In quantitative analysis of repeated coronary angiograms, a variable vasomotor tone of the epicardial coronary arteries may influence the accuracy of the results. Therefore, we evaluated the extent and reproducibility of coronary artery dilation with nitrocompounds. In 32 patients with coronary artery disease, the vasodilatory response of angiographically normal coronary segments to different nitrocompounds was analyzed with the computer-assisted contour detection system CAAS. Twenty patients received 5 mg or 10 mg of isosorbide dinitrate sublingually. After 10 to 15 min, a maximal diameter increase was measured with an average of 16 +/- 11% (5 mg: P less than 0.01) and 28 +/- 13% (10 mg: P less than 0.001) from control. Another 12 patients received 0.025 mg per kg body weight of SIN-1, the active metabolite of molsidomine, as an intravenous infusion over 5 min. A comparable maximal dilation (29 +/- 5%; P less than 0.001) occurred after 10 to 15 min and could not be enhanced further with 0.8 mg nitroglycerin administered sublingually (28 +/- 7%; n.s.). One hour after square root of Sin-1, coronary dilation was still 24 +/- 8% (P less than 0.001 compared with control), and 0.8 mg of nitroglycerin sublingually reestablished the previous maximal dilation of 28 +/- 8%. We conclude that high doses of nitrocompounds induce a reproducible maximal coronary dilation that eliminates a substantial source of error in quantitative analysis of repeated coronary angiograms. At present, sublingual administrations of either 10 mg isosorbide dinitrate once or 0.8 mg nitroglycerin repeatedly seem to represent the easiest practicable modes to achieve maximal coronary vasodilation for an adequate period.