A multicenter randomized double-blind withdrawal study was conducted to compare the efficacy of nifedipine to that of placebo in vasospastic angina. Following a 2-week single-blind nifedipine baseline period, during which nifedipine was maintained at prestudy levels, 38 patients, 19 taking placebo and 19 continuing nifedipine therapy, either completed a 4-week randomized phase or were prematurely withdrawn because of therapeutic failure. During the randomized phase, an increase in median anginal frequency (2.8 attacks/wk, p less than 0.003) and nitroglycerin usage (0.5 tablets/wk, p less than 0.03) occurred only in the placebo group. The randomized phase was prematurely terminated because of anginal exacerbation in 7 of 19 placebo patients (37%) (only 1 patient receiving nifedipine [p = 0.02] experienced anginal exacerbation). Double-blind therapy was judged effective in 16 patients (84%) receiving nifedipine and in 3 patients (16%) receiving placebo (p less than 0.001). Nifedipine was well tolerated. This study establishes the efficacy of nifedipine in the treatment of variant and validates previous clinical experience.