BACKGROUND: Results from a retrospective analysis of the Studies of Left Ventricular Dysfunction (SOLVD) study suggest that angiotensin-converting-enzyme (ACE) inhibitors may be less effective in patients receiving aspirin. We aimed to confirm or refute this theory. METHODS: We used the Peto-Yusuf method to undertake a systematic overview of data for 22060 patients from six long-term randomised trials of ACE inhibitors to assess whether aspirin altered the effects of ACE inhibitor therapy on major clinical outcomes (composite of death, myocardial infarction, stroke, hospital admission for congestive heart failure, or revascularisation). FINDINGS: Baseline characteristics, and prognosis in patients allocated placebo, differed strikingly between those who were and were not taking aspirin at baseline. Results from analyses of all trials, except SOLVD, did not suggest any significant differences between the proportional reductions in risk with ACE inhibitor therapy in the presence or absence of aspirin for the major clinical outcomes (p=0.15), or in any of its individual components, except myocardial infarction (interaction p=0.01). Overall, ACE inhibitor therapy significantly reduced the risk of the major clinical outcomes by 22% (p<0.0001), with clear reductions in risk both among those receiving aspirin at baseline (odds ratio 0.80, [99% CI 0.73-0.88]) and those who were not (0.71 [99% CI 0.62-0.81], interaction p=0.07). INTERPRETATION: Considering the totality of evidence on all major vascular outcomes in these trials, there is only weak evidence of any reduction in the benefit of ACE-inhibitor therapy when added to aspirin. However, there is definite evidence of clinically important benefits with respect to these major clinical outcomes with ACE-inhibitor therapy, irrespective of whether concomitant aspirin is used.


Journal article



Publication Date





1037 - 1043


Aged, Analysis of Variance, Angiotensin-Converting Enzyme Inhibitors, Anticoagulants, Aspirin, Drug Therapy, Combination, Female, Heart Failure, Humans, Logistic Models, Male, Middle Aged, Myocardial Infarction, Myocardial Revascularization, Patient Admission, Prognosis, Proportional Hazards Models, Prospective Studies, Randomized Controlled Trials as Topic, Retrospective Studies, Risk Factors, Stroke, Treatment Outcome, Ventricular Dysfunction, Left