This post hoc analysis of the Warfarin Re-Infarction Study evaluates the effect of long-term anticoagulant therapy in different subgroups after acute myocardial infarction (MI). The study population comprised 1214 patients. The mean duration of treatment was 37 months. The overall significance of prespecified prognostic factors was assessed by univariate survival analyses. Those risk factors that yielded a statistically significant result were evaluated with regard to response to treatment in a stratified manner. After startification, heterogeneity across the strata was found to pertain to the effect of treatment with warfarin in subjects with prior MI and diabetes mellitus. Hence, mortality was not found to be influenced favorably by warfarin therapy in patients with previous MI. Likewise, recurrent MI was not significantly reduced by warfarin therapy in patients with prior MI or diabetes mellitus. Although not statistically significant, increasing age was associated with less benefit from therapy. The findings persisted also after controlling for possible confounders in a Cox regression model. Thus, our data suggest a lack of a beneficial effect by warfarin therapy in subjects with prior MI or diabetes mellitus, when the therapy is given for the sole purpose of secondary prophylaxis of MI. Furthermore, a trend toward an attenuated effect of therapy was found among the oldest patients.

(Arch Intern Med. 1992;152:993-997)