/41,886 (77.4) 9470/41,886 (22.6)

b.0001 b.0001 b.0001

(85.2) (11.5) (3.2) (0.1) (62, 80)

351/411 (85.4) 48/411 (11.7) 12/411 (2.9) 0/411 (0.0) 477, 70 (62, 80)

(120, 145) (67, 80)

477, 130 (119, 143) 477, 78 (70, 84)

1139, 129 (115, 140) 1139, 75 (68, 80)

(45.6) (4.8)

211/477 (44.2) 29/477 (6.1)

578/1140 (50.7) 58/1140 (5.1)

2690, 68 (59, 76)

Presentation and in-hospital features ACS presentation UA/NSTEMI 2298/2671 STEMI 373/2671 Killip class on presentation I 1978/2322 II 266/2322 III 75/2322 IV 3/2322 Heart rate at randomization, 2681, 70 beats/min Systolic BP at randomization, mm Hg 2687, 130 Diastolic BP at randomization, mm Hg 2687, 75 ECG changes (for UA/NSTEMI) ST depression 1226/2688 Transient ST elevation 128/2688 Laboratory values at randomization Creatinine, mg/dL 2498, 1.0 Hemoglobin, g/dL 2487, 13.3 In-hospital procedures PCI 956/1882 CABG 96/1883

477, 69 (61, 76)

1140, 71 (62, 78)

739/999 199/999 54/999 7/999 1138, 75

(74.0) (19.9) (5.4) (0.7) (66, 85)

(0.8, 1.3) 427, 1.0 (0.8, 1.2) 1060, 1.1 (0.9, 1.4) (12.1, 14.5) 425, 13.6 (12.2, 14.6) 1039, 13.1 (11.7, 14.2) (50.8) (5.1)

152/339 (44.8) 38/340 (11.2)

233/642 (36.3) 40/648 (6.2)

42,386, 64 (56, 71)

33,057/35,592 (92.9) 2172/35,592 (6.1) 329/35,592 (0.9) 34/35,592 (0.1) 42,307, 70 (63, 80) 42,325, 130 (119, 140) 42,325, 78 (70, 83) 18,021/42,366 (42.5) 2890/42,365 (6.8) 38,244, 0.9 (0.8, 1.1) 38,029, 13.8 (12.7, 14.9) 18,612/32,216 (57.8) 1911/32,243 (5.9)

b.0001 b.0001

b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b.0001 b0.0001

Values are n/N (%) or n, median (IQR). Abbreviations: CHF, congestive heart failure; PAD, peripheral artery disease; UA, unstable angina; BP, blood pressure; ECG, electrocardiogram.

63) had stroke, and 13.8% (n = 436) had CV death as a second event. Among 2,690 patients with a first MI event and 477 patients with a first stroke event, 32.2% (n = 866) and 32.1% (n = 153), respectively, had a second event (MI, stroke, CV death, or non-CV death). The median time to second event was 35 days after initial MI vs 13 days after initial stroke. Figure 3 shows the distribution of second events according to first MI vs stroke. Patients with a first MI event and a second event (n = 866) most

commonly had a second MI event (48.0%; n = 416) or CV death (40.9%; n = 354), with median first-to-second-event times (IQRs) of 72 (26, 173) and 5 (1, 34) days, respectively. Patients with an initial stroke and a second event (n = 153) most commonly had CV death (53.6%; n = 82) soon after the first stroke event (median 9 days; IQR 2, 34 days), while non-CV death was the next most common second event (17.6% [n = 27]; median time to event 24 days; IQR 4, 141 days).

American Heart Journal Volume 0, Number 0

Hess et al 5

Figure 1

First recurrent CV events. Curves of cumulative cause-specific risk for MI vs stroke vs CV death as the first recurrent event are depicted.

Discussion Among more than 45,000 initially stabilized patients enrolled in 4 recent post-ACS trials evaluating the long-term use of novel antithrombotic therapies, 9.2% of patients experienced a recurrent CV event during a median follow-up of approximately 1 year with a differential timing and trajectory of incidence for CV death, MI, or stroke as the first event. Certain baseline clinical characteristics were more likely to be associated with stroke vs MI as the first event (older age, prior stroke/TIA, prior atrial fibrillation, and elevated diastolic blood pressure), whereas prior PCI was more likely to be associated with first MI vs stroke. Approximately one-third of patients with a first nonfatal stroke or MI event experienced a second ischemic event with differential timing and distribution of second recurrent events for those with a first stroke vs MI. Previous analyses from smaller databases have associated recurrent MI and stroke events in the post-ACS setting with increased mortality and have identified some clinical factors associated with an increased risk for these events. 21-23 However, most studies focused on a single type of CV event and only considered the time to first event. Others investigated the impact of antithrombotic therapies on total recurrent events, rather than just first CV events, 24-26 but these analyses did not differentiate patient characteristics associated with the first events, did not evaluate relationships between first and subsequent events, and did not characterize how the type of nonfatal ischemic event influences the timing and frequency of

downstream events. In contrast, we simultaneously assessed multiple ischemic event types from an integrated database of 4 global CV trials that included patients from up to 50 countries, and we investigated the differential occurrence of downstream events after a first nonfatal event. We also examined the secondary prevention phase of post-ACS treatment that was not confounded by in-hospital revascularization procedures or use of parenteral antithrombotic medications. In summary, our results provide novel insights into the occurrence and prognostic significance of CV events that are commonly incorporated into the composite primary end points of large-scale post-ACS trials. Although stroke has been included in the primary composite end points for most contemporary post-ACS trials, stroke is distinct from other CV events such as MI and CV death. First, stroke occurs less frequently than MI or CV death in the post-ACS setting, so the impact of stroke in typical time-to-first-event analyses for composite primary end points is proportionally smaller than that of other component end points. 13-18 Second, we found a significantly greater frequency of stroke vs MI as the first event among patients with older age, atrial fibrillation, and prior stroke/TIA and a greater frequency of MI vs stroke for patients with prior PCI. These results highlight the clinical differences between patients with a first nonfatal post-ACS stroke vs MI and may inform efforts to enhance expected event rates when developing protocols for future post-ACS trials. Third, stroke end point definitions in post-ACS trials evaluating antithrombotic

6 Hess et al

American Heart Journal Month Year

Figure 2

Factors associated with MI vs stroke as first nonfatal CV event. The forest plot demonstrates associations of clinical characteristics with MI vs stroke. A differential P value of b.05 indicates that the association is significantly different in magnitude and/or direction across the 2 end points. HR indicates hazard ratio; UA, unstable angina; CHF, congestive heart failure; PAD, peripheral artery disease; BP, blood pressure.

American Heart Journal Volume 0, Number 0

Hess et al 7

Table II. Second events among patients with a first nonfatal MI or stroke MI as first event (n = 2690)

Any event MI Stroke CV death Non-CV death

Stroke as first event (n = 477)

n (%)

Time (first to second event), d, median (IQR)

n (%)

Time (first to second event), d, median (IQR)

866 (32.2) 416 (15.5) 40 (1.5) 354 (13.2) 56 (2.1)

35 72 22 5 117

153 (32.1) 21 (4.4) 23 (4.8) 82 (17.2) 27 (5.7)

13 28 24 9 24

Second event

(5, 137) (26, 173) (2, 61) (1, 34) (28, 300)

Figure 3

(3, (3, (8, (2, (4,

59) 62) 82) 34) 141)

therapies typically include ischemic and hemorrhagic subtypes (as was the case for the 4 trials evaluated in this analysis) and can thus reflect both efficacy (reduced ischemic risk) and safety (increased bleeding risk). Nonetheless, because predictors of post-ACS intracranial hemorrhage are similar to those identified in our study for all stroke events 27 and because most stroke events in our analysis were ischemic strokes, the differentiation of stroke subtypes would not fundamentally change our conclusions. Finally, the downstream impact of nonfatal stroke is more ominous than for nonfatal MI, with higher rates of CV and non-CV death after first stroke vs MI. Therefore, although the pathophysiology of post-ACS stroke events may differ substantially from post-ACS MI events, stroke events have important consequences in the post-ACS setting that have been delineated by our analyses. Our results have implications for improving post-ACS care. Although practice guidelines recommend post-ACS risk stratification, 28,29 currently available tools, such as the Global Registry of Acute Coronary Events risk score for 6-month mortality, 28 are not intended to predict the risk of nonfatal or second (subsequent) events after an initial nonfatal CV event. We demonstrated that approximately one-third of patients with a first nonfatal MI or stroke event have a second event that tends to occur relatively soon after the first event, suggesting that treatment strategies after a first nonfatal ischemic events should be aggressive and focused on mitigating early adverse consequences of the first event. In addition, almost one-third of CV deaths in our study were second events shortly after an initial nonfatal MI or stroke. Therefore, the efficacy of novel therapies for post-ACS treatment may be more comprehensively and accurately assessed by accounting for all recurrent CV events within primary end point analyses (rather than just accounting for the time to first event) and for the timing and clustering of second events after the first nonfatal event.

Distribution of second recurrent events according to type of first nonfatal CV event. The distribution of event types among patients with a second recurrent event according to initial MI vs stroke is shown. Among patients with a first nonfatal MI, 416, 40, 354, and 56 patients experienced subsequent MI, stroke, CV death, or non-CV death, respectively. Among patients with a first nonfatal stroke, 21, 23, 82, and 27 patients experienced subsequent MI, stroke, CV death, or non-CV death, respectively.

Our findings also illustrate issues surrounding the use of composite efficacy outcomes in clinical trials. Although the use of these end points can increase statistical power and lead to reductions in sample size, challenges arise when the component end points have differential post–event clinical impact, when treatments affect component end points differently, and when patients experience multiple recurrent ischemic events. 30-32 The higher relative incidence of death after stroke vs MI as the first nonfatal event in our analysis demonstrates that these component end points have different prognostic implications in the post-ACS setting. Statistical methods that account for the clinical relevance of component end points through differential weighting and for the incorporation of all recurrent ischemic events have been developed and should be explored and considered for evaluating efficacy outcomes in future post-ACS trials to more broadly assess the impact of experimental treatments. [30,32]

Limitations Several limitations should be acknowledged. First, we excluded patients with ischemic events within the first 7 days after index ACS hospitalization because of differences in trial enrollment periods and our desire to focus on primarily spontaneous events occurring after a common starting time point. Second, there were

8 Hess et al

differences in major trial inclusion/exclusion criteria: STEMI patients were only included in PLATO and APPRAISE-2, and medically managed non–ST-segment elevation ACS patients were exclusively enrolled in TRILOGY ACS, whereas the other 3 trials included patients who commonly underwent in-hospital revascularization. Nonetheless, a sensitivity analysis including trial as a covariate when evaluating factors differentially associated with stroke vs MI showed similar results. Third, patients with prior stroke or TIA were excluded from TRILOGY ACS because of contraindicated prasugrel use in this population but were included in the other 3 trials, and this covariate was significantly associated with stroke as the first event. However, all trials excluded patients requiring concomitant treatment with an oral anticoagulant and likely excluded patients with atrial fibrillation with high stroke risk, so the potential impact of different exclusion criteria across trials was likely mitigated. Fourth, we could not assess post–stroke neurologic disability, as Rankin scores were not routinely assessed or recorded. Fifth, we did not assess the impact of randomized antithrombotic therapies and concomitant background medications because of the numerous combinations of drugs and dosages that would confound such analyses. Finally, these results that were generated from analyses of participants in clinical trials may not be generalizable to the broader population of ACS patients treated in routine practice.

Conclusions In this study of 4 contemporary, global trials evaluating post-ACS antithrombotic therapies, 9.2% of patients had a recurrent CV event over a median follow-up of almost 1 year, and approximately one-third of patients with a first nonfatal event experienced a second event. We demonstrated differences in the clinical profile, risk factors, and prognosis for patients with first nonfatal stroke vs MI. These findings should inform clinical decision making during secondary prevention treatment for ACS patients and have implications for the use and interpretation of composite efficacy end points in future post-ACS trials.

Disclosures Dr Hess reports receiving research funding from Gilead Sciences. Dr Tricoci reports receiving research grants from Merck, Sanofi-Aventis, CSL, and Regeneron, and consulting fees or honoraria from Merck and CSL. Dr Mahaffey's financial disclosures before August 1, 2013, can be viewed at https://www.dcri.org/about-us/conflictof-interest/Mahaffey-COI_2011-2013.pdf; disclosures after August 1, 2013, can be viewed at http://med. stanford.edu/profiles/kenneth-mahaffey. Dr James receives institutional research grants from AstraZeneca, Eli Lilly, Bristol-Myers Squibb, Terumo Inc, Medtronic,

American Heart Journal Month Year

and Vascular Solutions; honoraria from The Medicines Company, AstraZeneca, Eli Lilly, Bristol-Myers Squibb, and IROKO; and consultant/advisory board fees from AstraZeneca, Eli Lilly, Merck, Medtronic, and Sanofi. Dr Alexander reports receiving research grants from Boehringer Ingelheim, Bristol-Myers Squibb, CSL Behring, National Institutes of Health, Regado Biosciences, Tenex Therapeutics, and Vivus Pharmaceuticals, and consulting fees or honoraria from Portola Pharmaceuticals, Sohmalution, and the VA Cooperatives Studies Program. Dr Held reports institutional research grants from AstraZeneca, Merck, GlaxoSmithKline, Roche, and Bristol-Myers Squibb/Pfizer, and honoraria/advisory board fees from AstraZeneca. Dr Lopes reports receiving research grants from Bristol-Myers Squibb and GlaxoSmithKline, and consulting fees or honoraria from Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, GlaxoSmithKline, Merck, Pfizer, and Portola Pharmaceuticals. Dr Fox reports receiving research grants from Eli Lilly, Bayer, Johnson & Johnson, and AstraZeneca; speaker's bureau payments from Bayer, Johnson & Johnson, AstraZeneca, and Sanofi-Aventis; and consulting/other payments from Eli Lilly, Bayer, Johnson & Johnson, AstraZeneca, Sanofi-Aventis, Boehringer Ingelheim, and Eli Lilly. Dr White reports receiving grant support from Sanofi-Aventis, Eli Lilly, The Medicines Company, NIH, Pfizer, Roche, Johnson & Johnson, Schering-Plough, Merck Sharpe & Dohme, AstraZeneca, GlaxoSmithKline, Daiichi Sankyo Pharma Development, and Bristol-Myers Squibb; he also participates in advisory boards for Merck Sharpe & Dohme, Roche, and Regado Biosciences. Dr Wallentin reports receiving research grants from AstraZeneca, Merck, Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, and GlaxoSmithKline; consulting fees from Abbott, Merck, Regado Biosciences, Athera Biotechnologies, Boehringer Ingelheim, AstraZeneca, GlaxoSmithKline, and Bristol-Myers Squibb/Pfizer; lecture fees from AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, and GlaxoSmithKline; honoraria from Boehringer Ingelheim, AstraZeneca, Bristol-Myers Squibb/Pfizer, and GlaxoSmithKline; and travel support from AstraZeneca, Bristol-Myers Squibb/Pfizer, and GlaxoSmithKline. Dr Armstrong reports receiving consulting fees from Eli Lilly, Hoffmann-La Roche, Merck, Axio Research, and Orexigen; grant support from Boehringer Ingelheim, Hoffmann-La Roche, Sanofi-Aventis, Scios, Ortho Biotech, Johnson & Johnson, Janssen Pharmaceuticals, GlaxoSmithKline, Amylin Pharmaceuticals, and Merck; and payment for developing educational presentations from AstraZeneca and Eli Lilly. Dr Harrington reports receiving research grants from Merck, AstraZeneca, Bristol-Myers Squibb, and Johnson & Johnson, and consulting fees from CSL, Bristol-Myers Squibb, Daiichi, Gilead, Johnson & Johnson, Merck, and MyoKardia. Dr Ohman reports receiving grant support and travel

American Heart Journal Volume 0, Number 0

expenses from Daiichi Sankyo and Eli Lilly; consulting fees from AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb, Gilead Sciences, Janssen Pharmaceuticals, Liposcience, Merck, Pozen, Hoffmann-La Roche, Sanofi-Aventis, The Medicines Company, and Web MD; grant support from Gilead Sciences; and lecture fees from Gilead Sciences, Boehringer Ingelheim, and The Medicines Company. Dr Roe receives research funding from Eli Lilly, Sanofi-Aventis, Daiichi Sankyo, Janssen Pharmaceuticals, Ferring Pharmaceuticals, the American College of Cardiology, the American Heart Association, and the Familial Hypercholesterolemia Foundation. He also receives consulting payments or honoraria from AstraZeneca, Boehringer Ingelheim, Merck, Amgen, PriMed, and Elsevier Publishers. All disclosures for DCRI authors are listed at https://www.dcri.org/ about-us/conflict-of-interest. The remaining authors have no disclosures. All authors have approved the final article. This work and the TRILOGY ACS study were funded by Daiichi Sankyo and Eli Lilly. PLATO was funded by AstraZeneca; APPRAISE-2 was funded by Bristol-Myers Squibb and Pfizer; and TRACER was funded by Merck & Co. The study sponsors had no role in the conception and design of this study or in creating the first draft of the manuscript.

Appendix. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.ahj.2017.01.016.

Hess et al 9

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

References 1. World Health Organization. The 10 leading causes of death in the world, 2000 and 2012. Fact Sheet No. 310. Updated May 2014. Available from, http://www.who.int/mediacentre/factsheets/ fs310/en/. 2. The task force on the management of stable coronary artery disease of the European Society of Cardiology. 2013 ESC guidelines on the management of stable coronary arteryEur Heart J 2013;34(38): 2949-3003. 3. Fihn SD, Blankenship JC, Alexander KP, et al. 2014 ACC/AHA/AATS/PCNA/SCAI/STS focused update of the guideline for the diagnosis and management of patients with stable ischemic heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines, and the American Association for Thoracic Surgery, Preventive Cardiovascular Nurses Association, Society for Cardiovascular Angiography and Interventions, and Society of Thoracic Surgeons. Circulation 2014;1309:1749-67. 4. European Stroke Organisation (ESO) Executive Committee, ESO Writing Committee. Guidelines for management of ischaemic stroke and transient ischaemic attack 2008. Cerebrovasc Dis 2008;25: 457-507. 5. Kernan WN, Ovbiagele B, Black HR, et al. Guidelines for the prevention of stroke in patients with stroke and transient ischemic

16. 17.

18. 19. 20.

21.

22.

23.

attack: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke 2014;45: 2160-236. Low-molecular-weight heparin during instability in coronary artery disease, Fragmin during Instability in Coronary Artery Disease (FRISC) study groupLancet 1996;347:561-8. Inhibition of platelet glycoprotein IIb/IIIa with eptifibatide in patients with acute coronary syndromes. The PURSUIT Trial Investigators. Platelet Glycoprotein IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin TherapyN Engl J Med 1998;339:436-43. A comparison of aspirin plus tirofiban with aspirin plus heparin for unstable angina. Platelet Receptor Inhibition in Ischemic Syndrome Management (PRISM) Study InvestigatorsN Engl J Med 1998;338: 1498-505. International, randomized, controlled trial of lamifiban (a platelet glycoprotein IIb/IIIa inhibitor), heparin, or both in unstable angina. The PARAGON Investigators. Platelet IIb/IIIa Antagonism for the Reduction of Acute coronary syndrome events in a Global Organization NetworkCirculation 1998;97:2386-95. CAPRIE Steering Committee. A randomised, blinded, trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). CAPRIE Steering Committee. Lancet 1996;348:1329-39. Yusuf S, Zhao F, Mehta SR, et al. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494-502. Wallentin L, Becker RC, Budaj A, et al. Ticagrelor versus clopidogrel in patients with acute coronary syndromes. N Engl J Med 2009;361: 1045-57. Alexander JH, Lopes RD, James S, et al. Apixaban with antiplatelet therapy after acute coronary syndrome. N Engl J Med 2011;365: 699-708. Tricoci P, Huang Z, Held C, et al. Thrombin-receptor antagonist vorapaxar in acute coronary syndromes. N Engl J Med 2012;366: 20-33. Roe MT, Armstrong PW, Fox KA, et al. Prasugrel versus clopidogrel for acute coronary syndromes without revascularization. N Engl J Med 2012;367:1297-309. Mega JL, Braunwald E, Wiviott SD, et al. Rivaroxaban in patients with a recent acute coronary syndrome. N Engl J Med 2012;366:9-19. Wiviott SD, Braunwald E, McCabe CH, et al. Prasugrel versus clopidogrel in patients with acute coronary syndromes. N Engl J Med 2007;357:2001-15. Thygesen K, Alpert JS, White HD. Universal definition of myocardial infarction. Circulation 2007;116:2634-53. Lunn M, McNeil D. Applying Cox regression to competing risks. Biometrics 1995;51:524-32. Budaj A, Flasinska K, Gore JM, et al. Magnitude of and risk factors for in-hospital and postdischarge stroke in patients with acute coronary syndromes: findings from a Global Registry of Acute Coronary Events. Circulation 2005;111:3242-7. Genereux P, Bernard S, Palmerini T, et al. Incidence, predictors, and impact of neurological events in non–ST-segment elevation acute coronary syndromes: the ACUITY trial. EuroIntervention 2015;11:399-406. Stone SG, Serrao GW, Mehran R, et al. Incidence, predictors, and implications of reinfarction after primary percutaneous coronary intervention in ST-segment-elevation myocardial infarction: the Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction Trial. Circ Cardiovasc Interv 2014;7:543-51. White HD, Huang Z, Tricoci P, et al. Reduction in overall occurrences of ischemic events with vorapaxar: results from TRACER. J Am Heart Assoc 2014;3. [pii:e001032].

American Heart Journal Month Year

10 Hess et al

24. Murphy SA, Antman EM, Wiviott SD, et al. Reduction in recurrent cardiovascular events with prasugrel compared with clopidogrel in patients with acute coronary syndromes from the TRITON-TIMI 38 trial. Eur Heart J 2008;29:2473-9. 25. Murphy SA, Cannon CP, Blazing MA, et al. Reduction in total cardiovascular events with ezetimibe/simvastatin post–acute coronary syndrome. J Am Coll Cardiol 2016; 67:353-61. 26. Mahaffey KW, Hager R, Wojdyla D, et al. Meta-analysis of intracranial hemorrhage in acute coronary syndromes: incidence, predictors, and clinical outcomes. J Am Heart Assoc 2015;4: e001512. 27. Roffi M, Patrono C, Collet JP, et al. 2015 ESC guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: Task Force for the Management of Acute Coronary Syndromes in Patients Presenting

28.

29.

30.

31.

without Persistent ST-Segment Elevation of the European Society of Cardiology (ESC). Eur Heart J 2016;37:267-315. The Task Force on the management of ST-segment elevation acute myocardial infarction of the European Society of Cardiology (ESC), Steg PG, James SK, et al. ESC guidelines for the management of acute myocardial infarction in patients presenting with ST-segment elevation. Eur Heart J 2012;33:2569-619. Bakal JA, Westerhout CM, Armstrong PW. Impact of weighted composite compared to traditional composite endpoints for the design of randomized controlled trials. Stat Methods Med Res 2015;24:980-8. Stolker JM, Spertus JA, Cohen DJ, et al. Rethinking composite end points in clinical trials: insights from patients and trialists. Circulation 2014;130:1254-61. Ciolino JD, Carter RE. Reanalysis or redefinition of the hypothesis? Eur Heart J 2015;36:340-1.