Adjusted Indirect Meta-Analysis of Aspirin Plus Warfarin at International Normalized Ratios 2 to 3 Versus Aspirin Plus Clopidogrel After Acute Coronary Syndromes

Adjusted Indirect Meta-Analysis of Aspirin Plus Warfarin at International Normalized Ratios 2 to 3 Versus Aspirin Plus Clopidogrel After Acute Coronary Syndromes Luca Testa, MDa,*, Giuseppe Biondi Zoccai, MDb, Italo Porto, MDc, Graziana Trotta, MDc, Pierfrancesco Agostoni, MDd, Felicita Andreotti, MDc, and Filippo Crea, MDc After acute coronary syndromes, the beneficial effect of aspirin plus clopidogrel (A ⴙ C) or aspirin plus dose-adjusted warfarin (A ⴙ W) compared with aspirin alone is well established. However, these regimens were never compared. To compare the risk– benefit profile of A ⴙ C versus A ⴙ W after acute coronary syndromes, major medical databases for randomized controlled trials comparing 1 of these combined approaches versus aspirin alone after an acute coronary syndrome (updated June 2006) were searched. Evaluated end points were major adverse events [MAEs: all-cause death, acute myocardial infarction [AMI], thromboembolic stroke, major bleeds, and overall risk of stroke [hemorrhagic or ischemic]). Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for (1) A ⴙ W versus aspirin alone, (2) A ⴙ C versus aspirin alone, and (3) A ⴙ W versus A ⴙ C using adjusted indirect meta-analysis. Thirteen studies were included, totaling 69,741 patients. Ten compared A ⴙ W versus aspirin alone and 3 compared A ⴙ C versus aspirin alone. Each combined approach yielded a significantly lower risk of MAEs, albeit an increased risk of major bleeds, compared with aspirin alone. No significant difference was found for A ⴙ W versus A ⴙ C for risk of overall MAEs, death, or AMI. However, A ⴙ W versus A ⴙ C was associated with a significantly lower risk of thromboembolic stroke (OR 0.53, 95% CI 0.31 to 0.88, number needed to treat 60) and all types of stroke (OR 0.58, 95% CI 0.35 to 0.94, p ⴝ 0.038), but also with increased risk of major bleeds (OR 1.9, 95% CI 1.2 to 2.8, number needed to harm 300). In conclusion, after an acute coronary syndrome, A ⴙ W and A ⴙ C are comparable in the prevention of MAEs, death, and AMI compared with aspirin alone. Allocating 100 patients to A ⴙ W (at international normalized ratio 2 to 3) with respect to A ⴙ C could prevent 17 thromboembolic strokes while causing 3 major bleeds. © 2007 Elsevier Inc. All rights reserved. (Am J Cardiol 2007;99:1637–1642)

Although the pivotal role of aspirin in secondary prevention after an acute coronary syndrome is currently accepted worldwide,1 recent studies showed an additive benefit for both clopidogrel2 and warfarin at international normalized ratio 2 to 33–5 on top of aspirin despite an increased risk of bleeding complications. Of note, no data are available on the direct comparison of these different approaches. Adjusted indirect meta-analyses enable the head-to-head comparison of treatments evaluated in different trials against a common reference treatment by adjustment of results of their direct comparisons to such a common control. The scientific validity of this approach is well established.6,7 We conducted this study aiming to provide an adjusted indirect assessment of the risk– benefit profile of 1 combination versus the other in patients recovering from an acute coronary syndrome.

a Institute of Cardiology, John Radcliffe Hospital, Oxford, United Kingdom; bDivision of Cardiology, University of Turin, Turin; cInstitute of Cardiology, Catholic University, Rome, Italy; and dAntwerp Cardiovascular Institute Middelheim, Antwerp, Belgium. Manuscript received December 24, 2006; revised manuscript received and accepted January 29, 2007. *Corresponding author: Tel.: 39-338-842-0721; fax: 44-018-65220585. E-mail address: [email protected] (L. Testa).

0002-9149/07/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.amjcard.2007.01.052

Methods BioMedCentral, CENTRAL, mRCT, and PubMed were searched without language restrictions (updated to September 2006) according to an established method.8 Pertinent trials were also searched in major recent international cardiology meetings. References of original and review articles, including a recent thorough meta-analysis (to which the reader is referred for extensive search, study abstraction, and validity assessment),5 were cross-checked. Specifically, inclusion criteria were (1) randomized allocation; (2) controlled comparison of aspirin plus warfarin at international normalized ratio 2 to 3 (A ⫹ W) or aspirin plus clopidogrel (A ⫹ C) versus aspirin alone or plus placebo (A ⫹ P) in the setting of acute coronary syndromes, and (3) intention-to-treat analysis. Exclusion criteria were (1) an equivocal treatment allocation process, (2) significant imbalances in major baseline characteristics among study groups, and (3) incomplete (⬍80%) follow-up. End points of interest were (1) the combined rate of major adverse events (MAEs), defined as all-cause death, nonfatal acute myocardial infarction (AMI), or nonfatal thromboembolic stroke; and (2) the combined rate of major bleeds, intra- and extracranial. Additional analyses were carried out for single end points and the overall risk of stroke (ischemic or hemorrhagic). www.AJConline.org

1638

The American Journal of Cardiology (www.AJConline.org)

Review Manager 4.2.5 (available from The Cochrane Collaboration at http://www.cochrane.org/***), OpenOffice. org (Sun Microsystems), and SPSS 11.0 for Windows (SPSS Inc., Chicago, Illinois) were used for analysis. Data regarding the comparison of A ⫹ W versus A ⫹ P (or aspirin alone) were available from a previous publication of our group,5 and no further trials were published or presented to date. Odds ratios (ORs) with 95% confidence intervals (95% CIs) were used as summary statistics. Binary outcomes from individual studies were combined with both random- and fixed-effect models. Z test values were 2 tailed. Hypothesis testing results were considered significant at p ⬍0.05. Cochrane Q heterogeneity tests were also computed and considered significant at p for heterogeneity ⬍0.10.9 As per protocol, we calculated the number needed to treat to prevent an MAE and the number needed to harm by causing an MB. The degree of inconsistency among studies (I2) was estimated according to Higgins et al,10 with scores ⬍25%, 25% to 75%, and ⬎75% representing low, moderate, or high inconsistency, respectively. The internal validity of included trials was appraised according to the Cochrane Collaboration criteria, in other words, judging the risk of selection, performance, attrition, and adjudication biases. Risk of bias was expressed as low (A), moderate (B), or high (C), and incomplete reporting leading to inability to ascertain the underlying risk of bias was scored as D. Allocation concealment was distinguished as adequate (A), unclear (B), inadequate (C), or not used (D).9 Sensitivity analyses were performed excluding trials 1 at time, from those with the lowest to those with the highest quality score, to assess the contribution of each study to the pooled estimates.9 The likelihood of publication or small-study bias was assessed using Egger’s test (p for significant asymmetry ⬍0.1).11 According to established methods, adjusted indirect comparisons with inverse variance weighting of pooled estimates were then performed for MAEs, major bleeds, and single end points. From random-effect ORs comparing A ⫹ W versus A ⫹ P and A ⫹ C versus A ⫹ P, we generated an interaction OR with 95% CI for the different combined treatments and Z score for 2-tailed hypothesis testing (p is significant if ⬍0.05). Specifically, interaction ORs with respective 95% CIs and inconsistency among trials are calculated according to the following formulas: ln (ORA⫹W vs A⫹C) ⫽ ln (ORA⫹W vs A⫹P) ⫺ ln (ORA⫹C vs A⫹P); var [ln (ORA⫹W vs A⫹C)] ⫽ var [ln (ORA⫹W vs A⫹P)] ⫹ var [ln (ORA⫹C vs A⫹P)]; and I2(A⫹W vs A⫹C) ⫽ (chi-square(A⫹W) ⫹ chi-square(A⫹C)) ⫻ 10/degress of freedom(A⫹W) ⫹ degrees of freedom(A⫹C).7,12 Finally, to appraise the magnitude of the population needed for a reliable head-to-head comparison, we computed a post hoc sample size aiming for 80% power and 2-tailed alpha of 0.05. Results Ten studies13–22 compared A ⫹ W at international normalized ratio 2 to 3 versus aspirin alone or A ⫹ P for a total of 7,836 patients.5 Of 234 potentially relevant citations, we finally included 3 studies comparing A ⫹ C versus A ⫹ P for a total of 61,905 patients2,23,24 (Figure 1). Main characteristics of

Figure 1. Flow diagram according to the QUOROM statement for the appraisal of study comparing A ⫹ C versus aspirin alone or A ⫹ P. RCT ⫽ randomized controlled trial. Table 1 Main characteristics of studies comparing aspirin plus dose-adjusted warfarin versus aspirin alone or plus placebo Trial ATACS14ⴱ ATACS15 Williams et al15 OASIS Pilot16 Huyhn et al17 OASIS-2 Warfarin Substudy18 ASPECT 219 APRICOT 220 WARIS 221 Zibaeenezhad et al22

n

Type of ACS

69 214 57 197 90 3,712

UAP, UAP, UAP, UAP, UAP, UAP

NSTEMI NSTEMI AMI NSTEMI NSTEMI

669 274 2414 140

UAP, AMI STEMI AMI AMI

Follow-Up (mo)

Aspirin Dose (mg/d)

3 3 2.5 3 12 5

325 162.5 150 325 80 “Standard”

12 3 48 12

80 80 75–160 100

ACS ⫽ acute coronary syndrome; APRICOT 2 ⫽ Antithrombotics in the Prevention of Reocclusion in Coronary Thrombolysis; ASPECT 2 ⫽ Antithrombotics in the Secondary Prevention of Events in Coronary Thrombosis-2; ATACS ⫽ Antithrombotic Therapy in Acute Coronary Syndromes; NSTEMI ⫽ non–ST-elevation myocardial infarction; OASIS ⫽ Organization to Assess Strategies for Ischemic Syndromes; STEMI ⫽ ST-elevation myocardial infarction; UAP ⫽ unstable angina pectoris; WARIS II ⫽ Warfarin, Aspirin, Reinfarction Study II.

included studies are listed in Tables 1 and 2. End point definitions were homogeneous among included studies. In a previous meta-analysis of our group, combined treatment with A ⫹ W (at international normalized ratio 2 to 3) compared with A ⫹ P was associated with a lower risk of MAEs (OR 0.73, 95% CI 0.63 to 0.84). Specifically, A ⫹ W was associated with a lower risk of AMI (OR 0.70, 95% CI

Coronary Artery Disease/Combined Approaches After ACS

1639

Table 2 Main characteristics of studies comparing aspirin plus clopidogrel versus aspirin alone or plus placebo Trial CURE23 CLARITY-TIMI2824 COMMIT2

n

Type of ACS

Follow-Up (mo)

Aspirin Dose (mg/d)

12,562 3,491 45,852

NSTE-ACS STEMI AMI

9–12 1 ⱕ1

75–325 150–325 day 1 by 75–162 162

ACS ⫽ acute coronary syndrome; CLARITY-TIMI28 ⫽ Clopidogrel as Adjunctive Reperfusion Therapy-Thrombolysis in Myocardial Infarction; COMMIT ⫽ Clopidogrel and Metoprolol in Myocardial infarction Trial; CURE ⫽ Clopidogrel in Unstable angina to prevent Recurrent Events; NSTE-ACS ⫽ non–ST-elevation acute coronary syndrome; STEMI ⫽ ST-elevation myocardial infarction.

Figure 2. Overall analysis of risk of (A) MAEs for study comparing A ⫹ C versus aspirin alone or A ⫹ P. Single study ORs and 95% CIs shown by squares and lines. Single study random-effect ORs and 95% CIs shown by squares and horizontal lines. Overall OR with 95% CI shown by diamonds. Overall analysis of risk of (B) death; (C) AMI; and (D) thromboembolic stroke. ASA ⫽ aspirin.

0.52 to 0.95) and thromboembolic stroke (OR 0.43, 95% CI 0.27 to 0.70), whereas all-cause mortality was not significantly affected. A ⫹ W was also associated with increased risk of major bleeds (OR 2.32, 95% CI 1.63 to 3.29).5 Results according to both random- and fixed-effect models did not differ significantly; thus, in the case of significant inconsistency, data are presented according to the randomeffect model. Combination therapy with A ⫹ C compared

Figure 3. Overall analysis of risk of (A) major bleeds for study comparing A ⫹ C versus aspirin alone or A ⫹ P. Single study ORs and 95% CIs shown by squares and lines. Single study random-effect overall OR with 95% CI shown by diamonds. Overall analysis of risk of (B) intracranial bleeds and (C) extracranial major bleeds.

with A ⫹ P was associated with an overall decrease in risk of MAEs (OR 0.86, 95% CI 0.78 to 0.94, p ⫽ 0.0008, p for heterogeneity ⫽ 0.13, I2 ⫽ 50.9%). Specifically, A ⫹ C significantly decreased the risk of all-cause death (OR 0.93, 95% CI 0.87 to 0.99, p ⫽ 0.02, p for heterogeneity ⫽ 0.95, I2 ⫽ 0%), AMI (OR 0.77, 95% CI 0.70 to 0.86, p ⬍0.01, p for heterogeneity ⫽ 0.58, I2 ⫽ 0%), and thromboembolic stroke (OR 0.81, 95% CI 0.68 to 0.98, p ⫽ 0.03, p for heterogeneity ⫽ 0.03, I2 ⫽ 10.3%, Figure 2). In absolute

1640

The American Journal of Cardiology (www.AJConline.org)

Figure 4. Overall analysis of risk of MAEs for study comparing A ⫹ W (left side) and A ⫹ C (right side) versus aspirin alone. Abbreviation as in Figure 2.

terms, A ⫹ C was associated with an overall risk of MAEs of 9.8% versus about 11% with A ⫹ P. Specifically, the risk of all-cause death decreased with A ⫹ C from 7.4% to 6.9%; risk of AMI, from 2.7% to 2.1%; and risk of thromboembolic stroke, from 1% to 0.8%. The number needed to treat for A ⫹ C by saving 1 MAE was 50. Compared with A ⫹ P, A ⫹ C was associated with a significantly increased risk of major bleeds (OR 1.22, 95% CI 1.01 to 1.48, p ⫽ 0.04, p for heterogeneity ⫽ 0.23, I2 ⫽ 31.1%). In detail, A ⫹ C was associated with an increased risk of extracranial bleeds (OR 1.31, 95% CI 1.11 to 1.54, p ⫽ 0.002, p for heterogeneity ⫽ 0.49, I2 ⫽ 0%), but not of intracranial bleeds (OR 0.96, 95% CI 0.69 to 1.33, p ⫽ 0.79, p for heterogeneity ⫽ 0.57, I2 ⫽ 0%, Figure 3). In absolute terms, the risk of major bleeds increased from 1% with aspirin alone to 1.2% with A ⫹ C. The risk of intracranial bleeds was 0.2% in both treatment modes, whereas the risk of extracranial bleeds increased from 0.8% to 1% with A ⫹ C. Given the small difference between the 2 strategies, no significant risk difference was found, thus making the number needed to harm calculation impossible. Combination therapy did not significantly affect overall risk of stroke (ischemic or hemorrhagic, OR 0.83, 95% CI 0.67 to 1.01, p ⫽ 0.07, p for heterogeneity ⫽ 0.23, I2 ⫽ 31.6%. No heterogeneity was found among trials comparing A ⫹ W at international normalized ratio 2 to 3 versus aspirin alone.5 Moderate inconsistency was present among studies comparing A ⫹ C versus aspirin alone in the analysis for MAEs and major bleeds, possibly consistent with the different sample size of the Clopidogrel and Metoprolol Infarction Trial (COMMIT) mega-trial because excluding the latter, such inconsistency became nonsignificant. The likelihood of bias in studies comparing A ⫹ W versus aspirin alone previously was assessed.5 All studies comparing A ⫹ C versus A ⫹ P were of good quality, in other words, at low risk of bias, according to Cochrane Collaboration criteria. Publication or small-study bias was excluded because p values for asymmetry according to MAEs and major bleeds were 0.4 and 0.3, respectively. On

Figure 5. Adjusted indirect comparison of A ⫹ W versus A ⫹ C. ORs and 95% CIs shown by squares and horizontal lines. MB ⫽ overall major bleed; TES ⫽ thromboembolic stroke.

sensitivity analysis, exclusion of any single trial did not substantially alter overall results. None of the possible relevant citations was excluded for incomplete follow-up. Thus, such predefined exclusion criteria could not have introduced a selection bias into results. Using adjusted indirect comparison, no significant differences were found between A ⫹ W and A ⫹ C in risk of MAEs (OR 0.84, 95% CI 0.71 to 1.20, p ⫽ 0.2; I2 ⫽ 32.5%; Figures 4 and 5), all-cause death (OR 1.06, 95% CI 0.85 to 1.31, p ⫽ 0.5, I2 ⫽ 5%), and nonfatal myocardial infarction (OR 0.9, 95% CI 0.66 to 1.24, p ⫽ 0.6, I2 ⫽ 6%, Figure 4). Notably, A ⫹ W was associated with a lower risk of thromboembolic stroke (OR 0.53, 95% CI 0.31 to 0.88, p ⫽ 0.03, I2 ⫽ 6.5%; Figure 4). In absolute terms, the risk of thromboembolic stroke decreased from 0.8% with A ⫹ C to 0.5% with A ⫹ W, with a number needed to treat to save 1 thromboembolic stroke of 60. The risk of major bleeds was increased by A ⫹ W com-

Coronary Artery Disease/Combined Approaches After ACS

1641

Figure 6. Overall analysis of risk of major bleeds for study comparing A ⫹ W (left side) and A ⫹ C (right side) versus aspirin (ASA) alone.

pared with A ⫹ C (OR 1.9, 95% CI 1.2 to 2.8, p ⫽ 0.005, I2 ⫽ 6%; Figures 5 and 6), mainly because of a significant increase in rate of extracranial bleeds (OR 1.8, 95% CI 1.02 to 3.2, p ⫽ 0.04, I2 ⫽ 5%), whereas the risk of intracranial bleeds was not significantly different (OR 3.14, 95% CI 0.6 to 16.1, p ⫽ 0.7, I2 ⫽ 2.7%; Figure 4). In absolute terms, the risk of major bleeds increased from 1.3% to 2.6%. Such increased risk of major bleeds was related to the increased risk of extracranial bleeds because it increased from 1% to 2.2% with A ⫹ W. The number needed to harm by causing 1 major bleed was 300. These estimates suggest that allocating 1,000 patients to A ⫹ W should avoid 17 thromboembolic strokes while causing 3 major bleeds. Of note, A ⫹ W was associated with a decrease in risk of overall stroke with an OR of 0.59 (95% CI 0.35 to 0.97, p ⫽ 0.03, I2 ⫽ 3%). In absolute terms, overall incidences of stroke (hemorrhagic or ischemic) were 0.7% for A ⫹ W and 1% for A ⫹ C (chi square test p ⫽ 0.03). Assuming the event rates for A ⫹ C and risk change associated with A ⫹ W obtained from our indirect metaanalysis (2,908 of 30,972 [9.4%] and 0.86, respectively)5 and aiming for 80% power with 2-tailed alpha of 0.05, we calculated that 14,462 patients (7,231 per group) should be enrolled in a hypothetical future trial to adequately compare the impact of A ⫹ W versus A ⫹ C for the occurrence of MAEs. Discussion The risk– benefit profile of the combined approach of A ⫹ W over aspirin alone in secondary prevention after an acute coronary syndrome is well established.3–5 Our data comprehensively assessed efficacy and safety of the double antiplatelet therapy with A ⫹ C compared with aspirin alone, showing a lower risk of the composite end point of MAEs, but also of the single end points of overall mortality, nonfatal AMI, and nonfatal thromboembolic stroke. Indirect adjusted comparisons showed no difference between A ⫹ W and A ⫹ C with regard to MAEs, overall mortality, and AMIs. Intriguingly, A ⫹ W was associated with a 50% lower risk of stroke (in particular thromboembolic stroke), but at the price of a higher risk of extracranial major bleeds.

Aspirin deserved its cornerstone position as a treatment for reducing cardiovascular events in patients with established vascular disease based on consistent results of the large number of randomized controlled trials.1 Clopidogrel, by offering a different mechanism of action to decrease platelet activation and aggregation, showed similar or even greater effectiveness in reducing the risk of vascular events in patients with established vascular disease.25 However, in the Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) trial, its beneficial effect was not additive to that of aspirin in a low-risk population.26 A possible benefit during a 28-month follow-up was observed in the subgroup of patients with symptomatic vascular disease (MAE rates 6.9% with A ⫹ C and 7.9% with A ⫹ P), but a suggestion of harm was disclosed in asymptomatic patients (death rates from cardiovascular causes 3.9% in those receiving clopidogrel and 2.2% in those receiving placebo).26 In contrast, in the setting of acute coronary syndromes, the addition of clopidogrel to aspirin showed incremental benefit.27 Also for warfarin (at international normalized ratio 2 and 3) on top of aspirin, a number of data showed superiority versus aspirin alone.3–5 Such findings translated in current guidelines with various levels of evidence and recommendation.28,29 Aspirin should be continued indefinitely in the absence of clear contraindication, and also warfarin when indicated (i.e., atrial fibrillation, prosthetic valve, and left ventricle thrombus in the absence of high bleeding risk). Notably, there is no indication for the addition of clopidogrel on top of aspirin for ⬎1 year.28,29 A limitation inherent to all meta-analyses is the potential heterogeneity among studies in terms of protocols (e.g., type of acute coronary syndrome and length of follow-up), patients, and sample sizes. Such diversities may lead to inaccurate conclusions. However, both the Cochrane Q heterogeneity test (which assesses heterogeneity among ORs and the validity of pooling results) and test of inconsistency (I2) indicated significant diversity only among studies comparing A ⫹ C versus aspirin alone in the analysis for MAEs, possibly consistent with a significant difference in sample sizes of included studies. Although indirect comparisons,

1642

The American Journal of Cardiology (www.AJConline.org)

adjusted meta-analyses with inverse variance weighting are useful instruments in the assessment of different treatment strategies because they are usually confirmed by later randomized trials.8,12,30 Moreover, such methods are receiving progressive attention and diffusion.31

15.

16.

Appendix Model for the calculation of natural logarithm (ln), variance (var), and inconsistency (I2) of competing interventions using adjusted indirect meta-analysis with inverse variance weighting. A ⫽ first intervention/drug; B ⫽ common control; C ⫽ second intervention/drug, chi-square ⫽ test for heterogeneity. (1) ln (ORA vs C) ⫽ ln (ORA vs B) ⫺ ln (ORC vs B); (2) 2. var [ln (ORA vs C)] ⫽ var [ln (ORA vs B)] ⫹ var [ln (ORA vs C)]; (3) I2(A vs C) ⫽ (chi-square(A) ⫹ chi-square(C)) ⫻ 10/degrees of freedom(A) ⫹ degrees of freedom(C); (4) Once calculated the variance, it is necessary to calculate the SE, in other words, the square root of the variance; (5) CIs at 95% of ln are then calculated as ln ⫹1,96 * SE (upper CI) and ln ⫺1.96 * SE (lower CI); (6) The OR for the comparison of A versus C can be calculated as the exponent of the ln; (7) CIs of ORs are finally calculated as the exponent of the CI of the ln; and (8) Zeta score for calculation of p value for significance is finally derived from ln/SE. 1. Antithrombotic Trialists’ Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ 2002;324:71– 86. 2. COMMIT Collaborative Group. Addition of clopidogrel to aspirin in 45852 patients with acute myocardial infarction: randomised placebocontrolled trial. Lancet 2005;366:1607–1621. 3. Anand SS, Yusuf S. Oral anticoagulants in patients with coronary artery disease: a meta analysis. JAMA 1999;282:2058 –2067. 4. Anand SS, Yusuf S. Oral anticoagulants in patients with coronary artery disease. J Am Coll Cardiol 2003;41(suppl):S62–S69. 5. Andreotti F, Testa L, Biondi-Zoccai G, Crea F. Aspirin plus warfarin compared to aspirin alone after acute coronary syndromes: a comprehensive meta-analysis of 25307 patients. Eur Heart J 2006;27:519 –526. 6. Patsopoulos NA, Analatos AA, Ioannidis JPA. Relative citation impact of various study designs in health sciences. JAMA 2005;293:2362–2366. 7. Song F, Altman D, Glenny A-M, Deeks JJ. Validity of indirect comparison for estimating efficacy of competing interventions: empirical evidence from published meta-analysis. BMJ 2003;326:472. 8. Biondi Zoccai GG, Agostoni P, Abbate A, Testa L, Burzotta F. A simple hint to improve Robinson and Dickersin’s highly sensitive PubMed search strategy for controlled clinical trials. Int J Epidemiol 2005;34:224 –225. 9. Clarke M, Oxman AD, eds. Cochrane Reviewers’ Handbook 4.1.3. In: The Cochrane Library, Issue 3, 2001. Oxford, UK: Update Software. Updated quarterly. 10. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–560. 11. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629 – 634. 12. Biondi-Zoccai GGL, Agostoni P, Abbate A, Testa L, Burzotta F, Lotrionte M, Crea F, Biasucci M, Vetrovec GW, Colombo A. Adjusted indirect comparison of intracoronary drug-eluting stents: evidence from a metaanalysis of randomized bare-metal-stent-controlled trias. Int J Cardiol 2005;100:119 –123. 13. Cohen M, Adams PC, Hawkins L, Bach M, Fuster V. Usefulness of antithrombotic therapy in resting angina pectoris or non Q wave myocardial infarction in preventing death and myocardial infarction (a pilot study from the Antithrombotic Therapy in Acute Coronary Syndromes study group). Am J Cardiol 1990;66:1287–1292. 14. Cohen M, Adams PC, Parry G, Xiong J, Chamberlain D, Wieczorek I, Fox KA, Chesebro JH, Strain J, Keller C. Combination antithrombotic therapy in unstable rest angina and non-Q-wave infarction in non prior

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

aspirin users. Primary end points analysis from the ATACS Trial. Circulation 1994;89:81– 88. Williams MJA, Morison IM, Parker JH, Stewart RAH. Progression of the culprit lesion in unstable coronary artery disease with warfarin and aspirin versus aspirin alone: preliminary study. J Am Coll Cardiol 1997;30:364 –369. Anand SS, Yusuf S, Pogue J, Weitz JI, Flather M, for the OASIS Pilot Study Investigators. Long-term oral anticoagulant therapy in patients with unstable angina or suspected non-Q-wave myocardial infarction. Organization to Assess Strategies for Ischemic Syndromes (OASIS) pilot study results. Circulation 1998;98:1064 –1070. Huynh T, Theroux P, Bogaty P, Nasmith J, Solymoss S. Aspirin, warfarin, or the combination for secondary prevention of coronary events in patients with acute coronary syndromes and prior coronary artery bypass surgery. Circulation 2001;103:3069 –3074. The Organization to Assess Strategies for Ischemic Syndromes (OASIS) Investigators. Effects of long term, moderate intensity oral anticoagulation in addition to aspirin in unstable angina. J Am Coll Cardiol 2001;37:475– 484. van Es RF, Jonker JJ, Verheugt FW, Deckers JW, Grobbee DE. Aspirin and Coumadin after acute coronary syndromes (the ASPECT-2 study): a randomised controlled trial. Lancet 2002;360:109 –113. Brouwer MA, van den Bergh PJ, Aengevaeren WR, Veen G, Luijten HE, Hertzberger DP, van Boven AD, Vromans R, Uijen G, Verheugt F. Aspirin plus coumarin versus aspirin alone in the prevention of reocclusion after fibrinolysis for acute myocardial infarction: results of the APRICOT 2 trial. Circulation 2002;106:659 – 665. Hurlen M, Abdelnoor M, Smith P, Erikssen J, Arnesen H. Warfarin, aspirin, or both after myocardial infarction. N Engl J Med 2002;347: 969 –974. Zibaeenezhad MJ, Mowla A, Sorbi MH. Warfarin and aspirin versus aspirin alone in patients with acute myocardial infarction: a pilot study. Angiology 2004;55:17–20. The Clopidogrel in Unstable Angina to Prevent Recurrent Events Trial Investigators. 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. Sabatine MS, Cannon CP, Gibson CM, Lopez-Sendon JL, Montalescot G, Theroux P, Claeys MJ, Cools F, Hill KA, Skene AM, McCabe CH, Braunwald E, for the CLARITY–TIMI 28 Investigators. Addition of clopidogrel to aspirin and fibrinolytic therapy for myocardial infarction with ST segment elevation. N Engl J Med 2005;352:1179 –1189. CAPRIE Steering Committee. A randomised, blinded trial of Clopidogrel versus Aspirin in Patients at Risk of Ischaemic Events (CAPRIE). Lancet 1996;348:1329 –1339. Bhatt DL, Fox KA, Hacke W, Berger PB, Black AR, Boden WE, Cacoub P, Cohen FA, Creager MA, Easton SD, Flatner MD, Hapfner SM, et al. Clopidogrel plus aspirin versus aspirin alone for the prevention of atherothrombotic events. N Engl J Med 2006;354:1706 –1717. Mehta SR, Yusuf S, Peters RJ, Bertrand M, Lewis BS, Natarajan MK, Malmberg K, Rupprecht H, Zhad F, et al. Effects of pretreatment with clopidogrel and aspirin followed by long-term therapy in patients undergoing percutaneous coronary intervention: the PCI-CURE study. Lancet 2001;358:527–533. ACC/AHA Guidelines for the Management of Patients With STElevation Myocardial Infarction. A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Revise the 1999 Guidelines for the Management of Patients With Acute Myocardial Infarction) Developed in Collaboration With the Canadian Cardiovascular Society. Available at www.americanheart.org. Accessed October 2006. ACC/AHA. Guidelines for the Management of Patients With Unstable Angina and Non-ST-Segment Elevation Myocardial Infarction. J Am Coll Cardiol 2000;36:970 –1062. Kastrati A, Mehilli J, von Beckerath N, Dibra A, Hausleiter J, Pache J, Schuhlen H, Schmitt C, Dirschinger J, Schomig A, for the ISARDESIRE Study Investigators. Sirolimus-eluting stent or paclitaxeleluting stent vs balloon angioplasty for prevention of recurrences in patients with coronary in-stent restenosis: a randomized controlled trial. JAMA 2005;293:165–171. Zhou Z, Rahme E, Pilote L. Are statins created equal? Evidence from randomized trials of pravastatin, simvastatin, and atorvastatin for cardiovascular disease prevention. Am Heart J 2006;151:273–281.