Figure 1. Study selection according to Preferred Reporting Items for Systematic Reviews and Meta-analyses.
Figure 2. Effects of the use of new-generation oral anticoagulant agents in patients receiving antiplatelet therapy after an acute coronary syndrome. A, Thrombolysis in myocardial infarction (TIMI) major bleeding events. B, Major and clinically relevant nonmajor bleeding events. C, Any bleeding event. The pooled odds ratios show significant increases in all 3 bleeding categories. Diamond indicates overall summary estimate for the analysis (width of the diamond represents the 95% CI); width of the shaded square, size of the population.
Figure 3. Effects of the use of new-generation oral anticoagulant agents in patients receiving antiplatelet therapy after an acute coronary syndrome. A, Overall mortality. B, Composite ischemic events. C, Definite or probable stent thrombosis. The pooled odds ratios show significant decreases in the risk for composite ischemic events and for definite or probable stent thrombosis, with no effect on overall mortality. Diamond indicates overall summary estimate for the analysis (width of the diamond represents the 95% CI); width of the shaded square, size of the population.
Figure 4. Effects of the use of different types and dosages of new-generation oral anticoagulant agents on the net clinical benefit in patients receiving antiplatelet therapy after an acute coronary syndrome. The forest plots show the net clinical benefit, calculated as the sum of composite ischemic events and major bleeding events, according to the tested daily doses (white background) and the pooled odds ratios for the trials with different drugs (gray background). P values represent the fixed-effects meta-regression analysis results for differences in effect sizes, with daily doses of the drug as the predictor. DTI indicates direct thrombin inhibitor; vertical bar, effect estimate for the analysis; horizontal line, 95% CI.
Komócsi A, Vorobcsuk A, Kehl D, Aradi D. Use of new-generation oral anticoagulant agents in patients receiving antiplatelet therapy after an acute coronary syndrome: systematic review and meta-analysis of randomized controlled trials. Arch Intern Med. Published online September 24, 2012. doi:10.1001/archinternmed.2012.4026
eAppendix. PubMed Search Syntax
eTable. Results of Fixed-Effects Meta-Regression Analyses of Trial-Level Characteristics and the Trial Outcomes
eFigure 1. Effect of New-Generation Oral Anticoagulants on the Frequency of Myocardial Infarction
eFigure 2. Funnel Plot Showing Effect Estimates as a Function of Standard Error Regarding Net Clinical Benefit
András Komócsi, András Vorobcsuk, Dániel Kehl, Dániel Aradi. Use of New-Generation Oral Anticoagulant Agents in Patients Receiving Antiplatelet Therapy After an Acute Coronary SyndromeSystematic Review and Meta-analysis of Randomized Controlled Trials. Arch Intern Med. 2012;172(20):1537–1545. doi:10.1001/archinternmed.2012.4026
Author Affiliations: Department of Interventional Cardiology, Heart Institute (Drs Komócsi, Vorobcsuk, and Aradi), and Department of Statistics and Econometrics, Institute of Applied Studies in Business and Economics, Faculty of Business and Economics (Mr Kehl), University of Pécs, Pécs, Hungary.
Background Despite receipt of dual antiplatelet therapy, patients after an acute coronary syndrome (ACS) remain at significant risk for thrombotic events. The role of orally activated Xa antagonist (anti-Xa) and direct thrombin inhibitors is debated in this setting. Our study objective was to evaluate the efficacy and safety of new-generation oral anticoagulant agents compared with placebo in patients receiving antiplatelet therapy after an ACS.
Methods Electronic databases were searched to identify prospective randomized placebo-controlled clinical trials that evaluated the effects of anti-Xa or direct thrombin inhibitors in patients receiving antiplatelet therapy after an ACS. Efficacy measures included stent thrombosis, overall mortality, and a composite end point of major ischemic events, while thrombolysis in myocardial infarction–defined major bleeding events were used as a safety end point. The net clinical benefit was calculated as the sum of composite ischemic events and major bleeding events.
Results For the period January 1, 2000, through December 31, 2011, we identified 7 prospective randomized placebo-controlled clinical trials that met the study criteria, involving 31 286 patients. Based on the pooled results, the use of new-generation oral anticoagulant agents in patients receiving antiplatelet therapy after an ACS was associated with a dramatic increase in major bleeding events (odds ratio, 3.03; 95% CI, 2.20-4.16; P < .001). Significant but moderate reductions in the risk for stent thrombosis or composite ischemic events were observed, without a significant effect on overall mortality. For the net clinical benefit, treatment with new-generation oral anticoagulant agents provided no advantage over placebo (odds ratio, 0.98; 95% CI, 0.90-1.06; P = .57).
Conclusion The use of anti-Xa or direct thrombin inhibitors is associated with a dramatic increase in major bleeding events, which might offset all ischemic benefits in patients receiving antiplatelet therapy after an ACS.
In the era of dual antiplatelet therapy, featuring the use of aspirin and adenosine diphosphate (ADP) receptor antagonists, patients after an acute coronary syndrome (ACS) remain at significant risk for recurrent thrombotic events.1 According to contemporary trials,2,3 the rate of occurrence of stroke, cardiovascular death, or myocardial infarction (MI) at 1 year exceeds 11% among patients with an ACS who take a second-generation thienopyridine in combination with aspirin. Because of many limitations associated with the use of clopidogrel bisulfate,4 antiplatelet agents were developed with quicker, more potent, and less variable P2Y12-receptor inhibition. The use of prasugrel and ticagrelor demonstrated that these benefits translated into better clinical outcomes, and new-generation ADP receptor antagonists became the standard of care in patients after ACS.2,3
However, the occurrence of thrombotic events after an ACS is not completely platelet dependent, and previous experience with the use of warfarin sodium suggested an incremental ischemic benefit when anticoagulant therapy was used in combination with aspirin.5 However, the clinical relevance and thrombotic benefits of such an approach are offset by the frequent drug and food interactions, the high rates of bleeding complications, and the narrow therapeutic window with the need for careful monitoring.
In recent years, specifically acting oral anticoagulant agents have been developed, providing a more reliable effect, without the need for laboratory monitoring. Compared with the use of warfarin or low-molecular-weight heparins, these agents showed improved net clinical outcomes when used in patients experiencing atrial fibrillation or undergoing orthopedic surgery.6- 9 These results set the stage for the possibility of using these agents as an adjunct to antiplatelet therapy in an attempt to reduce thrombotic complications in patients after an ACS. Numerous prospective randomized placebo-controlled trials have been performed to evaluate the benefits of such an approach in patients after an ACS; however, the results are heterogeneous, and studies were commonly underpowered for separate outcomes. Therefore, we aimed to perform a systematic review and meta-analysis to gain more statistical power to study the clinical efficacy and safety of new-generation oral anticoagulant agents (anti-Xa or direct thrombin inhibitors [DTIs]) compared with placebo in patients receiving antiplatelet therapy after an ACS.
We searched for prospective randomized placebo-controlled clinical trials that evaluated the clinical efficacy and safety of an anticoagulant protocol that included oral anti-Xa and DTI in patients receiving antiplatelet therapy after an ACS. Searches were performed using electronic databases (PubMed, www.clinicaltrials.gov, and SciVerse Scopus) for relevant studies published between January 1, 2000, and December 31, 2011. We also performed manual searches of the reference lists of studies, reviews, editorials, and letters, as well as related conference proceedings. No language restriction was used. The search keywords included the following terms: acute coronary syndrome, oral anticoagulation, apixaban OR edoxaban OR darexaban OR rivaroxaban OR otamixaban OR YM466, and dabigatran OR argatroban OR ximelagatran. Details of the PubMed search syntax are given in the eAppendix.
Excluded were investigations of vitamin K antagonists, trials of parenteral anticoagulant agents, and studies that evaluated the clinical effects of more potent antiplatelet protocols (higher dosages or third agents and new-generation P2Y12 ADP-receptor or protease-activated receptor 1 [PAR-1] antagonists). We also excluded studies that aimed to compare only the biological efficacy of anticoagulant protocols.
Article selection and data abstraction were performed independently by 2 of us (A.K. and A.V.). Disagreement was resolved by consensus of all authors. The following details were recorded for each study: author, study name, exclusion criteria, year of publication, phase and design of the study, and total daily dose of the oral anticoagulant agents used, as well as the number of patients and their sex and mean age at baseline. Also extracted were patient characteristics, including the presence of diabetes mellitus and their clinical presentation (rates of unstable angina pectoris, ST-segment elevation MI, and non–ST-segment elevation MI), as well as data on concomitant antiplatelet medication, the duration of the follow-up period, and the rate of trial drug discontinuation.
We recorded data about cardiovascular end points and bleeding event frequencies. Efficacy end points in the analysis included overall mortality, definite or probable stent thrombosis, and the composite end point of death, MI, ischemic stroke, or severe recurrent ischemia. The net clinical benefit was calculated as the sum of composite ischemic events and thrombolysis in MI (TIMI) major bleeding events. Data on overall mortality and the composite efficacy end point were used according to each study definition, while definite or probable stent thrombosis was defined according to criteria by the Academic Research Consortium.10 The main safety end point was TIMI major bleeding events. We also recorded the rates of any bleeding event and of major and clinically relevant nonmajor bleeding events according to the definitions used in each trial.
Statistical analysis was performed using available software (RevMan 5.1.6; Cochrane Information Management System). Our study adhered to Meta-analysis of Observational Studies in Epidemiology guidelines.11
Considering that the included trials studied different drugs and different treatment effects, we prespecified the use of a random-effects model to assess effect sizes. The reported event frequencies were used to calculate odds ratios (95% CIs). Log odds ratios were pooled with inverse variance weighting. To decrease the effect of the inclusion of small studies, a sensitivity analysis with the fixed-effects model was also performed. The following prespecified subgroup analyses were also performed: anti-Xa and DTI trials, phase 2 and phase 3 trials, and completed trials and prematurely discontinued trials. The degree of inconsistency across studies was quantified using the I2 statistic. The Cochrane Q heterogeneity test (χ2 test) was also applied. These data are reported as I2 percentages, along with P values from the χ2 test. The likelihood of publication bias was assessed visually by generating a funnel plot for the composite end point and mathematically by using the Egger test. P ≤ .05 was considered statistically significant.
Seven prospective randomized placebo-controlled clinical trials12- 18 involving 31 286 patients (range, 1279-15 526 patients per trial) were included in the analysis (Figure 1). The primary characteristics of the 7 included trials are listed in Table 1 and Table 2. Trial patients were admitted to the hospital because of unstable angina pectoris, ST-segment elevation MI, or non–ST-segment elevation MI, and enrollment in the trial was completed within 1 week of admission in anti-Xa trials and within 2 weeks of admission in DTI trials.
The selected trials excluded patients with severe cardiac, renal, or liver insufficiency and patients with excessive risk for bleeding, including those requiring long-term oral anticoagulation. Trial definitions were heterogeneous for prior stroke (Table 1), and the aspirin dosages ranged from 75 to 325 mg/d. All but 1 trial included patients receiving dual antiplatelet therapy, and the frequency of administration ranged from 0.0% to 96.6% across trials. The rates of percutaneous coronary intervention also varied considerably, ranging from 8.3% to 74.6% across trials. The 20-mg arms of the Apixaban for Prevention of Acute Ischemic and Safety Events (APPRAISE) trial,14 APPRAISE-2 trial,15 and APPRAISE-JAPAN trial (http://www.clinicaltrials.gov/ ct2/show/NCT00852397?term=NCT00852397&rank=1) were stopped prematurely because of a lack of observed benefit from the trial drugs. Because no data were available on the latter trial, it was excluded from the analysis.
Duration of the follow-up periods ranged from 3 months to 5 years, with a total of 52 936 patient-years, and few patients were lost to follow-up analysis. The rate of trial drug discontinuation was high, occurring in more than one-fourth of patients (mean, 26.4%; range, 12.0%-44.0%) in the active treatment arms (Table 2).
Based on the pooled results, the use of new-generation oral anticoagulant agents in patients receiving antiplatelet therapy after an ACS was associated with a dramatic increase in major bleeding events, with 2- to 3-fold higher odds based on the definitions of bleeding events used in the trials (Figure 2). In contrast, significant but moderate reductions in the risk for definite or probable stent thrombosis or for composite ischemic events were observed, without a significant effect on overall mortality (Figure 3). The use of new-generation oral anticoagulant agents significantly reduced the risk for MI (eFigure 1). When both composite ischemic events and major bleeding events were taken into account, the use of new-generation oral anticoagulant agents showed no difference in the net clinical benefit (Figure 4).
The observed effects were consistent among most evaluated outcomes except for the heterogeneity across composite ischemic events and in the case of major and clinically relevant nonmajor bleeding events (Figures 2 and 3). Clinical outcomes showed no association with possible trial-level confounders (eTable). Subgroup and sensitivity analyses showed results consistent with those of the overall analysis (Table 3).
Types and dosages of various oral anticoagulant agents were further analyzed to evaluate possible interactions with the net clinical benefit. No significant interaction was found according to the type of anticoagulant or the dosing (Figure 4).
To analyze the likelihood of publication bias, funnel plots were constructed. No analysis demonstrated a skewed distribution pattern, suggesting a lack of publication bias in the study. The Egger test results confirmed the visual evaluation of the plots (eFigure 2).
This meta-analysis involving 31 286 patients provides a comprehensive assessment of the risk-benefit profile for the use of new-generation oral anticoagulant agents in patients receiving antiplatelet therapy after an ACS. Our main finding is that the ischemic benefits achieved with the use of these new agents were offset by the higher risk for major bleeding event. The increased risk for bleeding may explain the lack of significant benefit in overall mortality, despite the lower rate of ischemic events. Because subgroup analyses revealed no meaningful interaction for the evaluated outcomes, the results herein refute the unrestricted use of new-generation oral anticoagulant agents in combination with antiplatelet therapy among patients after an ACS. The present results are in agreement with multiple sources of evidence (discussed herein) that suggest an increased risk for bleeding associated with long-term use of oral anticoagulation in patients receiving dual antiplatelet therapy after an ACS.
In the past decade, a similar risk-benefit profile was observed with the use of warfarin in combination with aspirin.5 In that meta-analysis, therapeutic-range anticoagulation with warfarin (international normalized ratio, 2-3) was associated with a significant reduction in ischemic events (odds ratio, 0.73; 95% CI, 0.63-0.84; P < .001) but triggered an increased risk for major bleeding. Because of the higher risk for bleeding and thrombotic events compared with dual antiplatelet therapy, long-term use of oral anticoagulation after an ACS has been restricted to the prevention of thromboembolic complications in special subsets of patients after an ACS (eg, those with atrial fibrillation or after pulmonary embolism). In patients experiencing atrial fibrillation or undergoing orthopedic surgery, new-generation oral anticoagulant agents successfully reduced bleeding complications, and the use of some agents (eg, apixaban and high-dose dabigatran) was associated with a decreased risk for thromboembolic events compared with warfarin or low-molecular-weight heparins.6- 9 However, our results suggest that the excess risk for bleeding associated with the use of novel anti-Xa agents and DTIs is prohibitively high among patients receiving dual antiplatelet therapy after an ACS. Beyond its prognostic significance, the occurrence of bleeding is also an important element of compliance. Greater frequency of major bleeding events may have been the primary impetus for the high discontinuation rates of anti-Xa agents and DTIs across the trials analyzed herein.
Most important, the observations herein among patients receiving antiplatelet therapy are not novel: the introductory use of clopidogrel and prasugrel was also accompanied by an increased risk for major bleeding complications unrelated to a coronary artery bypass graft (CABG).3,19 While the relative increase in the annual rate of TIMI non-CABG bleeding events was approximately 25% with the introduction of new P2Y12 ADP-receptor antagonists,2,3 we found 2- to 3-fold higher odds for bleeding events associated with the use of new-generation oral anticoagulant agents. Because the prognostic significance of bleeding complications is as serious as the occurrence of ischemic events after percutaneous coronary intervention,20,21 the 200% relative risk increase in major bleeding complications observed herein seems unacceptable. Moreover, the introductory use of the novel potent antiplatelet agents (clopidogrel, prasugrel, and ticagrelor) was associated with a significantly improved net clinical profile compared with the standard treatment of patients after an ACS. For example, the administration of clopidogrel prevented 21 major ischemic events at the cost of 10 major bleeding episodes per 1000 patients treated in the Clopidogrel in Unstable Angina to Prevent Recurrent Events trial.19 Similarly, the use of ticagrelor prevented 19 ischemic events at the cost of 6 non-CABG bleeding events in the Platelet Inhibition and Patient Outcomes trial,2 while the administration of prasugrel prevented 22 ischemic events at the cost of 6 major non-CABG bleeding events in the Trial to Assess Improvement in Therapeutic Outcomes by Optimizing Platelet Inhibition With Prasugrel.3 The findings of our meta-analysis suggest that the administration of new-generation oral anticoagulant agents in patients receiving dual antiplatelet therapy after an ACS can prevent 13 major ischemic events at the cost of 9 TIMI major bleeding episodes. Lack of net clinical benefit was supported by the neutral odds ratio (0.98; 95% CI, 0.90-1.06; P = .57) for the overall analysis (Figure 4). Similar results were found in 2 large phase 3 prospective randomized placebo-controlled clinical trials in our meta-analysis: in the Anti-Xa Therapy to Lower Cardiovascular Events in Addition to Aspirin With or Without Thienopyridine in Subjects With Acute Coronary Syndromes trial,13 the administration of anti-Xa agents prevented 18 ischemic events at the cost of 15 major bleeding events, while in the APPRAISE-2 trial,15 the use of apixaban prevented 4 ischemic events at the cost of 15 major bleeding episodes. Therefore, on the basis of the available evidence, we conclude that the extremely high risk for bleeding complications offsets any ischemic benefits associated with the administration of new-generation oral anticoagulant agents to patients receiving antiplatelet therapy after an ACS, resulting in a neutral net clinical benefit. These findings are somewhat similar to the effect of adding warfarin in combination with antiplatelet agents, which has never been the treatment of choice for patients after an ACS.5
The excess risk for bleeding associated with the use of new-generation oral anticoagulant agents in patients receiving dual antiplatelet therapy after an ACS may also explain why the overall mortality remained unaffected in our analysis.
Acute coronary syndrome trials6- 8,23 that showed a mortality benefit usually demonstrated equivalent or less strikingly increased rates of bleeding complications compared with their respective controls than was observed herein with new-generation anticoagulants vs placebo. Moreover, dose-dependent relationships seem to exist for the efficacy and safety profiles of each anticoagulant. However, when the net clinical benefit was evaluated for the types and dosages of various oral anticoagulant agents, none of the tested drugs showed superior results compared with placebo. These results suggest that the unrestricted use of new-generation oral anticoagulant agents as an adjunct to dual antiplatelet therapy after an ACS cannot be recommended. Similar to the use of PAR-1 antagonists,24 the unrestricted use of triple antithrombotic therapy in patients after an ACS may not be the right way forward. However, limited data suggest that such potent strategies may be beneficial in high-risk subsets of patients after an ACS.25,26
Notably, a substantial proportion (6%-21%) of patients after an ACS require long-term oral anticoagulation treatment to prevent thromboembolic complications due to atrial fibrillation, prosthetic heart valves, left ventricular thrombus, or pulmonary embolism.27,28 In patients with such conditions, the current option is to use warfarin, which is associated with a significantly increased risk for major bleeding when used in combination with dual antiplatelet therapy.29 The value of using new-generation oral anticoagulant agents in this subset is unknown because such patients were generally excluded a priori from the analyzed randomized clinical trials. The results of subgroup analysis from the Efficacy and Safety of the Oral Direct Thrombin Inhibitor Ximelagatran in Patients With Recent Myocardial Damage trial12 suggest that patients who developed atrial fibrillation after enrollment had a significantly higher risk for ischemic events; however, ximelagatran proved to be highly effective in reducing vascular events in this special subset of patients.30
Beyond the higher risk for bleeding, ongoing concern exists about an increased risk for MI associated with the use of dabigatran compared with warfarin. This warning was expressed in the initial findings of the Randomized Evaluation of Long-term Anticoagulation Therapy trial.6 The results of a recently repeated analyses suggest a lack of significant difference between the 2 drugs31,32; however, these findings contrast with the results of a meta-analysis33 among 7 trials of dabigatran, which shows a significantly higher risk for MI compared with warfarin. In patients receiving dual antiplatelet therapy after an ACS, our results indicate that the use of new-generation oral anticoagulant agents is effective in reducing MI compared with placebo. However, the MI rates were higher with the use of dabigatran and darexaban, raising concerns about whether anti-Xa agents and DTIs can provide the same benefit. Notably, the Canadian Cardiovascular Society guidelines34 on atrial fibrillation recommend the use of warfarin instead of dabigatran in patients at high risk for coronary events. However, specific recommendations on this topic seem premature because no trial to date has compared the use of dabigatran vs warfarin in patients with atrial fibrillation after an ACS.
Our meta-analysis has some limitations. First, several of the trials analyzed herein had premature trial drug terminations and short durations of the follow-up periods. Therefore, the effect of treatment duration on the observed risks and benefits is not well established by these trials. Second, we evaluated different agents that were used in various dosages across the trials. Because the efficacy and safety profiles of these agents are dose dependent, the variability in dosages might be an important source of distortion in the observed point estimates. Third, the definitions of composite outcomes for ischemic and bleeding events differed across trials. This may explain some of the observed heterogeneity for these end points.
In conclusion, the use of anti-Xa or direct thrombin inhibitors is associated with a dramatic increase in major bleeding events among patients receiving antiplatelet therapy after an ACS. The significant but moderate reductions in ischemic events observed in our meta-analysis were offset by the higher bleeding risk, resulting in a neutral net clinical benefit. However, because the use of new-generation P2Y12 ADP-receptor antagonists may result in greater reductions of ischemic events, with substantially lower risk for bleeding complications, the role of oral anticoagulant agents after an ACS is debatable.
Correspondence: András Komócsi, MD, PhD, Department of Interventional Cardiology, Heart Institute, University of Pécs, 13 Ifjúság útja, H-7624 Pécs, Hungary (firstname.lastname@example.org).
Accepted for Publication: June 11, 2012.
Published Online: September 24, 2012. doi:10.1001/archinternmed.2012.4026
Author Contributions: Drs Komócsi, Vorobcsuk, and Aradi had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Komócsi, Vorobcsuk, Kehl, and Aradi. Acquisition of data: Komócsi, Vorobcsuk, and Aradi. Analysis and interpretation of data: Komócsi, Vorobcsuk, Kehl, and Aradi. Drafting of the manuscript: Komócsi, Vorobcsuk, Kehl, and Aradi. Critical revision of the manuscript for important intellectual content: Komócsi, Kehl, and Aradi. Statistical analysis: Komócsi, Vorobcsuk, Kehl, and Aradi. Study supervision: Komócsi and Aradi.
Conflict of Interest Disclosures: Drs Komócsi and Aradi report receiving lecture fees from DSI/Lilly.
Funding/Support: This research was supported by grant 83464 from the Hungarian Scientific Research Funds (Dr Komócsi).