The correlation was examined between the angina-producing arteries predicted to be responsible for the sites of ST segment elevation during spontaneous ischemic attacks and the arteries in which spasm was induced by intracoronary injection of either acetylcholine or ergonovine in patients with variant angina. From 1991 January to 1996 June, 42 patients with variant angina, 40 men and 2 women with a mean age of 61.6 years old, underwent the acetylcholine provocation test within 2 weeks of observation of the last ST segment elevation. After discontinuation of antianginal agents for at least 24 hours, a bolus of acetylcholine was injected in incremental doses of 20, 50 micrograms (occasionally 80 micrograms) into the right coronary artery and of 20, 50 and 100 micrograms incrementally into the left coronary artery to provoke coronary spasm. Intracoronary injection of ergonovine was added in nine patients, in whom intracoronary injection of acetylcholine failed to document coronary spasm on the arteries predicted to be responsible for the sites of ST segment elevation during anginal attacks. Ergonovine was injected in total doses of 40 micrograms into the right coronary artery and of 64 micrograms into the left coronary artery. Positive spasm was defined as induction of more than 99% reversible stenosis. The correlation between the arteries predicted to be responsible for the sites of ST segment elevation during attacks and the vessels in which spasm was induced by acetylcholine test was 78.6% for all patients and 80.0% for all sites of ST segment elevation. By adding the ergonovine test after the acetylcholine test, the correlation increased to 95.2% for all patients and 95.6% for all sites of ST segment elevation. The correlation observed agrees with previous studies in which the ergonovine test was performed in patients without induced spasm by intracoronary injection of acetylcholine and that the super-imposed ergonovine test is useful for diagnosing patients with variant angina.