In a multicenter trial of aspirin or sulfinpyrazone in the treatment of unstable angina, we examined the possible importance to the outcome of mentioning potential side effects in the consent form. Inclusion, in two of the three centers, of a statement outlining possible gastrointestinal side effects resulted in a sixfold increase (P less than 0.001) in the number of subjects in these centers withdrawing from the study because of subjective, minor gastrointestinal symptoms. Major gastrointestinal complications such as peptic ulcer or bleeding as diagnosed by study physicians were similar in the three centers. Furthermore, no patient discontinued therapy because of subjective, nongastrointestinal side effects. Post hoc analysis suggests that the inclusion of gastrointestinal side effects in the consent form may have increased the likelihood of patients attributing gastrointestinal symptoms to drug therapy, leading to subsequent withdrawal from the study.