The efficacy and safety of oral nifedipine and diltiazem were compared in 20 patients with stable angina pectoris with use of a placebo run-in, randomized, double-blind titration to maximal effect crossover protocol. The effects of treatment withdrawal were also analyzed. All patients received placebo for 2 weeks and were then randomly assigned to receive either diltiazem or nifedipine. A 2 week drug titration phase in which patients received either diltiazem (180 to 360 mg/day) or nifedipine (30 to 120 mg/day) in three divided doses was followed by a 1 week maintenance phase. Patients then received placebo for 1 to 2 weeks, followed by crossover to the other treatment regimen and a second placebo washout period of 1 week. Patients (n = 13) who remained symptomatic on both diltiazem and nifedipine during the monotherapy periods entered a 3 week combination treatment phase, followed by a final 1 week placebo washout period. Frequency of angina, nitroglycerin consumption, exercise tolerance (Naughton protocol), and frequency of daily episodes of ST segment deviations on the electrocardiogram (1 mm of ST segment depression persisting for at least 1 min with and without chest pain) on an ambulatory electrocardiographic monitor were assessed during the baseline placebo, active monotherapy, placebo withdrawal, and combination treatment phases. Plasma drug levels were also measured. Compared with initial placebo values, the frequency of angina and the amount of nitroglycerin treatment were reduced by both diltiazem (p less than .001) and nifedipine (p less than .02). Diltiazem was more effective than nifedipine in reducing angina (p less than .02). Exercise duration increased with both drugs (p less than .0001). Diltiazem was significantly better than nifedipine in reducing the episodes of ST segment depression on the ambulatory monitor (p less than .01). Diltiazem reduced the resting heart rate (p less than .01); both drugs reduced the resting blood pressure and rate-pressure product. Overall, combination therapy was more effective in patients who did not maximally respond to diltiazem or nifedipine alone with respect to anginal and exercise variables and in reducing blood pressure at rest and during exercise. Plasma drug levels could not predict an individual patient's treatment response. Diltiazem may increase nifedipine drug levels when the drugs are combined. Fewer side effects were observed with diltiazem than nifedipine; the most side effects were seen with combination treatment. There were no apparent withdrawal effects observed with either treatment regimen.(ABSTRACT TRUNCATED AT 400 WORDS)