You are here

Article Text

Article menu
Coronary artery disease
Original article
Characterising and predicting bleeding in high-risk patients with an acute coronary syndrome
  1. Razi Khan1,
  2. Renato D Lopes2,
  3. Megan L Neely2,
  4. Susanna R Stevens2,
  5. Robert A Harrington3,
  6. Rafael Diaz4,
  7. Frank Cools5,
  8. Petr Jansky6,
  9. Gilles Montalescot7,
  10. Dan Atar8,
  11. Jose Lopez-Sendon9,
  12. Marcus Flather10,
  13. Danny Liaw11,
  14. Lars Wallentin12,
  15. John H Alexander2,
  16. Shaun G Goodman13
  17. for the Apixaban for Prevention of Acute Ischemic Safety Events (APPRAISE)-2 Steering Committee and Investigators
  1. 1Montreal Heart Institute, University of Montreal, Montreal, Quebec, Canada
  2. 2Duke Clinical Research Institute, Duke Medicine, Durham, North Carolina, United States
  3. 3Stanford School of Medicine, Stanford, California, USA
  4. 4ECLA Estudios Cardiológicos Latinoamérica, Rosario, Argentina
  5. 5Academisch Ziekenhuis Klina, Brasschaat, Belgium
  6. 6University Hospital Motol, Prague, Czech Republic
  7. 7Institute of Cardiology, Pitié-Salpêtrière University Hospital, Paris, France
  8. 8Oslo University Hospital, and Institute of Clinical Sciences, University of Oslo, Oslo, Norway
  9. 9University Hospital La Paz, Madrid, Spain
  10. 10Clinical Trials and Evaluation Unit, Royal Brompton Hospital and Harefield NHS Foundation Trust, London, UK
  11. 11Bristol-Myers Squibb, Princeton, New Jersey, USA
  12. 12Uppsala Clinical Research Center, Uppsala University, Uppsala, Sweden
  13. 13Terrence Donnelly Heart Centre, St Michael's Hospital, University of Toronto, and the Canadian Heart Research Centre, Toronto, Ontario, Canada
  1. Correspondence to Dr Renato D Lopes, Duke Clinical Research Institute, Duke University Medical Center, 2400 Pratt Street, Room 0311 Terrace Level, Durham NC 27705, USA; renato.lopes{at}duke.edu

Abstract

Objective In the Apixaban for Prevention of Acute Ischemic Events (APPRAISE-2) trial, the use of apixaban, when compared with placebo, in high-risk patients with a recent acute coronary syndrome (ACS) resulted in a significant increase in bleeding without a reduction in ischaemic events. The aim of this analysis was to provide further description of these bleeding events and to determine the baseline characteristics associated with bleeding in high-risk post-ACS patients.

Methods APPRAISE-2 was a multinational clinical trial including 7392 high-risk patients with a recent ACS randomised to apixaban (5 mg twice daily) or placebo. Bleeding including Thrombolysis in Myocardial Infarction (TIMI) major or minor bleeding, International Society on Thrombosis and Haemostasis (ISTH) major or clinically relevant non-major (CRNM) bleeding, and any bleeding were analysed using an on-treatment analysis. Kaplan–Meier curves were plotted to describe the timing of bleeding, and a Cox proportional hazards model was used to identify predictors of ISTH major or CRNM bleeding and any bleeding. Median follow-up was 241 days.

Results The proportion of patients who experienced TIMI major or minor, ISTH major or CRNM, and any bleeding was 1.5%, 2.2% and 13.3%, respectively. The incidence of bleeding was highest in the immediate post-ACS period (0.11 in the first 30 days vs 0.03 after 30 days events per 1 patient-year); however, >60% of major bleeding events occurred >30 days after the end of the index hospitalisation. Gastrointestinal bleeding was the most common cause of major bleeding, accounting for 45.9% of TIMI major or minor and 39.5% of ISTH major or CRNM bleeding events. Independent predictors of ISTH major or CRNM bleeding events included older age, renal dysfunction, dual oral antiplatelet therapy, smoking history, increased white cell count and coronary revascularisation.

Conclusions When compared with placebo, the use of apixaban is associated with an important short-term and long-term risk of bleeding in high-risk post-ACS patients, with gastrointestinal bleeding being the most common source of major bleeding. The baseline predictors of major bleeding appear to be consistent with those identified in lower-risk ACS populations with shorter-term follow-up.

Clinical trial No NCT00831441.

https://doi.org/10.1136/heartjnl-2014-307346

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    In recent years, there has been considerable evidence that in-hospital and short-term bleeding events after acute coronary syndrome (ACS) portend worsened prognosis.1–4 Much less is known about the characteristics and predictors of long-term bleeding after ACS and percutaneous coronary intervention (PCI). In particular, with the addition of newer antiplatelets and anticoagulants to post-ACS therapy,5–7 the identification of patient subsets predisposed to longer-term bleeding outcomes has become increasingly important. While different risk models have been created to assess bleeding risk in ACS,2 ,8 ,9 these typically estimate only in-hospital or relatively short-term bleeding risk.

    In the Apixaban for Prevention of Acute Ischemic Events (APPRAISE-2) trial, high-risk post-ACS patients were randomised to receive the oral factor Xa inhibitor apixaban or placebo on top of aspirin (with or without ADP receptor inhibitor therapy).10 This study provides a large cohort of patients at high-risk for recurrent ischaemic events in which to further define the risk of bleeding and to evaluate baseline characteristics associated with a higher longer-term bleeding risk. The objectives of these analyses were to (1) provide a detailed description of bleeding (incidence, type, timing and location) in high-risk patients with a recent ACS and (2) identify the predictors of longer-term bleeding in these high-risk patients with ACS.

    Methods

    APPRAISE-2 trial design, population and main results

    The design and main results of the APPRAISE-2 trial have been published.10 In brief, APPRAISE-2 was a clinical trial that included 7392 patients with recent ACS from 858 centres and 39 countries. Enrolled patients had at least two high-risk characteristics predisposing them to recurrent ischaemic events, including age >65 years, diabetes mellitus, myocardial infarction within previous 5 years, cerebrovascular accident, peripheral vascular disease, clinical heart failure, LV dysfunction (EF ≤40%), impaired renal function (creatinine clearance <60 mL/min) or absence of revascularisation for indexed ACS prior to randomisation.

    Patients were randomised 1:1 to receive apixaban 5 mg twice daily or placebo. The median follow-up duration was 240 days (IQR 132–352). All included patients gave written consent prior to entry. The study was approved by an institutional review board or a national ethics committee at each participating centre.

    After approximately 7000 patients had been enrolled, the independent Data Monitoring Committee recommended that the trial be stopped due to an excess of clinically important bleeding events, predominantly intracerebral haemorrhage (ICH), with apixaban without a counterbalancing reduction in ischaemic events.

    Outcome definitions

    The primary safety endpoint of the APPRAISE-2 trial was major bleeding, defined by the Thrombolysis in Myocardial Infarction (TIMI) bleeding criteria.11 A composite of TIMI major and minor bleeding was a secondary prespecified outcome. TIMI major bleeding included any ICH or reductions in haemoglobin ≥50 g/L or haematocrit >15%. TIMI minor bleeding was defined by the presence of observed blood loss and a drop in haemoglobin of 30–50 g/L or haematocrit of 10%–15%; spontaneous gross haematuria or haematemesis (>120 mL), even if the haemoglobin or haematocrit drop was <30 g/L or <10%; or an unobserved loss of ≥40 g/L in haemoglobin or ≥12% in haematocrit. Additional safety outcomes included a composite of major and clinically relevant non-major (CRNM) bleeding, defined by the International Society on Thrombosis and Haemostasis (ISTH) bleeding criteria. Major ISTH bleeding was defined as a fall in haemoglobin of ≥20 g/L or leading to transfusion of ≥2 units of packed red blood cells, symptomatic bleeding in a critical organ, or a fatal bleed.12 CRNM bleeding was defined as a clinically relevant bleeding event that did not meet the criteria for a major bleed but led to hospital admission, physician-guided medical or surgical therapy or change in antithrombotic therapy. Finally, any bleeding was also included as an endpoint.

    Statistical analysis

    The location(s) of each bleeding event was reported for all patients and then separately by treatment groups for TIMI major or minor bleeding, ISTH major or CRNM bleeding, and for any bleeding. In addition, Kaplan–Meier curves were plotted by treatment for both ‘short-term’ (≤30 days) and ‘long-term’ (0 days–end of study) follow-up to describe the timing of each bleeding event. Baseline patient characteristics were reported among patients with no bleeding events and among those with any, TIMI major or minor, or ISTH major or CRNM bleeding events. To identify independent predictors of bleeding events in this population, a Cox proportional hazards model for time-to-first-event was built for any bleeding and for ISTH major or CRNM bleeding. Variable selection was used to identify potential predictors from an a priori selected list of baseline characteristics (see online supplementary data). As the frequency of bleeding was previously established as being higher in the apixaban (vs placebo) group,10 we sought to identify additional predictors of bleeding in this high-risk ACS population. Therefore, study treatment was included as a stratification variable when selecting predictors in order to account for the established difference in bleeding risk among the treatment arms. Variable selection was performed using the Fast FSR procedure,13 which tunes traditional forward selection by controlling the false selection rate (FSR) of uninformative covariates at a prespecified level; here, set to be 0.05. The FSR was defined as the proportion of uninformative covariates selected into the model out of the total number of covariates selected. Traditional forward selection was performed by setting the alpha-to-enter at a prespecified level, typically taken to be 0.05, but no control on an error rate can be guaranteed in finite samples when using this tuning criterion which can translate into non-parsimonious models with covariates that are not truly predictive of the outcome.

    Results

    Rate, type, timing and location of bleeding events

    In the overall APPRAISE-2 cohort, the proportion of patients who experienced TIMI major or minor bleeding, ISTH major or CRNM, and any bleeding was 1.5%, 2.2% and 13.3%, respectively. Packed red blood cell transfusion was required for 3.0% of patients. The incidence of major and any bleeding events was highest in the immediate post-ACS period (≤30 days). During this ‘short-term’ period, ISTH, TIMI and any bleeding event rates were 0.11/patient-year, 0.08/patient-year and 0.88/patient-year, respectively (table 1). In contrast, rates of ‘late’ bleeding, occurring between 30 days post-indexed ACS event and trial termination, for ISTH, TIMI and any bleeding events were 0.03/patient-year, 0.02/patient-year and 0.14/patient-year, respectively. However, overall >60% of major bleeding events, characterised by TIMI major or minor and ISTH major or CRNM criteria, and >50% of any bleeding events occurred >30 days after ACS presentation. Treatment with apixaban was associated with significantly higher rates of TIMI major or minor, ISTH major or CRNM, and any bleeding events in both short-term and long-term follow-up (figures 1⇓–3).

    View this table:
    Table 1

    Frequency and timing of bleeding events

    Figure 1
    Figure 1

    Kaplan–Meier curves for (A) short-term and (B) long-term International Society on Thrombosis and Haemostasis (ISTH) major/ clinically relevant non-major bleeding events. Short-term: ≤30 days. Long-term: 0 days–end of study.

    Figure 2
    Figure 2

    Kaplan–Meier curves for (A) short-term and (B) long-term Thrombolysis in Myocardial Infarction (TIMI) major/minor bleeding events. Short-term: ≤30 days. Long-term: 0 days–end of study.

    Figure 3
    Figure 3

    Kaplan–Meier curves for (A) short-term and (B) long-term any bleeding events. Short-term: ≤30 days. Long-term: 0 days–end of study.

    In both the overall APPRAISE-2 cohort and in the apixaban arm, gastrointestinal bleeding was the most common cause of both TIMI major or minor and ISTH major or CRNM bleeding (table 2), accounting for 45.9% of TIMI major or minor and 39.5% of ISTH major or CRNM bleeding in the overall cohort, and 51.3% of TIMI major or minor and 43.6% of ISTH major or CRNM bleeds in the apixaban group, but only 31.0% of TIMI major or minor bleeds and 28.9% of ISTH major or CRNM bleeds in the placebo group. Unclassified (‘other’) bleeding events were the most common type of all bleeding events, with an incidence of 25.4% in the combined cohort. Epistaxis, bruising, gastrointestinal bleeding and haematomas were the most common types of any bleeding events in both treatment arms. The inclusion of occult bleeding events in the unclassified group may have resulted in underestimation of gastrointestinal bleeding, as they are often gastrointestinal in origin.

    View this table:
    Table 2

    Bleeding by type, location and treatment

    There was a fourfold greater number of ICH in the apixaban group compared with the placebo group (12 (0.3%) vs 3 (<0.1%)) (table 2). Although infrequent, fatal bleeding events were more common in the apixaban group (5 (0.1%) vs 0 (0.0%)). Epistaxis, gastrointestinal bleeding and haematuria also appeared to be more frequent in patients receiving apixaban. Finally, patients receiving apixaban required more transfusion compared with those receiving placebo (147 (4.0%) vs 76 (2.1%)).

    Predictors of bleeding outcomes

    Patients with bleeding events tended to be older, were more likely to have had an indexed ACS bleeding event, a history of smoking (current or prior), hypercholesterolaemia, impaired renal function, hypertension and atrial fibrillation (table 3). In addition, patients who experienced bleeding events were more likely to have undergone PCI, received dual oral antiplatelet therapy prior to randomisation and had a biomarker-positive ACS.

    View this table:
    Table 3

    Baseline characteristics by bleeding group*

    In a multivariable model, increasing age, impaired renal function, smoking, an elevated white cell count and use of dual oral antiplatelet therapy identified at the time of the index ACS event were independent predictors of ISTH major or CRNM bleeding (table 4). Patients who underwent coronary revascularisation compared with those medically managed for their index ACS event were also at higher risk of bleeding. Similar characteristics were identified in a multivariable model predicting any bleeding (table 5); in addition, history of bleeding events during index ACS hospitalisation, hypercholesterolaemia, atrial fibrillation, elevated diastolic blood pressure (>80 mm Hg) and treatment with prasugrel were independent predictors of any bleeding.

    View this table:
    Table 4

    Predictive model for ISTH major or CRNM bleeding

    View this table:
    Table 5

    Predictive model for any bleeding

    Discussion

    In this large population of patients post ACS and at high risk for recurrent ischaemic events in whom a majority were receiving dual oral antiplatelet therapy, the addition of the oral anticoagulant apixaban resulted in an increased risk of bleeding when compared with placebo. Bleeding occurred more frequently with apixaban in a variety of locations, including mucosal surfaces (eg, epistaxis, gingival), skin (eg, bruising, haematoma), gastrointestinal, genitourinary (eg, haematuria), intracranial and others. Multivariable models demonstrated that, in addition to study treatment itself, several clinical factors (eg, increasing age, history of smoking, renal dysfunction, elevated white cell count) and aspects of prerandomisation management (eg, use of dual oral antiplatelet therapy, coronary revascularisation) were associated with bleeding events over the course of the trial.

    Bleeding outcomes in APPRAISE-2

    In the overall APPRAISE-2 trial population, the proportion of patients who experienced TIMI major or minor and ISTH major or CRNM bleeding was approximately 2% while any bleeding occurred in approximately 13% of patients, indicating that the majority of bleeding events in the trial were minor. There was a marked increase in the rate of major and any bleeding events shortly (≤30 days) after ACS. Rates of bleeding were at least 50-fold higher in the short compared with ‘late’ post-ACS indexed hospitalisation period. However, the risk of late bleeding persisted after the initial indexed event and bleeding occurring >30 days after ACS accounted for more than half of all TIMI, ISTH and any bleeding events in the study. Treatment with apixaban was associated with significantly higher rates of bleeding events in both short-term and long-term follow-up. Gastrointestinal bleeding was the most common source of major bleeding in APPRAISE-2.

    In APPRAISE-2, use of apixaban was associated with an increase in fatal bleeds and a fourfold greater number of ICH when compared with placebo. While direct comparisons are not available, and the patient populations, concomitant therapies, study drugs and doses, and treatment durations differed, it is interesting to note that risk of bleeding (including fatal bleeding) and ICH in patients receiving apixaban was comparable with the threefold to fourfold increase in bleeding observed with other factor Xa and direct thrombin inhibitors in a large Phase III and several smaller Phase II studies, respectively.7 ,14–16 Further, a previous meta-analysis indicated that the addition of warfarin to aspirin after ACS was associated with an almost twofold increased risk of bleeding when compared with aspirin alone.17 Thus, risk of bleeding in the high ischaemic risk patients with ACS on oral anticoagulation together with single or dual antiplatelet therapy in APPRAISE-2 did not appear to be substantially different from that observed in previous studies of similar or lower ischaemic risk post-ACS populations.

    Predictors of major bleeding in APPRAISE-2

    In APPRAISE-2, patients with any bleeding or ISTH-classified bleeding events tended to be older and were more likely to have additional comorbidities (eg, smoking, hypertension, impaired renal function and atrial fibrillation). In addition, patients experiencing bleeding events were more likely to have an elevated white cell count and have undergone PCI for the indexed event and received dual oral antiplatelet therapy prior to randomisation. Indeed, the APPRAISE-2 trial included higher-risk patients with ACS and provided a longer-term bleeding assessment when compared with previous large, randomised and observational studies where predictors of bleeding were identified.8 ,18 ,19 A number of common predictors of short-term bleeding after ACS and/or PCI have been derived from pooled analysis of clinical trials and from large observational cohorts.8 ,9 ,20 ,21 Common predictors of short-term bleeding included older age, female sex, body weight, renal dysfunction, history of diabetes, baseline anaemia, prior vascular disease, use of oral anticoagulants, presenting blood pressure and congestive heart failure or shock on admission. Thus, characteristics that predispose to bleeding appear to be similar across the spectrum of moderate-risk to high-risk patients with ACS and consistent between ‘real world’ and clinical trial populations. Additionally, characteristics of patients with ACS which predispose to short-term bleeding appear to confer risk for long-term bleeding events as well.

    Both increased age and renal dysfunction have been consistently correlated with higher risk of post-ACS bleeding.8 ,18 ,19 ,22 Age >75 years has been associated with 1.9-fold to 3.3-fold increased risk of non-coronary artery bypass grafting (CABG) major bleeding in patients receiving contemporary ACS treatment.23 ,24 Similarly, impaired renal function has been associated with a more than fourfold increase in non-CABG major bleeding.25 Additionally, the risk of bleeding appears to rise incrementally by 4%–21% for every 5-year increase in age and by 6% for every 10 mL/min reduction in creatinine clearance.9 ,20

    Dual oral antiplatelet therapy has also been shown to increase risk of bleeding after ACS.26 When compared with aspirin alone, there is an estimated 0.6%–1.2% increase in the absolute risk of bleeding with the addition of clopidogrel;26 use of the newer P2Y12 receptor inhibitors prasugrel and ticagrelor, while more effective in reducing cardiovascular events, were also shown to further increase risk of bleeding post ACS when compared with clopidogrel.5 ,6

    Management with an invasive revascularisation strategy for index events has also been shown to increase risk of bleeding after ACS,21 ,27 and post-ACS bleeding was more likely with early revascularisation regardless of the anticoagulation therapy used.27 Early invasive therapy is an independent predictor of major bleeding, with an OR of >1.6 when compared with a conservative treatment strategy for ACS,21 which may be attributable to increased use of anticoagulation and access site bleeding.28

    The effects of prior smoking on bleeding predisposition in patients with ACS and those at high risk of vascular disease have been inconsistent. While some observational studies have shown a lower risk of bleeding among recent smokers receiving clopidogrel, in the APPRAISE-2 cohort, smoking was an independent predictor of increased bleeding risk. Smoking may increase bleeding risk by the induction of cytochrome P450 system enzymes resulting in increased conversion of clopidogrel to its active metabolite, thereby augmenting platelet inhibition.29 Finally, consistent with previous reports, we observed that an elevated white cell count was associated with an increased risk of short-term bleeding.20

    Ischaemic events with combined novel oral anticoagulants and dual antiplatelet therapy in ACS

    In contrast to APPRAISE-2, the Anti-Xa Therapy to Lower Cardiovascular Events in Addition to Standard Therapy in Subjects with Acute Coronary Syndrome–Thrombolysis in Myocardial Infarction 51 (ATLAS ACS 2–TIMI 51) study did find a significant reduction in major adverse cardiovascular events with rivaroxaban superimposed on standard ACS therapy.7 However, it should be noted that the other major clinical trials involving adjunctive treatment of novel oral anticoagulants in this setting, including the Randomized Dabigatran Etexilate Dose-Finding Study in Patients With Acute Coronary Syndromes (REDEEM) with dabigatran16 and the Thrombin Receptor Antagonist for Clinical Event Reduction in Acute Coronary Syndrome (TRACER) with vorapaxar, found increased bleeding with additional therapy without reductions in primary composite ischaemic endpoints.30 It is possible that an increase in major bleeding may have contributed to greater ischaemic events in the treatment arms of these trials resulting in an absence of effect on cardiovascular outcomes observed in these trials.31 Therefore, it is possible that a lower dose of apixaban may have reduced major bleeding events and consequently reduced ischaemic outcomes.

    Limitations to anticoagulation in targeting ischaemia alone

    Although the risk of recurrent ischaemia after ACS is increased in patients with high-risk features, so too is the predisposition towards bleeding events. Major bleeding events after ACS have even been shown to be prognostic for recurrent ischaemia, with rates of future ischaemic events being threefold to fivefold higher after bleeding.3 ,18 ,31 This is unsurprising as risk factors for bleeding identified in this study, including age, smoking, white cell count and renal impairment, have all been associated with increased risk of future ischaemic events.20 Similarly, patient characteristics such as heart rate, systolic blood pressure, congestive heart failure and renal function are included in both the Global Registry of Acute Coronary Events (GRACE) ischaemic and CRUSADE (Can Rapid risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/AHA guidelines) bleeding stratification scores.8 ,32 Considering the substantial overlap in risk factors for bleeding and ischaemia, attempts to reduce ischaemic events in higher-risk patients with increasingly aggressive pharmacological therapies, such as apixaban added to standard therapy including frequent use of dual oral antiplatelet therapy in APPRAISE-2, will likely come at a cost of increased bleeding with resultant prognostic implications.8

    Conclusion

    In the APPRAISE-2 trial, apixaban was associated with a clinically important risk of bleeding after ACS in high-risk patients when compared with placebo. This risk was highest immediately after ACS, but remained elevated throughout the course of the trial. The majority of TIMI and ISTH major bleeding events were gastrointestinal. Treatment with apixaban was associated with an increased risk of major bleeding, ICH and fatal bleeding when compared with placebo, but this bleeding risk appears comparable with that noted in previous trials with other oral anticoagulants and lower-risk populations for bleeding. Predictors of ISTH major or CRNM bleeding events in this patient population included older age, renal impairment, white cell count, smoking history, use of dual oral antiplatelet therapy and coronary revascularisation, matching predictors in lower-risk patients with ACS and determinants of short-term bleeding risk.

    As risk factors for major bleeding have been identified as being similar to those for future ischaemic events, use of adjunctive antithrombotic therapy aimed at reducing ischaemic events in higher-risk patients will also likely be accompanied by increased bleeding, as observed in the APPRAISE-2 trial.

    Key messages

    What is already known on this subject?

    • Bleeding events after acute coronary syndrome (ACS) are associated with worse prognosis.

    What might this study add?

    • The addition of the oral factor Xa inhibitor, apixaban, to standard therapy increased risk of major bleeding after ACS in high-risk patients when compared with placebo. The risk of bleeding was greatest immediately after ACS (first 30 days), but remained high throughout the course of the trial. Baseline characteristics that predispose to bleeding shortly after ACS similarly increase risk of long-term bleeding.

    How might this impact on clinical practice?

    • The addition of antithrombotic therapy to prevent future ischaemic events after ACS will likely be counterbalanced by an increase in bleeding risk. In part, this is attributable to the considerable overlap in baseline risk factors predisposing to both ischaemic and bleeding events.

    References

    1. ↵
      1. Rao SV,
      2. O'Grady K,
      3. Pieper KS, et al
      . Impact of bleeding severity on clinical outcomes among patients with acute coronary syndromes. Am J Cardiol 2005;96:1200–6. doi:10.1016/j.amjcard.2005.06.056
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Mehran R,
      2. Pocock SJ,
      3. Nikolsky E, et al
      . A risk score to predict bleeding in patients with acute coronary syndromes. J Am Coll Cardiol 2010;55:2556–66. doi:10.1016/j.jacc.2009.09.076
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Eikelboom JW,
      2. Mehta SR,
      3. Anand SS, et al
      . Adverse impact of bleeding on prognosis in patients with acute coronary syndromes. Circulation 2006;114:774–82. doi:10.1161/CIRCULATIONAHA.106.612812
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Mehran R,
      2. Pocock SJ,
      3. Stone GW, et al
      . Associations of major bleeding and myocardial infarction with the incidence and timing of mortality in patients presenting with non-ST-elevation acute coronary syndromes: a risk model from the ACUITY trial. Eur Heart J 2009;30:1457–66. doi:10.1093/eurheartj/ehp110
      OpenUrlAbstract/FREE Full Text
    5. ↵
      1. Wiviott SD,
      2. Braunwald E,
      3. McCabe CH, et al
      . Prasugrel versus clopidogrel in patients with acute coronary syndromes. N Engl J Med 2007;357:2001–15. doi:10.1056/NEJMoa0706482
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Wallentin L,
      2. Becker RC,
      3. Budaj A, et al
      . Ticagrelor versus clopidogrel in patients with acute coronary syndromes. N Engl J Med 2009;361:1045–57. doi:10.1056/NEJMoa0904327
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Mega JL,
      2. Braunwald E,
      3. Wiviott SD, et al
      . Rivaroxaban in patients with a recent acute coronary syndrome. N Engl J Med 2012;366:9–19. doi:10.1056/NEJMoa1112277
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Subherwal S,
      2. Bach RG,
      3. Chen AY, et al
      . Baseline risk of major bleeding in non-ST-segment-elevation myocardial infarction: the CRUSADE (Can Rapid risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/AHA Guidelines) Bleeding Score. Circulation 2009;119:1873–82. doi:10.1161/CIRCULATIONAHA.108.828541
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Mathews R,
      2. Peterson ED,
      3. Chen AY, et al
      . In-hospital major bleeding during ST-elevation and non-ST-elevation myocardial infarction care: derivation and validation of a model from the ACTION Registry®-GWTG™. Am J Cardiol 2011;107:1136–43. doi:10.1016/j.amjcard.2010.12.009
      OpenUrlCrossRefPubMed
    10. ↵
      1. Alexander JH,
      2. Lopes RD,
      3. James S, et al
      . Apixaban with antiplatelet therapy after acute coronary syndrome. N Engl J Med 2011;365:699–708. doi:10.1056/NEJMoa1105819
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Chesebro JH,
      2. Knatterud G,
      3. Roberts R, et al
      . Thrombolysis in Myocardial Infarction (TIMI) trial, phase I: a comparison between intravenous tissue plasminogen activator and intravenous streptokinase: clinical findings through hospital dis- charge. Circulation 1987;76:142–54. doi:10.1161/01.CIR.76.1.142
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Schulman S,
      2. Kearon C
      . Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost 2005;3:692–4. doi:10.1111/j.1538-7836.2005.01204.x
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Boos DD,
      2. Stefanski LA,
      3. Wu Y
      . Fast FSR variable selection with applications to clinical trials. Biometrics 2008;65:692–700. doi:10.1111/j.1541-0420.2008.01127.x
      OpenUrlPubMed
    14. ↵
      1. Mega JL,
      2. Braunwald E,
      3. Mohanavelu S, et al
      . Rivaroxaban versus placebo in patients with acute coronary syndromes (ATLAS ACS-TIMI 46): a randomised, double-blind, phase II trial. Lancet 2009;374:29–38. doi:10.1016/S0140-6736(09)60738-8
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Steg PG,
      2. Mehta SR,
      3. Jukema JW, et al
      . RUBY-1: a randomized, double-blind, placebo-controlled trial of the safety and tolerability of the novel oral factor Xa inhibitor darexaban (YM150) following acute coronary syndrome. Eur Heart J 2011;32:2541–54. doi:10.1093/eurheartj/ehr334
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Oldgren J,
      2. Budaj A,
      3. Granger CB, et al
      . Dabigatran vs. placebo in patients with acute coronary syndromes on dual antiplatelet therapy: a randomized, double-blind, phase II trial. Eur Heart J 2011;32:2781–9. doi:10.1093/eurheartj/ehr113
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Andreotti F,
      2. Testa L,
      3. Biondi-Zoccai GG, et al
      . Aspirin plus warfarin compared to aspirin alone after acute coronary syndromes: an updated and comprehensive meta-analysis of 25,307 patients. Eur Heart J 2006;27:519–26. doi:10.1093/eurheartj/ehi485
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Pocock SJ,
      2. Mehran R,
      3. Clayton TC, et al
      . Prognostic modeling of individual patient risk and mortality impact of ischemic and hemorrhagic complications assessment from the acute catheterization and urgent intervention triage strategy trial. Circulation 2010;121:43–51. doi:10.1161/CIRCULATIONAHA.109.878017
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Feit F,
      2. Voeltz MD,
      3. Attubato MJ, et al
      . Predictors and impact of major hemorrhage on mortality following percutaneous coronary intervention from the REPLACE-2 Trial. Am J Cardiol 2007;100:1364–9. doi:10.1016/j.amjcard.2007.06.026
      OpenUrlCrossRefPubMed
    20. ↵
      1. Mehran R,
      2. Pocock S,
      3. Nikolsky E, et al
      . Impact of bleeding on mortality after percutaneous coronary intervention results from a patient-level pooled analysis of the REPLACE-2 (randomized evaluation of PCI linking angiomax to reduced clinical events), ACUITY (acute catheterization and urgent intervention triage strategy), and HORIZONS-AMI (harmonizing outcomes with revascularization and stents in acute myocardial infarction) trials. JACC Cardiovasc Interv 2011;4:654–64. doi:10.1016/j.jcin.2011.02.011
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Moscucci M,
      2. Fox KA,
      3. Cannon CP, et al
      . Predictors of major bleeding in acute coronary syndromes: the Global Registry of Acute Coronary Events (GRACE). Eur Heart J 2003;24:1815–23. doi:10.1016/S0195-668X(03)00485-8
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Lopes RD,
      2. White JA,
      3. Tricoci P, et al
      . Age, treatment, and outcomes in high-risk non-ST-segment elevation acute coronary syndrome patients: insights from the EARLY ACS trial. Int J Cardiol 2013;167:2580–7. doi:10.1016/j.ijcard.2012.06.053
      OpenUrlCrossRefPubMed
    23. ↵
      1. Husted S,
      2. James S,
      3. Becker RC, et al
      . Ticagrelor versus clopidogrel in elderly patients with acute coronary syndromes: a substudy from the prospective randomized PLATelet inhibition and patient Outcomes (PLATO) trial. Circ Cardiovasc Qual Outcomes 2012;5:680–8. doi:10.1161/CIRCOUTCOMES.111.964395
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. Roe MT,
      2. Goodman SG,
      3. Ohman EM, et al
      . Elderly patients with acute coronary syndromes managed without revascularization: insights into the safety of long-term dual antiplatelet therapy with reduced-dose prasugrel versus standard-dose clopidogrel. Circulation 2013;128:823–33. doi:10.1161/CIRCULATIONAHA.113.002303
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Barthélémy O,
      2. Cayla G,
      3. Silvain J, et al
      . Optimal time for catheterization in NSTE-ACS patients with impaired renal function: Insights from the ABOARD Study. Int J Cardiol 2013;167:2646–52. doi:10.1016/j.ijcard.2012.06.126
      OpenUrlCrossRefPubMed
    26. ↵
      1. Yusuf S,
      2. Zhao F,
      3. 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. doi:10.1056/NEJMoa010746
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Kadakia MB,
      2. Desai NR,
      3. Alexander KP, et al
      . Use of anticoagulant agents and risk of bleeding among patients admitted with myocardial infarction: a report from the NCDR ACTION Registry—GWTG (National Cardiovascular Data Registry Acute Coronary Treatment and Intervention Outcomes Network Registry—Get With the Guidelines). JACC Cardiovasc Interv 2010;3:1166–77. doi:10.1016/j.jcin.2010.08.015
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Kinnaird TD,
      2. Stabile E,
      3. Mintz GS, et al
      . Incidence, predictors, and prognostic implications of bleeding and blood transfusion following percutaneous coronary interventions. Am J Cardiol 2003;92:930–5. doi:10.1016/S0002-9149(03)00972-X
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Bliden KP,
      2. Dichiara J,
      3. Lawal L, et al
      . The association of cigarette smoking with enhanced platelet inhibition by clopidogrel. J Am Coll Cardiol 2008;52:531–3. doi:10.1016/j.jacc.2008.04.045
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Tricoci P,
      2. Huang Z,
      3. Held C, et al
      . Thrombin-receptor antagonist vorapaxar in actue coronary syndromes. N Engl J Med 2012;366:20–33. doi:10.1056/NEJMoa1109719
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Manoukian SV,
      2. Feit F,
      3. Mehran R, et al
      . Impact of major bleeding on 30-day mortality and clinical outcomes in patients with acute coronary syndromes: an analysis from the ACUITY trial. J Am Coll Cardiol 2007;49:1362–8. doi:10.1016/j.jacc.2007.02.027
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Granger CB,
      2. Goldberg RJ,
      3. Dabbous OH, et al
      . Predictors of hospital mortality in the global registry of acute coronary events. Arch Intern Med 2003;163:2345–53. doi:10.1001/archinte.163.19.2345
      OpenUrlCrossRefPubMedWeb of Science

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • Contributors All authors made substantial contributions to the conception or design of the work, or the acquisition, analysis or interpretation of data. RK, RDL, MLN, SRS, LW, JHA and SGG drafted the work or revised it critically for important intellectual content. All authors gave final approval of the version published. RK, RDL and SGG agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

    • Funding The APPRAISE-2 study (ClinicalTrials.gov NCT00831441) and this analysis were funded by Bristol-Myers Squibb and Pfizer.

    • Competing interests None declared.

    • Ethics approval Ethics Committees and IRBs at participating centres.

    • Provenance and peer review Not commissioned; externally peer reviewed.