Twelve consecutive patients (10 men and 2 women, mean +/- standard deviation age 49 +/- 9 years) with chest pain, angiographically normal coronary arteries and coronary artery spasm documented by methylergometrine testing received a single oral dose of molsidomine (4 mg) or nifedipine (10 mg) in a randomized, double-blind, crossover fashion at a 24-hour interval. Coronary artery spasm was documented during coronary angiography in 6 patients (left anterior descending artery, 3; right coronary artery, 2; left circumflex, 1). In the remaining 6 patients, coronary artery spasm was documented by a positive methylergometrine test performed at the bedside, which provoked ST-segment elevation in the inferior (n = 3), anterior (n = 1) or lateral (n = 2) leads. Ninety minutes after administration of the study medication, methylergometrine testing was performed at the bedside, using incremental doses of up to 0.4 mg of methylergometrine. After molsidomine, 10 patients (83%) had a negative and 2 had a positive test; after nifedipine, 9 patients (75%) had a negative and 3 a positive test. Only 1 patient had a methylergometrine test that remained positive after either molsidomine or nifedipine. Therefore, molsidomine appears as effective as nifedipine in suppressing methylergometrine-induced coronary artery spasm in patients with variant angina. In addition, patients not responding to 1 of the study medications may respond to the other.