CLINICAL RESEARCH STUDY

Effect of Aspirin on Mortality in the Primary Prevention of Cardiovascular Disease

Nina Raju, MD, MsC,a Magdalena Sobieraj-Teague, MBBS,b Jack Hirsh, MD,c Martin ODonnell, MD, PhD,c John Eikelboom, MD, MsCc
a Haematology Unit, Queensland Pathology and Department of Internal Medicine, Prince Charles Hospital, Brisbane, Australia; bSA Pathology, Flinders Medical Centre, Adelaide, Australia; cThrombosis Medicine, McMaster University, Hamilton, Canada.

ABSTRACT OBJECTIVE: The lack of a mortality benet of aspirin in prior meta-analyses of primary prevention trials of cardiovascular disease has contributed to uncertainty about the balance of benets and risks of aspirin in primary prevention. We performed an updated meta-analysis of randomized controlled trials of aspirin to obtain best estimates of the effect of aspirin on mortality in primary prevention. METHODS: Eligible articles were identied by searches of electronic databases and reference lists. Outcomes of interest were all-cause mortality, cardiovascular mortality, myocardial infarction, stroke, and bleeding. Data were pooled from individual trials using the DerSimonian-Laird random-effects model, and results are presented as relative risk (RR) and 95% condence intervals (CIs). RESULTS: Nine randomized controlled trials enrolling 100,076 participants were included. Aspirin reduced all-cause mortality (RR 0.94; 95% CI, 0.88-1.00), myocardial infarction (RR 0.83; 95% CI, 0.69-1.00), ischemic stroke (RR 0.86; 95% CI, 0.75-0.98), and the composite of myocardial infarction, stroke, or cardiovascular death (RR 0.88; 95% CI, 0.83-0.94), but did not reduce cardiovascular mortality (RR 0.96; 95% CI, 0.84-1.09). Aspirin increased the risk of hemorrhagic stroke (RR 1.36; 95% CI, 1.01-1.82), major bleeding (RR 1.66; 95% CI, 1.41-1.95), and gastrointestinal bleeding (RR 1.37; 95% CI, 1.15-1.62). A lack of availability of patient-level data precluded exploration of benets and risks of aspirin in key subgroups. CONCLUSION: Aspirin prevents deaths, myocardial infarction, and ischemic stroke, and increases hemorrhagic stroke and major bleeding when used in the primary prevention of cardiovascular disease. 2011 Elsevier Inc. All rights reserved. The American Journal of Medicine (2011) 124, 621-629 KEYWORDS: Aspirin; Meta-analysis; Primary prevention

Aspirin reduces nonfatal and fatal cardiovascular events in patients with a history of symptomatic cardiovascular disease, and it is widely accepted that the benets of aspirin outweigh the increased risk of bleeding in this setting.1,2 In patients without a history of cardiovascular disease, aspirin has not been shown to reduce mortality,3-10 and it is uncer-

Funding: None. Conict of Interest: None of the authors have any conicts of interest associated with the work presented in this manuscript. Authorship: All authors had access to the data and played a role in writing this manuscript. Requests for reprints should be addressed to Magdalena SobierajTeague, MBBS, Flinders Medical Centre, Flinders Drive, Bedford Park, SA 5042, Australia. E-mail address: magdalena.sobieraj-teague@health.sa.gov.au

tain whether the reduction in myocardial infarction and ischemic stroke outweighs the increased risk of bleeding. Uncertainty about the balance between the risks and the benets of aspirin in the primary prevention of cardiovascular disease is reected in conicting recommendations by guideline panels.11-15 The 2002 and 2007 American Heart Association guidelines and the 2008 American College of Chest Physicians guidelines recommend aspirin for primary prevention of cardiovascular disease in patients with a 10-year risk of coronary heart disease of 10%,11,12,14 the 2005 Joint British Societies guidelines recommend aspirin for patients with a 10-year risk of cardiovascular events of 20%,16 and the 2007 European Society of Cardiology guidelines recommend aspirin for patients with a markedly increased 10-year risk of cardiovascular disease mortality.13

0002-9343/$ -see front matter 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.amjmed.2011.01.018

622

The American Journal of Medicine, Vol 124, No 7, July 2011

The 2009 US Preventive Services Taskforce guideline panel Cochrane library (to May 2010) using the terms aspirin, makes separate recommendations for aspirin according to acetylsalicylic acid, cardiovascular disease, myocardial inpatient age and sex, and provides thresholds of 10-year farction, stroke, cerebrovascular disease, mortality, death, coronary heart disease risk (men) and stroke risk (women) survival, randomized trial, controlled trial, random, preat which the number of nonfatal cardiovascular events and vent, and primary prevention. Bibliographies of journal arserious bleeds is closely balticles were hand-searched, and a anced.15 The 2009 Antithrombotic related article PubMed search Trialists Collaboration overview was performed to identify addiCLINICAL SIGNIFICANCE interprets the primary prevention tional relevant articles. We also data differently. By contrasting searched the National Institutes of In patients without a history of cardiothe lack of a signicant mortality Health Clinical Trials Registry vascular disease, long-term treatment benet of aspirin in primary pre(www.clinicaltrials.gov) and conwith aspirin reduces all-cause mortality, vention trials with the 10% reductacted experts to identify unpubmyocardial infarction, and ischemic tion in mortality in the secondary lished studies. stroke, and increases hemorrhagic stroke prevention trials, the Antithromand major bleeding. botic Trialists Collaboration writStudy Selection ing group concluded that aspirin is The reduction in all-cause mortality Two investigators (MST, NR) of uncertain net value in patients independently evaluated studies should be taken into account by cliniwithout documented cardiovascufor inclusion using predened cians and guideline panels when making lar disease because the reduction criteria. Relevance was assessed recommendations about the use of asin ischemic events is closely using a hierarchic approach pirin for primary prevention of cardiomatched by the increase in major based on title, abstract, and full vascular disease. bleeding, irrespective of an inditext publication. Disagreement viduals baseline risk of cardioon study eligibility was resolved vascular disease.10 by discussion and involvement Since the publication of the aforementioned guideof a third reviewer (JWE). lines,10-16 the results of 3 additional randomized controlled To be eligible for inclusion, studies had to meet the trials of aspirin for the primary prevention of cardiovascular following criteria: (a) randomized controlled trial; (b) indisease involving more than 7000 subjects have been pubclude adults without a history of symptomatic cardiovasculished.17-19 Subjects in these trials did not have a history of lar disease ( 95% of enrolled participants); (c) compare symptomatic cardiovascular disease but were at increased aspirin (any dose) with placebo or no aspirin treatment risk of cardiovascular events because of diabetes or a refor the prevention of cardiovascular disease; and (d) duced ankle-brachial index. report at least one of the following outcomes: all-cause We performed a meta-analysis of all randomized conmortality, cardiovascular mortality, myocardial infarctrolled trials of aspirin for the primary prevention of cartion, stroke, and bleeding. Randomized controlled trials diovascular disease to obtain best estimates of aspirin comin which aspirin was combined with a second antithrompared with placebo or no aspirin on all-cause and botic agent were not included, unless there were separate cardiovascular mortality and on other major cardiovascular placebo and aspirin-only treatment groups, in which case outcomes, including myocardial infarction, stroke, and only the data from these groups were included. bleeding.

MATERIALS AND METHODS

We adopted the Preferred Reporting Items for Systematic reviews and Meta-Analyses guidelines for the reporting of this meta-analysis.20,21 A protocol was prospectively developed detailing the study objectives, primary and secondary outcomes, criteria for study selection, approach to assessing study quality, data synthesis, and statistical analyses.

Data Extraction and Quality Assessment

Two investigators (MST, NR) independently extracted data on study design, participant characteristics, eligibility criteria, intervention and comparator, quality, and the following outcomes: all-cause mortality, cardiovascular mortality, major cardiovascular events, myocardial infarction, all-cause stroke, ischemic stroke, hemorrhagic stroke, major bleeding, and gastrointestinal bleeding. We accepted investigators denitions of outcomes and did not retrospectively reclassify events. Authors were contacted in cases in which study methodology or study data required clarication. Risk of bias was assessed using criteria adapted from the Cochrane Methods Group Guidelines on Systematic Reviews of Interventions.22 These criteria include proper generation of the treatment allocation sequence; proper con-

Data Sources and Searches

We identied relevant published randomized controlled trials comparing aspirin with placebo or no aspirin treatment in individuals without a history of symptomatic cardiovascular disease. Two investigators (MST, NR) independently searched MEDLINE (1966 to May 2010), EMBASE (1980 to May 2010), CINAHL (1982 to May 2010), and the

Raju et al

Effect of Aspirin on Mortality

623

Figure 1

Process of study selection.

cealment of the allocation sequence; blinding of participants, investigators, and outcome assessors; completeness of follow-up; intention to treat analysis; and treatment compliance.

Data Synthesis and Analysis

Interobserver agreement for full text study selection was measured using Cohens unweighted kappa statistic. Results are presented using relative risk, and all effect estimates are presented with 95% condence intervals (CIs). Relative risks for the prespecied primary and secondary outcomes were calculated by pooling individual trial data with the DerSimonian-Laird random-effects model using Cochrane Collaboration Review Manager software, version 5.0 (The Nordic Cochrane Centre, Copenhagen, The Cochrane Collaboration, 2008). Results obtained with a random-effects model were compared with those obtained using a xed-effects model. A 2-sided P value of .05 was considered statistically signicant. Heterogeneity was assessed using the I2 statistic and chi-square test, and potential sources of statistical

heterogeneity were explored by examining differences in trial populations, the dose of aspirin, co-interventions, outcome denitions, and trial quality. Sensitivity analyses were performed by exclusion of studies deemed to be of lower quality (open-label studies, incomplete follow-up), studies using a higher dose of aspirin ( 150 mg/d), and studies completed before 2000 to assess the robustness of our results.

RESULTS Study Selection

The process of study selection is outlined in Figure 1. Reviewer agreement at the level of study selection from full text articles as assessed by Cohens unweighted kappa was 0.86 (standard error 0.14).

Study Characteristics
Nine studies involving 100,076 participants were included. Study and participant characteristics are summarized in Table 1. Three of the nine studies, the British Doctors Trial

624
14% 70 y 7% 70 y 32% 65 y N/A 50% 65 y 10% 65 y 54% 65 y 52% 60 y N/A

The American Journal of Medicine, Vol 124, No 7, July 2011

BDT, British Doctors Trial; PHS, Physicians Health Study; TPT, Thrombosis Prevention Trial; HOT, Hypertension Optimal Treatment study; PPP, Primary Prevention Project trial; WHS, Womens Health Study; JPAD, Japanese Primary Prevention of Atherosclerosis with Aspirin for Diabetes trial; POPADAD, Prevention Of Progression Of Arterial Disease And Diabetes trial; AAA, Aspirin for Asymptomatic Atherosclerosis trial; DM, diabetes mellitus; HT, hypertension; N/A, not available.

Current Smokers (%)

(BDT),23 Physicians Health Study (PHS),24 and Thrombosis Prevention Trial (TPT),25 did not include women, whereas the Womens Health Study (WHS)26 exclusively enrolled women. The Prevention of Progression of Arterial Disease and Diabetes (POPADAD) trial17 and Japanese Primary Prevention of Atherosclerosis with Aspirin for Diabetes (JPAD)18 recruited only patients with diabetes mellitus. The Aspirin for Asymptomatic Atherosclerosis (AAA)19 study enrolled men and women without a history of symptomatic cardiovascular disease but with an anklebrachial index 0.95. Two studies, BDT23 and the Hypertension Optimal Trial (HOT),27 included a small proportion of participants ( 5%) with a history of ischemic heart disease, stroke, or transient ischemic attack.

Mean Age(y) HT (%) DM (%)

Elderly

2 2 8 2 17 3 100 100 3

10 9 100 26 68 26 58 N/A N/A

N/A N/A 61.5 57.3 64.4 54.6 64.5 60.3 62.0

Risk of Bias
With the exception of the BDT, PHS, and WHS,23,24,26 which did not specify the method of allocation concealment, all studies reported the process of sequence generation and concealment of allocation. The PHS, TPT, WHS, POPADAD, AAA, and HOT17,19,24-27 were doubleblinded studies, and the remaining 3 trials, the BDT, JPAD, and Primary Prevention Project (PPP),18,23,28 were openlabel. Although the open-label design is unlikely to have biased the main outcome of mortality, it may have inuenced ascertainment of nonfatal cardiovascular outcomes, in particular when outcomes were self-reported. In all trials, outcome adjudicators were blinded to treatment allocation. All studies reported completeness of follow-up. Vital status was determined for all participants in the BDT, PHS, and TPT.23-25 Loss to follow-up was highest (7.6%) in the JPAD study,18 followed by the HOT (2.6%),27 and was less than 1% in the other studies. The numbers lost to follow-up exceeded the total number of primary events in the HOT, WHS, and JPAD.18,26,27 Treatment adherence was not reported for the WHS26 and ranged from 50% to 93% for the other studies with the lowest adherence in the POPADAD trial.17 All studies were analyzed according to an intention-totreat principle. Three studies were terminated prematurely, the PHS and PPP24,28 because of the apparent therapeutic advantage of aspirin, and the AAA19 because of the improbability of nding a difference in the primary end point and an increased risk of bleeding with aspirin.

Mean Follow-Up (y) Male (%) Additional Therapies

Study and Participant Characteristics

300-500 mg/d 325 mg alternate day 75 mg/d 75 mg/d 100 mg/d 100 mg alternate day 81 mg or 100 mg/d 100 mg/d 100 mg/d

Aspirin Dose

Open label Double blinded Double blinded Double blinded Open label Double blinded Open label Double blinded Double blinded

Design

None Beta carotene Felodipine other agents to achieve target BP Warfarin Vitamin E Vitamin E and beta carotene None Antioxidant None

100 100 53 100 42 0 55 44 28

6 5 3.8 6.4 3.6 10.1 4.4 (median) 6.7 (median) 8.2

31 11 16 41 15 13 21 31 33

Outcomes
All-Cause Mortality. Aspirin therapy was associated with a 6% reduction in all-cause mortality (3.65% vs 3.74%, RR 0.94; 95% CI, 0.88-1.0) (Figure 2, Table 2). With the exception of TPT,25 the point estimate in each study suggested a reduction in mortality, although none of the studies were powered to detect a mortality benet of aspirin. There was no statistical evidence of heterogeneity among the studies for the outcome of all-cause mortality (I2 0%, 2 1.70, P .99).

Patients (n) Study Name (Publication Year)

BDT23 (1988) PHS24(1989) HOT27 (1998) TPT25 (1998) PPP28 (2001) WHS26 (2005) JPAD18 (2008) POPADAD17 (2008) AAA19 2010

Table 1

5139 22,071 18,790 2540 4495 39,876 2539 1276 3350

Raju et al

Effect of Aspirin on Mortality

Aspirin Placebo/ No treatment Risk Ratio Study or Subgroup Events Total Events Total M-H, Random, 95% CI Year 1.1.1 All-cause mortality BDT PHS HOT TPT PPP WHS POPADAD JPAD AAA Subtotal (95% CI) 270 3429 217 11037 284 9399 113 1268 62 2226 609 19934 94 638 34 1262 176 1675 50868 151 227 305 110 78 642 101 38 186 1710 11034 9391 1272 2269 19942 638 1277 1675 49208 0.89 [0.74, 1.08] 0.96 [0.79, 1.15] 0.93 [0.79, 1.09] 1.03 [0.80, 1.32] 0.81 [0.58, 1.13] 0.95 [0.85, 1.06] 0.93 [0.72, 1.21] 0.91 [0.57, 1.43] 0.95 [0.78, 1.15] 0.94 [0.88, 1.00] 1988 1989 1998 1998 2001 2005 2008 2008 2010 Risk Ratio M-H, Random, 95% CI

625

Total events 1859 1838 Heterogeneity: Tau = 0.00; Chi = 1.70, df = 8 (P = 0.99); I = 0% Test for overall effect: Z = 2.00 (P = 0.05) 1.1.2 Cardiovascular mortality BDT PHS HOT TPT PPP WHS JPAD POPADAD AAA Subtotal (95% CI) 148 3429 81 11037 133 9399 49 1268 17 2226 120 19934 1 1262 43 638 35 1675 50868 79 83 140 49 31 126 10 35 30 1710 11034 9391 1272 2269 19942 1277 638 1675 49208 0.93 [0.72, 1.22] 0.98 [0.72, 1.32] 0.95 [0.75, 1.20] 1.00 [0.68, 1.48] 0.56 [0.31, 1.01] 0.95 [0.74, 1.22] 0.10 [0.01, 0.79] 1.23 [0.80, 1.89] 1.17 [0.72, 1.89] 0.96 [0.84, 1.09] 1988 1989 1998 1998 2001 2005 2008 2008 2010

Total events 627 583 Heterogeneity: Tau = 0.01; Chi = 9.87, df = 8 (P = 0.27); I = 19% Test for overall effect: Z = 0.66 (P = 0.51) 0.5 0.7 1 1.5 2 Favours Aspirin Favours control

Figure 2 aspirin.

Relative risk of all-cause mortality and cardiovascular mortality in participants treated with and without

Cardiovascular Mortality. There was no reduction in the risk of cardiovascular mortality in participants assigned to aspirin (Figure 2, Table 2). Only the JPAD18 showed a reduction in cardiovascular death with aspirin, but this was based on only 11 events. Major Cardiovascular Events (Composite of Cardiovascular Mortality, Nonfatal Myocardial Infarction, and Nonfatal Stroke). Aspirin use was associated with a 12% reduction in major cardiovascular events (3.66% vs 3.98%, RR 0.88; 95% CI, 0.83-0.94) with no statistical evidence of heterogeneity (I2 0%, 2 7.56, P .48) (Figure 3, Table 2). The reduction in major cardiovascular events was primarily attributable to a reduction in myocardial infarction. Myocardial Infarction (Composite of Fatal and Nonfatal Myocardial Infarction). The largest measured benet of aspirin was in the prevention of myocardial infarction with a relative risk reduction of 17% (1.68% vs 1.91%, RR 0.83; 95% CI, 0.69-1.00) (Figure 3, Table 2). Heterogeneity in

71%, 2 27.51, myocardial infarction results (I2 P .0006) was not explained by differences in trial populations, aspirin dose, or methodological quality of the studies (data not shown), but statistical heterogeneity was no longer evident after removing the PHS from the analyses.24 The PHS was stopped early because of benet, which can result in an inated estimate of treatment effect.29 Stroke (Composite of Fatal and Nonfatal Stroke). Aspirin reduced the risk of ischemic stroke (0.97% vs 1.10%, RR 0.86; 95% CI, 0.75-0.98, P for heterogeneity 0.48), but this was offset by a concomitant increase in hemorrhagic stroke (0.27% vs 0.19%, RR 1.36; 95% CI, 1.011.82, P for heterogeneity 0.63). Thus, there was no reduction in all-cause stroke (Figures 3 and 4, Table 2). Bleeding. Aspirin increased the risk of major bleeding (RR 1.66; 95% CI, 1.41-1.95) with no statistical evidence 0%, 2 6.01, P .42) (Figure of heterogeneity (I2 4). Aspirin also increased the risk of gastrointestinal bleeding. Statistical heterogeneity for the outcome of

626
Table 2 Outcome All-cause mortality Cardiovascular mortality Major cardiovascular events Myocardial infarction All-cause stroke Ischemic stroke Hemorrhagic stroke Gastrointestinal bleed Major bleed
CI Condence interval.

The American Journal of Medicine, Vol 124, No 7, July 2011

Pooled Estimates of the Benets and Risks of Aspirin in Primary Prevention of Cardiovascular Disease Total Events in Aspirin Arm (%) 1859 627 1861 854 720 403 112 1947 406 (3.65) (1.23) (3.66) (1.68) (1.42) (0.97) (0.27) (4.10) (0.83) Total Events in NonAspirin Arm (%) 1838 583 1957 942 733 439 76 1560 234 (3.74) (1.18) (3.98) (1.91) (1.49) (1.10) (0.19) (3.28) (0.50) Pooled Relative Risk (95% CI) 0.94 0.96 0.88 0.83 0.93 0.86 1.36 1.37 1.66 (0.88-1.00) (0.84-1.09) (0.83-0.94) (0.69-1.00) (0.82-1.05) (0.75-0.98) (1.01-1.82) (1.15-1.62) (1.41-1.95) Relative Risk Reduction (95% CI) 6% 4% 12% 17% 7% 14% 36% 37% 66% (0-12) ( 9 to 16) (6-17) (0-31) ( 5 to 18) (2-25) ( 1 to 82) ( 15 to 62) ( 41 to 95)

gastrointestinal bleeding was not explained by differences in study populations, aspirin dose, or study quality (data not shown). Sensitivity Analyses. Sensitivity analysis performed by excluding studies considered to be at higher risk of bias (open label: BDT, JPAD, and PPP;18,23,28 high loss to follow-up ( 5%): JPAD;18 early termination of study: PHS, PPP, AAA19,24,28) demonstrated a consistent treatment effect of aspirin on all-cause mortality (data not shown). Similarly, consistent results were seen when studies using aspirin at a dosage of 150 mg/d (BDT, PHS)23,24 were excluded (RR 0.94; 95% CI, 0.88-1.01) and when studies published before 2000 were excluded (BDT, PHS, HOT, TPT)23-25,27 (data not shown). Assessment of Publication Bias. Beggs adjusted-rank correlation test30 provided no evidence of publication bias for any of the outcomes that we examined (data not shown).

DISCUSSION
Our meta-analysis of 9 aspirin primary prevention trials involving 100,076 patients demonstrates that long-term aspirin compared with placebo or no aspirin reduces all-cause mortality, myocardial infarction, and ischemic stroke; does not reduce cardiovascular mortality; and increases hemorrhagic stroke, major bleeding, and gastrointestinal bleeding. In the absence of a mortality benet of aspirin, treatment guidelines for the use of aspirin for primary prevention of cardiovascular disease have based their recommendations on the trade-off between the reduction of nonfatal myocardial infarction and the increase in bleeding. Applying this trade-off in clinical practice can be challenging because the balance between a reduction in nonfatal ischemic events and an increase in bleeding is strongly inuenced by individual patient values and preferences. The authors of the 2009 Antithrombotic Trialists Collaboration overview suggested that the benets

and harms of aspirin were so nely balanced that aspirin was of uncertain value in patients without previous cardiovascular disease, irrespective of their baseline cardiovascular risk.10 Our results demonstrating that aspirin reduces all-cause mortality in individuals without a history of symptomatic cardiovascular disease tilts the balance in favor of the use of aspirin for primary prevention and should be taken into account by clinicians and future treatment guidelines. The most likely reason why our meta-analysis reported a reduction in mortality when aspirin is used for the primary prevention of cardiovascular disease whereas previous meta-analyses did not is that our meta-analysis included more patients and more outcome events, thereby improving the precision of the estimates of treatment effect. The 3 recently published trials (POPADAD, JPAD, AAA) included in this meta-analysis involved patients with diabetes or asymptomatic atherosclerosis who were underrepresented in previous trials. Further information regarding the efcacy and safety of aspirin for primary prevention of cardiovascular disease and in specic populations will be provided by the results of ongoing randomized controlled trials of aspirin in the elderly (Aspirin in Reducing Events in the Elderly trial),31 patients with diabetes (Aspirin and Simvastatin Combination for Cardiovascular Events Prevention Trial in Diabetes; A Study of Cardiovascular Events in Diabetes),32,33 and in different ethnic groups (eg, Japanese Primary Prevention Project).34 We propose 2 possible explanations for why aspirin reduces all-cause mortality but not cardiovascular mortality. The rst possible explanation is that aspirin reduces both cardiovascular and noncardiovascular deaths so that when the data for these 2 outcomes are combined, a signicant reduction in all-cause mortality becomes evident. Aspirin has been reported to reduce mortality in patients with cancer,35,36 and the Antithrombotic Trialists Collaboration meta-analysis also reported fewer noncardiovascular deaths in patients receiving aspirin compared with placebo for the secondary prevention of

Raju et al

Effect of Aspirin on Mortality

Aspirin Placebo/ No treatment Events Total Events Total Study or Subgroup 1.2.1 Cardiovascular events BDT PHS HOT TPT PPP WHS POPADAD JPAD AAA Subtotal (95% CI) 289 3429 320 11037 315 9399 112 1268 47 2226 477 19934 638 127 40 1262 134 1675 50868 147 388 368 147 71 522 132 46 136 1710 11034 9391 1272 2269 19942 638 1277 1675 49208 Risk Ratio M-H, Random, 95% CI Year 0.98 [0.81, 1.19] 0.82 [0.71, 0.95] 0.86 [0.74, 0.99] 0.76 [0.61, 0.97] 0.67 [0.47, 0.97] 0.91 [0.81, 1.03] 0.96 [0.77, 1.20] 0.88 [0.58, 1.33] 0.99 [0.78, 1.24] 0.88 [0.83, 0.94] 1988 1989 1998 1998 2001 2005 2008 2008 2010 Risk Ratio M-H, Random, 95% CI

627

1957 1861 Total events Heterogeneity: Tau = 0.00; Chi = 7.56, df = 8 (P = 0.48); I = 0% Test for overall effect: Z = 3.96 (P < 0.0001) 1.2.2 Myocardial infarction BDT PHS HOT TPT PPP WHS JPAD POPADAD AAA Subtotal (95% CI) 169 3429 139 11037 82 9399 69 1268 19 2226 198 19934 12 1262 638 76 90 1675 50868 88 239 127 98 28 193 14 69 86 1710 11034 9391 1272 2269 19942 1277 638 1675 49208 0.96 [0.75, 1.23] 0.58 [0.47, 0.72] 0.65 [0.49, 0.85] 0.71 [0.52, 0.95] 0.69 [0.39, 1.23] 1.03 [0.84, 1.25] 0.87 [0.40, 1.87] 1.10 [0.81, 1.50] 1.05 [0.78, 1.40] 0.83 [0.69, 1.00] 1988 1989 1998 1998 2001 2005 2008 2008 2010

942 Total events 854 Heterogeneity: Tau = 0.05; Chi = 27.51, df = 8 (P = 0.0006); I = 71% Test for overall effect: Z = 1.99 (P = 0.05) 1.2.3 Stroke BDT PHS HOT TPT PPP WHS JPAD POPADAD AAA Subtotal (95% CI) 91 3429 119 11037 146 9399 18 1268 16 2226 221 19934 28 1262 638 37 44 1675 50868 39 98 148 26 24 266 32 50 50 1710 11034 9391 1272 2269 19942 1277 638 1675 49208 1.16 [0.80, 1.69] 1.21 [0.93, 1.58] 0.99 [0.79, 1.24] 0.69 [0.38, 1.26] 0.68 [0.36, 1.28] 0.83 [0.70, 0.99] 0.89 [0.54, 1.46] 0.74 [0.49, 1.12] 0.88 [0.59, 1.31] 0.93 [0.82, 1.05] 1988 1989 1998 1998 2001 2005 2008 2008 2010

733 Total events 720 Heterogeneity: Tau = 0.01; Chi = 10.22, df = 8 (P = 0.25); I = 22% Test for overall effect: Z = 1.19 (P = 0.23) 1.2.4 Ischemic stroke BDT PHS TPT PPP WHS POPADAD JPAD AAA Subtotal (95% CI) 61 3429 91 11037 10 1268 16 2226 170 19934 638 3 22 1262 30 1675 41469 29 82 18 22 221 5 25 37 1710 11034 1272 2269 19942 638 1277 1675 39817 1.05 [0.68, 1.63] 1.11 [0.82, 1.49] 0.56 [0.26, 1.20] 0.74 [0.39, 1.41] 0.77 [0.63, 0.94] 0.60 [0.14, 2.50] 0.89 [0.50, 1.57] 0.81 [0.50, 1.31] 0.86 [0.75, 0.98] 1988 1989 1998 2001 2005 2008 2008 2010

439 403 Total events Heterogeneity: Tau = 0.00; Chi = 6.55, df = 7 (P = 0.48); I = 0% Test for overall effect: Z = 2.25 (P = 0.02) 0.5 0.7 1 1.5 2 Favours aspirin Favours control

Figure 3

Ischemic cardiovascular events in participants treated with and without aspirin.

cardiovascular disease.10 The second possible explanation is that cardiovascular deaths may have been misattributed to noncardiovascular causes. A potential limitation of our meta-analysis is that the included studies were conducted over a 30-year period during

which there have been major advances in cardiovascular prevention and an increase in the concomitant use of other effective primary prevention strategies, such as blood pressure control and lipid-lowering therapy. Despite an increase in the use of other cardiovascular prevention therapies, the event rates in

628

The American Journal of Medicine, Vol 124, No 7, July 2011

Aspirin Placebo/ No treatment Risk Ratio Study or Subgroup Events Total Events Total M-H, Random, 95% CI Year 1.3.1 Hemorrhagic stroke BDT PHS TPT PPP WHS JPAD POPADAD AAA Subtotal (95% CI) 13 3429 23 11037 2 1268 10 2226 51 19934 6 1262 2 638 5 1675 41469 6 12 0 3 41 7 3 4 1710 11034 1272 2269 19942 1277 638 1675 39817 1.08 [0.41, 2.84] 1.92 [0.95, 3.85] 5.02 [0.24, 104.37] 3.40 [0.94, 12.33] 1.24 [0.83, 1.88] 0.87 [0.29, 2.57] 0.67 [0.11, 3.98] 1.25 [0.34, 4.65] 1.36 [1.01, 1.82] 1988 1989 1998 2001 2005 2008 2008 2010 Risk Ratio M-H, Random, 95% CI

Total events 112 76 Heterogeneity: Tau = 0.00; Chi = 5.27, df = 7 (P = 0.63); I = 0% Test for overall effect: Z = 2.03 (P = 0.04) 1.3.2 Major bleeding BDT PHS HOT TPT PPP WHS AAA Subtotal (95% CI) 29 3429 48 11037 136 9399 8 1268 24 2226 127 19934 34 1675 48968 7 28 78 4 6 91 20 1710 10979 9391 1278 2269 19942 1675 47244 2.07 [0.91, 4.71] 1.71 [1.07, 2.72] 1.74 [1.32, 2.30] 2.02 [0.61, 6.68] 4.08 [1.67, 9.96] 1.40 [1.07, 1.83] 1.70 [0.98, 2.94] 1.66 [1.41, 1.95] 1988 1989 1998 1998 2001 2005 2010

Total events 406 234 Heterogeneity: Tau = 0.00; Chi = 6.01, df = 6 (P = 0.42); I = 0% Test for overall effect: Z = 6.14 (P < 0.00001) 1.3.3 Gastrointestinal bleeding PHS TPT HOT PPP WHS POPADAD JPAD AAA Subtotal (95% CI) 842 11037 22 1268 107 9399 17 2226 910 19934 28 638 12 1262 9 1675 47439 696 10 55 5 751 31 4 8 11034 1272 9391 2269 19942 638 1277 1675 47498 1.21 [1.10, 1.33] 2.21 [1.05, 4.64] 1.94 [1.41, 2.69] 3.47 [1.28, 9.38] 1.21 [1.10, 1.33] 0.90 [0.55, 1.49] 3.04 [0.98, 9.39] 1.13 [0.44, 2.91] 1.37 [1.15, 1.62] 1989 1998 1998 2001 2005 2008 2008 2010

Total events 1947 1560 Heterogeneity: Tau = 0.02; Chi = 18.30, df = 7 (P = 0.01); I = 62% Test for overall effect: Z = 3.58 (P = 0.0003) 0.1 0.2 0.5 1 2 5 10 Favours aspirin Favours control

The HOT study did not report hemorrhagic stroke. The JPAD and POPADAD studies did not report major bleeding and BDT did not report gastrointestinal bleeding.

Figure 4 Relative risk of hemorrhagic stroke, major bleeding, and gastrointestinal bleeding in participants treated with and without aspirin.

placebo-treated patients and the relative benets of aspirin compared with placebo or no aspirin were fairly consistent over time.

CONCLUSIONS
Our results demonstrate a consistent pattern of reduced mortality in all of the aspirin primary prevention trials and a signicant,

albeit modest, reduction in all-cause mortality when the data are pooled. This reduction in all-cause mortality tilts the balance between the benets and risks of treatment in favor of the use of aspirin. The results of our meta-analysis are in agreement with those of the 2009 Antithrombotic Trialists Collaboration metaanalysis in secondary prevention and demonstrate that aspirin also is benecial in the primary prevention population.

Raju et al

Effect of Aspirin on Mortality

629
with diabetes and asymptomatic peripheral arterial disease. BMJ. 2008;337:a1840. Ogawa H, Nakayama M, Morimoto T, et al. Low-dose aspirin for primary prevention of atherosclerotic events in patients with type 2 diabetes: a randomized controlled trial. JAMA. 2008;300:2134-2141. Fowkes FG, Price JF, Stewart MC, et al. Aspirin for prevention of cardiovascular events in a general population screened for a low ankle brachial index: a randomized controlled trial. JAMA. 2010;303:841-848. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. Ann Intern Med. 2009;151:W65-94. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med. 2009;151:264-269, W64. Higgins JPT GS, eds. Assessment of study quality. Cochrane Handbook for Systematic Reviews of Interventions 4.2.6 [updated September 2006];Section 6. Available at: http://www.cochrane.org/resources/ handbook/hbook.htm. Accessed March 6, 2009. Peto R, Gray R, Collins R, et al. Randomised trial of prophylactic daily aspirin in British male doctors. Br Med J (Clin Res Ed). 1988;296: 313-316. Final report on the aspirin component of the ongoing Physicians Health Study. Steering Committee of the Physicians Health Study Research Group. N Engl J Med. 1989;321:129-135. Thrombosis Prevention Trial: randomised trial of low-intensity oral anticoagulation with warfarin and low-dose aspirin in the primary prevention of ischaemic heart disease in men at increased risk. The Medical Research Councils General Practice Research Framework. Lancet. 1998;351:233-241. Ridker PM, Cook NR, Lee IM, et al. A randomized trial of low-dose aspirin in the primary prevention of cardiovascular disease in women. N Engl J Med. 2005;352:1293-1304. Hansson L, Zanchetti A, Carruthers SG, et al. Effects of intensive bloodpressure lowering and low-dose aspirin in patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT) randomised trial. HOT Study Group. Lancet. 1998;351:1755-1762. de Gaetano G. Low-dose aspirin and vitamin E in people at cardiovascular risk: a randomised trial in general practice. Collaborative Group of the Primary Prevention Project. Lancet. 2001;357:89-95. Bassler D, Briel M, Montori VM, et al. Stopping randomized trials early for benet and estimation of treatment effects: systematic review and meta-regression analysis. JAMA. 2010;303:1180-1187. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088-1101. Nelson M, Reid C, Beilin L, et al. Rationale for a trial of low-dose aspirin for the primary prevention of major adverse cardiovascular events and vascular dementia in the elderly: Aspirin in Reducing Events in the Elderly (ASPREE). Drugs Aging. 2003;20:897-903. De Berardis G, Sacco M, Evangelista V, et al. Aspirin and Simvastatin Combination for Cardiovascular Events Prevention Trial in Diabetes (ACCEPT-D): design of a randomized study of the efcacy of lowdose aspirin in the prevention of cardiovascular events in subjects with diabetes mellitus treated with statins. Trials. 2007;8:21. A study of cardiovascular events in diabetes (ASCEND). Available at: http://www.ctsu.ox.ac.uk/ascend. Accessed May 12, 2010. Japanese primary prevention project with aspirin in the elderly with one or more risk factors of vascular events: JPPP. http://clinicaltrials. gov/ct2/show/NCT00225849. Accessed May 12, 2010. Chan AT, Ogino S, Fuchs CS. Aspirin use and survival after diagnosis of colorectal cancer. JAMA. 2009;302:649-658. Holmes MD, Chen WY, Li L, Hertzmark E, Spiegelman D, Hankinson SE. Aspirin intake and survival after breast cancer. J Clin Oncol. 28:1467-1472.

References
1. Yusuf S, Reddy S, Ounpuu S, Anand S. Global burden of cardiovascular diseases: part I: general considerations, the epidemiologic transition, risk factors, and impact of urbanization. Circulation. 2001;104: 2746-2753. 2. 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. 3. Hart RG, Halperin JL, McBride R, Benavente O, Man-Son-Hing M, Kronmal RA. Aspirin for the primary prevention of stroke and other major vascular events: meta-analysis and hypotheses. Arch Neurol. 2000;57:326-332. 4. Sanmuganathan PS, Ghahramani P, Jackson PR, Wallis EJ, Ramsay LE. Aspirin for primary prevention of coronary heart disease: safety and absolute benet related to coronary risk derived from meta-analysis of randomised trials. Heart. 2001;85:265-271. 5. Hayden M, Pignone M, Phillips C, Mulrow C. Aspirin for the primary prevention of cardiovascular events: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med. 2002;136:161-172. 6. Bredie SJ, Wollersheim H, Verheugt FW, Thien T. Low-dose aspirin for primary prevention of cardiovascular disease. Semin Vasc Med. 2003;3:177-184. 7. Eidelman RS, Hebert PR, Weisman SM, Hennekens CH. An update on aspirin in the primary prevention of cardiovascular disease. Arch Intern Med. 2003;163:2006-2010. 8. Bartolucci AA, Howard G. Meta-analysis of data from the six primary prevention trials of cardiovascular events using aspirin. Am J Cardiol. 2006;98:746-750. 9. Berger JS, Roncaglioni MC, Avanzini F, Pangrazzi I, Tognoni G, Brown DL. Aspirin for the primary prevention of cardiovascular events in women and men: a sex-specic meta-analysis of randomized controlled trials. JAMA. 2006;295:306-313. 10. Baigent C, Blackwell L, Collins R, et al. Aspirin in the primary and secondary prevention of vascular disease: collaborative meta-analysis of individual participant data from randomised trials. Lancet. 2009; 373:1849-1860. 11. Pearson TA, Blair SN, Daniels SR, et al. AHA Guidelines for Primary Prevention of Cardiovascular Disease and Stroke: 2002 Update: Consensus Panel Guide to Comprehensive Risk Reduction for Adult Patients Without Coronary or Other Atherosclerotic Vascular Diseases. American Heart Association Science Advisory and Coordinating Committee. Circulation. 2002;106:388-391. 12. Buse JB, Ginsberg HN, Bakris GL, et al. Primary prevention of cardiovascular diseases in people with diabetes mellitus: a scientic statement from the American Heart Association and the American Diabetes Association. Circulation. 2007;115:114-126. 13. Graham I, Atar D, Borch-Johnsen K, et al. European guidelines on cardiovascular disease prevention in clinical practice: executive summary. Fourth Joint Task Force of the European Society of Cardiology and other societies on cardiovascular disease prevention in clinical practice (constituted by representatives of nine societies and by invited experts). Eur J Cardiovasc Prev Rehabil. 2007;14(Suppl 2):E1-40. 14. Patrono C, Baigent C, Hirsh J, Roth G. Antiplatelet drugs: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest. 2008;133(6 Suppl):199S-233S. 15. Aspirin for the prevention of cardiovascular disease. U.S. Preventive Services Task Force recommendation statement. Ann Intern Med. 2009;150:396-404. 16. JBS 2. Joint British Societies guidelines on prevention of cardiovascular disease in clinical practice. Heart. 2005;91(Suppl 5):v1-52. 17. Belch J, MacCuish A, Campbell I, et al. The Prevention of Progression of Arterial Disease and Diabetes (POPADAD) trial: factorial randomised placebo controlled trial of aspirin and antioxidants in patients 18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30. 31.

32.

33. 34.

35. 36.