Sex-related differences in response to aspirin in cardiovascular disease: an untested hypothesis

Charles H Hennekens*, Danielle Hollar and Colin Baigent  About the authors

The recently published aspirin component of the Women's Health Study (WHS) compared low-dose aspirin (100 mg on alternate days) with placebo among 39,876 apparently healthy female healthcare providers for an average duration of around 10 years.¹ Aspirin significantly lowered the risk of a first nonfatal stroke by 19% (relative risk [RR], 0.81, 95% CI 0.67–0.97; P = 0.02), but had no significant effect on the risk of first nonfatal myocardial infarction (MI; RR 1.01, 95% CI 0.83–1.24).¹ This result was surprising because it appeared to contradict the existing evidence from randomized trials of aspirin in primary prevention. A previous meta-analysis of five published primary prevention trials, predominantly among men, had shown that aspirin significantly reduced the occurrence of nonfatal MI by about a third (RR 0.68, 95% CI 0.59–0.79), but had no significant effect on nonfatal stroke (RR 1.06, 95% CI 0.87–1.29).² The editorial accompanying the WHS opined that "The investigators of the Physicians' Health Study and the Women's Health Study have produced two landmark studies on the use of aspirin in asymptomatic persons, the results of which segregate according to sex."³ It is our view, however, that any conclusion that there are sex-related differences in response to aspirin for the prevention of cardiovascular disease in apparently healthy men and women might be premature.

In order to explore the specific effects of aspirin on the coronary and cerebral vasculature, we have analyzed a subset of trials that provided individual patient data for the Antiplatelet Trialists' (APT) Collaboration.⁴ Detailed data from six trials comparing aspirin with placebo among 11,265 patients with a history of MI, and from nine trials comparing aspirin with placebo among 6,765 patients with a history of cerebrovascular disease (i.e. a transient ischemic attack or stroke) were available for analysis. Aspirin reduced the risk of a major coronary event—nonfatal MI or coronary death—to a similar degree in men and women (19% reduction in men [SE = 5] and 25% reduction in women [SE = 12]; ² test for heterogeneity with 1 degree of freedom = 0.3; P = 0.6), and also reduced the risk of stroke (17% reduction in men [SE = 7] and 22% reduction in women [SE = 11]; ² test for heterogeneity with 1 degree of freedom = 0.2; P = 0.7). These findings provide evidence against a clinically important difference in the antiplatelet efficacy of aspirin in men and women with previous vascular disease. Thus, any putative explanation for the apparently disparate findings on coronary heart disease and stroke occurrence in men and women within the six primary prevention trials of aspirin versus placebo would have to account for the lack of any cogent evidence of such differences in the meta-analysis of the secondary prevention trials. Sex-specific differences in salicylate levels after aspirin administration⁵ are an unlikely explanation for these findings, as aspirin irreversibly inhibits platelet thromboxane biosynthesis before it is hydrolyzed in the liver. When compared with development in men, however, atherosclerosis develops later in women and affects cerebral arteries to a greater extent than coronary vessels. So there could be biological differences in the contribution of thromboxane-mediated platelet aggregation in the pathogenesis of atherosclerosis and thrombosis.

While basic research and the findings of new aspirin trials should advance our knowledge, how should healthcare providers respond to the uncertainty in the meantime? The most clinically relevant point is that aspirin should be considered for patients in whom the expected vascular benefits exceed the risks of serious gastrointestinal bleeding. In the context of secondary prevention, where the risk of a serious vascular event generally surpasses 5% per year, the benefits of aspirin clearly outweigh the risks, irrespective of sex.⁴ For primary prevention of vascular disease, however, the average risk of a serious vascular event is much lower, typically less than 1% per year. Since all available randomized evidence suggests that aspirin reliably prevents a first MI, but might have more limited effects on stroke, the clinician's task is to identify the individuals who are at above average risk of coronary heart disease. In practice, this means identifying individuals in whom a risk stratification tool (e.g. the Framingham Risk Score)^{2, 6, 7} predicts an annual rate of coronary heart disease above a certain threshold; however, the US Preventive Services Task Force set the threshold at 0.6% per year,⁶ and the American Heart Association suggest 1.0% per year.⁷ Most women remain at low risk of coronary heart disease until aged approximately 65–70 years, whereupon the annual risk rises rapidly, so that elderly patients (i.e. those aged 70 years or over) form the largest identifiable group of women who might benefit from aspirin. Unfortunately, only around 10% of women enrolled in the WHS were aged 65 years or over at trial entry. Nevertheless, subgroup analyses based on small numbers of events in these elderly women were consistent with the hypothesis that aspirin reduces the risk both of first MI (41 MIs in patients who received aspirin versus 62 who received placebo, RR 0.66, 95% CI 0.44–0.97; P = 0.04) and first ischemic stroke (53 strokes in patients who received aspirin versus 75 who received placebo, RR 0.70 95% CI 0.49–1.00; P = 0.05). The statistical imprecision of these results, however, means that there is substantial uncertainty at this point as to whether aspirin has a negligible or a substantial benefit among elderly women. Hence we cannot compare any possible benefits of aspirin with the probable risk of bleeding, which also increases substantially in older people and is associated with higher case fatality.⁸ Similarly, since only approximately 10% of men in the five previous primary prevention trials were aged 70 years or over, we cannot determine the effects of aspirin among older men with adequate precision; so identical arguments apply to the use of aspirin in older men. In order to resolve the current uncertainty, data are needed from large-scale, randomized trials of aspirin in the elderly. This is planned, for example, in the Aspirin in Reducing Events in the Elderly (ASPREE) trial,⁹ which aims to randomize 15,000 people aged 70 years or over to 100 mg aspirin daily or placebo.

In summary, the hypothesis formulated in the editorial accompanying the aspirin component of the WHS warrants further investigation; however, it is not supported by evidence from randomized trials in secondary prevention, nor is it plausibly explained by differences in the sex-specific pharmacokinetic or pharmacodynamic profile of aspirin. To paraphrase TH Huxley, the idea might yet prove to be another beautiful hypothesis slain by ugly facts,¹⁰ but this will not emerge until we have the results of large-scale, randomized trials of aspirin versus placebo in primary prevention of cardiovascular disease in elderly people.

Competing interests

CH Hennekens is funded by the Department of Biomedical Science and Center of Excellence at Florida Atlantic University (FAU). He is Principal Investigator of two investigator initiated research grants funded by Bayer HealthCare, testing the effects of aspirin dose on platelet biomarkers, inflammatory markers, nitric oxide formation and endothelial function. He is a consultant for Agatston Research Institute, Amgen, AstraZeneca, Bayer, Bristol-Myers Squibb, Chattem, Deaco, the FDA, GlaxoSmithKline, McNeil, the NIH, Novartis, Pfizer, Reliant, TAP and UpToDate. D Hollar receives funding from the Department of Biomedical Science and Center of Excellence at Florida Atlantic University. She is Project Director of two investigator initiated research grants funded by Bayer HealthCare.

ARTICLE NAVIGATION -ARTICLE TOOLS

ISSUE NAVIGATION - THIS ISSUE

• Table of Contents
• Previous article
• Next article