In The Lancet today, the Antithrombotic Trialists' (ATT) Collaboration reports on an individual participant data meta-analysis of primary prevention with aspirin.1 The ATT investigators should be commended for this labour-intensive work on more than 95 000 individuals from six randomised clinical trials. The main added value of the present report compared with an earlier aggregate data meta-analysis on the same six trials is that it had access to individual participant data,2 which allowed the ATT investigators to estimate the magnitude of several risk factors for selected outcomes. Factors such as age, male sex, diabetes mellitus, and high blood pressure were not only associated with ischaemic complications, but also with bleeding, although somewhat less strongly. The ATT investigators reported that the risk of serious vascular events dropped from 0·57% to 0·51% per year by the use of aspirin, but that it increased the risk of major bleeds from 0·07% to 0·10% per year. They therefore concluded that aspirin is of uncertain value for primary prevention.
A major difference in reporting between the two studies is that the investigators of the aggregate data meta-analysis primarily reported effectiveness data for women and men separately, whereas the present report does not. The ATT investigators do provide these data, but reasoned that the proportional reduction in specific vascular outcomes did not differ significantly between men and women when adjustment for multiple comparisons was made. Hence, they do not differentiate between women and men in their recommendations.
Is this the optimum approach? First, at the level of single comparisons there were several tests for the different effects of aspirin in women and men that were significant. For major coronary events, the p value for heterogeneous effectiveness between the sexes was 0·03, that for first stroke (any type) was 0·004, and that for probable ischaemic stroke was 0·005. The latter two remained significant when the ATT global test for heterogeneity was used. The ATT investigators, however, dismissed these differences with the use of a strict multiple comparisons criterion, whereas published work for multiple comparisons is at least as heterogeneous itself.3—5 Second, the ATT investigators obtained information from secondary prevention trials in which different effects of aspirin between the sexes were virtually absent. We think that one should be very cautious with borrowing of data from patients with clinically manifest disease. We are therefore not willing to give the secondary prevention data as much weight in the assessment of the primary prevention data as the ATT investigators do. Third, incidence rates of myocardial infarction, ischaemic stroke, intracranial haemorrhage, and gastrointestinal bleeding differ between the sexes, which affects the balance between benefits and risks.
How can we use today's study in daily practice? For example, should the first author of this Comment (55-year-old man, slightly elevated cholesterol, no other vascular risk factors) take an aspirin a day to keep the doctor away? The ATT investigators try to balance benefits (less serious vascular events) and risks (more haemorrhages), but without taking into account the consequences of both harms and benefits on quality of life.
We thus decided to update our previous cost-effectiveness analysis6 with the new data provided by the ATT investigators. We have used the slightly different risk ratios for cardiac events and ischaemic stroke for women and men separately. Moreover, in individuals for whom we used an increased risk of cardiovascular disease we let the risk of haemorrhage increase proportionally on the basis of the new ATT observations. The table summarises for which individuals an incremental cost-effectiveness ratio of €20 000 or less would be obtained; these individuals are probably the best candidates for prevention with aspirin. These ratios are the most reliable estimates since all data have been incorporated in the cost-effectiveness analysis, but these ratios come with their own uncertainty (see the sensitivity analysis in table 5 of our previous report).6 The new estimates from the cost-effectiveness analysis imply that the first author does not need to take an aspirin a day. It should be emphasised that our model also took the (dis)utility of taking aspirin into account and our sensitivity analyses indicated that the results of our analyses were sensitive to it. Patients might not wish to be medicalised7—such considerations are important in the decision to take aspirin or not. Whether statins should be preferred above aspirin is a different and difficult question that needs careful consideration too. Apart from drug treatment, one must not forget the importance of lifestyle changes, such as cessation of smoking, healthy diet, and regular exercise.

Recommended aspirin use on basis of updated cost-effectiveness analysis6

References

1 Antithrombotic Trialists' (ATT) Collaboration. 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.
2 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-specific meta-analysis of randomized controlled trials. JAMA 2006; 295: 306-313.
3 Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology 1990; 1: 43-46.
4 Bland JM, Altman DG. Multiple significance tests: the Bonferroni method. BMJ 1995; 310: 170.
5 Perneger TV. What's wrong with Bonferroni adjustments. BMJ 1998; 316: 1236-1238.
6 Greving JP, Buskens E, Koffijberg H, Algra A. Cost-effectiveness of aspirin treatment in the primary prevention of cardiovascular disease events in subgroups based on age, gender and varying cardiovascular risk. Circulation 2008; 117: 2875-2883.
7 Christie B. Guidelines on treating risk factors turn healthy people into patients, doctors say. BMJ 2006; 333: 988.


E' disponibile un set di diapositive tratte dall'articolo:

alert-08-Lancet_2009_373_1821.ppt (222 KB)