To examine whether intracoronary injections of acetylcholine induce coronary artery spasm in patients with vasospastic angina, incremental doses (20, 30 and 50 micrograms) were injected directly into the coronary arteries in 12 patients with variant angina (Group A: rest angina with electrocardiographic ST-segment elevation during attacks), 19 with vasospastic angina (Group B: rest angina and/or effort angina with variable threshold in the treadmill exercise stress test), 11 with organic coronary artery stenosis but without angina (Group C), and 14 without coronary artery disease (Group D). A temporary cardiac pacemaker was positioned in the right ventricle. Coronary artery spasm was defined as severe vasoconstriction (greater than or equal to 90% of reduction in the luminal diameter) with chest pain and/or ischemic changes in the electrocardiogram. Intracoronary injection of acetylcholine induced spasm of at least one coronary artery in all 12 patients (100%) of Group A, in 18 (95%) of Group B, in two (18%) of Group C, and in two (14%) of Group D. Thus, the sensitivity of this method for inducing coronary spasm was 100% in group A, 95% in Group B, and 97% in Group A plus Group B. The specificity for inducing spasm was 86% in Group D, and 84% in Group C and Group D. When acetylcholine was injected separately into the left and right coronary arteries, spasm of both the coronary arteries was observed in two (40%) of Group A, in five (33%) of Group B, and none (0%) of Group C and Group D. Acetylcholine (20 micrograms) induced coronary spasm in 10 (83%) of Group A and only in nine (47%) of Group B.(ABSTRACT TRUNCATED AT 250 WORDS)