Nine patients with severe coronary artery disease and disabling angina receiving either regular metoprolol or oxprenolol, together with glyceryl trinitrate tablets as required for chest pain, were studied. Nifedipine 10 mg three times per day was compared to placebo in a double blind randomized control trial, using patient diary cards and exercise tests. The number of recorded episodes of angina during the placebo period of 15.0 +/- 2.1 (mean +/- s.e.m.) per patient per week was significantly reduced to 11.2 +/- 2.5 during the nifedipine period (P less than 0.05). Similarly, the recorded number of glyceryl trinitrate tablets consumed during the placebo period of 12.6 +/- 2.1 was significantly reduced to 9.1 +/- 2.0 tablets per patient per week during the nifedipine period (P less than 0.05). There was a statistically significant increase in both the duration of exercise to onset of chest pain (from 241 +/- 16.3 seconds on placebo to 306 +/- 38.4 seconds on nifedipine (P less than 0.05)) and the total work performed to the onset of chest pain during the nifedipine period compared to the control and placebo periods (P less than 0.05). There was a significant increase (P less than 0.05) in exercise time before the onset of appreciable ST depression (greater than 1 mm) on exercise testing during the nifedipine period (66.2 +/- 4.2 sec) compared to the control period (51.2 +/- 3.0 sec) and placebo periods (58.7 +/- 3.5 sec). Although nifedipine was generally well tolerated, one patient experienced a severe episode of angina whilst taking the drug, which required admission to hospital.