The effects of intravenous magnesium on exercise-induced angina were examined in 15 patients with variant angina and in 13 patients with stable effort angina and were compared with those of placebo. Symptom-limited bicycle exercise and thallium-201 myocardial scintigraphy were performed after intravenous administration of 0.27 mmol/kg body weight of magnesium sulfate and after placebo on different days. In all patients, serum magnesium levels after administration of magnesium sulfate were about twofold higher than levels after placebo. Exercise-induced angina associated with transient ST segment elevation occurred in 11 patients with variant angina receiving placebo and in only 2 of these patients receiving magnesium (p less than 0.005). On the other hand, exercise-induced angina was not suppressed by magnesium in any patient with stable effort angina. In these patients there was no significant difference in exercise duration after administration of placebo versus after administration of magnesium. The size of the perfusion defect as measured by thallium-201 scintigraphy was significantly less in patients with variant angina receiving magnesium than that in those receiving placebo (p less than 0.001), whereas it was not significantly different in patients with stable effort angina receiving placebo versus magnesium. In conclusion, exercise-induced angina is suppressed by intravenous magnesium in patients with variant angina but not in patients with stable effort angina. This beneficial effect of magnesium in patients with variant angina is most likely due to improvement of regional myocardial blood flow by suppression of coronary artery spasm.