The limitations of the Primary Prevention Program (PPP) diabetes substudy have been previously discussed in the original article (1) and in the accompanying editorial (2). Nevertheless, the comments of Dr. de Gaetano (3) require some important additional considerations.
1) The results of the substudy are not the outcome of a post hoc analysis, since an oversampling of diabetic patients was planned before the study started to specifically explore the role of aspirin in these patients.
2) Dr. de Gaetano suggests that the lower-than-expected effect of aspirin could have also emerged if other subgroups, namely patients with hypertension or hypercholesterolemia, would have been examined. This is not the case—aspirin was effective in reducing the risk for total cardiovascular events in both subgroups (OR 0.66, 95% CI 0.50–0.86 and 0.75, 0.52–1.09 for patients with hypertension and hypercholesterolemia, respectively). The benefit of aspirin in these subgroups was even greater after the exclusion of patients with diabetes (0.57, 0.41–0.80 and 0.65, 0.43–1.00, respectively).
3) The lower-than-expected effect of aspirin in individuals with diabetes was consistently found across the whole spectrum of cardiovascular end points considered, and it is highly unlikely that this coherence can be simply due to the play of chance.
4) The results of the PPP trial should not be considered in isolation but evaluated in the context of the existing evidence, which is surprisingly scant. A recent meta-analysis documented a significant effect of antiplatelet therapy in a broad range of high-risk subgroups but failed to show a clear benefit in diabetic patients, with a nonsignificant 7% proportional reduction in serious vascular events (4). Within the meta-analysis, results relative to aspirin were mainly derived from the Early Treatment Diabetic Retinopathy Study (ETDRS), the only one specifically conducted in 3,711 diabetic patients (5). In this trial, treatment with aspirin for an average of 5 years was associated with a nonsignificant 9% reduction in serious vascular events (vascular death, nonfatal myocardial infarction, or nonfatal stroke). Our data are highly consistent with the existing evidence, showing a nonsignificant 10% reduction in the risk of the same end point, as compared with a 41% reduction in nondiabetic individuals.
5) There is a general consensus that primary prevention should be recommended on the basis of the overall cardiovascular risk of an individual patient, rather than on the presence of specific risk factors. To this respect, patients at high cardiovascular risk are by definition heterogeneous, since they often carry several risk factors at the same time, and we cannot see any contradiction in stating that aspirin is effective in a broad range of high-risk patients, with presumably the only exception of individuals with diabetes.
6) Aspirin resistance is only one of the possible explanations for the lower efficacy of aspirin in individuals with diabetes. To our knowledge, the problem of aspirin resistance in the presence of diabetes has never been adequately addressed. To this respect, it seems to us a serious hypothesis to explore, rather than a fashionable issue.
7) We have clearly stated that our data cannot be considered as a conclusive proof against the use of aspirin in patients with diabetes. We believe that the main merit of our study was simply to raise the problem—before the data were published, it seemed that a general consensus was present about the efficacy of aspirin for the primary prevention of cardiovascular events in diabetes. It is now clear, and also Dr. de Gaetano seems to agree, that additional, large-scale trials are needed. It should also be considered that the vast majority of diabetic patients are already treated with ACE inhibitors and/or statins. Whether aspirin adds any benefit in these individuals remains to be proved.
