We studied the effect of propranolol administration on risk assessment based on submaximal exercise testing performed early after myocardial infarction. A total of 70 patients with recent infarction underwent modified Bruce treadmill testing with simultaneous measurement of expired gases in the absence of antianginal agents including beta-antagonists. Among these, 31 patients who had at least one of the following abnormalities--ST depression greater than or equal to 1 mm (22 patients), chest pain (four patients), or treadmill time less than 360 seconds (12 patients)--were studied in a randomized double-blind fashion and received either placebo or 240 mg of propranolol/day. A total of 28 patients completed the randomized phase and were able to undergo repeat exercise testing an average of 3.4 +/- 1.8 days later. Randomized groups were equivalent at baseline except for a higher peak oxygen consumption and carbon dioxide production (p less than 0.05) in the propranolol compared with the placebo group; these differences were taken into account in statistical analyses of the study data. Resting heart rate (59 +/- 1.2 versus 82 +/- 4.2 beats/min) and peak heart rate x systolic blood pressure (14,208 +/- 496 versus 20,075 +/- 1,062) were both significantly less (p less than 0.01) after propranolol than after placebo. Eight of nine patients treated with placebo maintained ST depression greater than or equal to 1 mm from the initial to the randomized exercise test, compared with only 4 of 13 receiving propranolol (p less than 0.01). In those with continued ST depression, time to positivity was significantly longer in those receiving propranolol compared with those taking placebo (538 +/- 73 versus 318 +/- 44 seconds, p less than 0.05). In contrast, the peak ratio between carbon dioxide production and oxygen consumption was higher in those receiving propranolol compared with those receiving placebo (0.93 +/- 0.04 versus 0.81 +/- 0.03, p less than 0.05). We conclude that propranolol therapy reduces evidence of ischemia and changes traditional estimates of potential cardiac risk derived from submaximal postinfarction exercise testing.