The effect of trimetazidine was evaluated in patients with stable angina by adding it to the other antianginal regimen in a double blind crossover design each of 8 week's duration. The method of evaluation made use of symptom recall, daily dairy of the intake of sublingual nitrates or of anginal discomfort and in some, symptom limited treadmill exercise stress test (EST). Thirty-six patients completed the trial. Symptom-wise, 16 patients could not differentiate the effect of the true tablet from the placebo. Eight had less and 12 had more angina while on the drug. Of the 17 evaluable EST, 9 showed no change in the degree of ischemic changes while 4 performed with less and 4 with more ischemia while on the drug. Symptom-wise and taking into account the pre and post trial periods, a placebo effect was not found to be dominant. It is concluded that trimetazidine does not improve angina among those already being treated with conventional doses of nitrates, beta and calcium blockers.