Constriction of atherosclerotic coronary segments during exercise may further reduce coronary flow reserve in patients with coronary artery disease. This could influence the linear regression analysis of the heart rate-related changes in ST-segment depression (ST/HR slope) thereby limiting the accuracy of this method in identifying the severity of the disease. To test this hypothesis, the exercise related ST/HR slopes on placebo were compared with those obtained during coronary vasodilation induced by a prostacyclin analogue (iloprost 6 ng kg-1 min-1) in 42 anginal patients with documented coronary artery disease. In seven of these, the same protocol was repeated during right heart catheterization. The overall diagnostic accuracy of the ST/HR slope on iloprost was better than on placebo in patients with advanced coronary artery disease. This was due mainly to a consistent rightward shift of the ST/HR slope in patients with one- and two-vessel, but not three-vessel disease or left main stem disease. The reason for the greater effects of iloprost on ST/HR slopes in patients with a lesser degree of atherosclerosis remains unclear. However, coronary blood flow was higher during drug infusion, which suggests that iloprost may prevent the occurrence of dynamic coronary events able to reduce the maximum coronary flow reserve during exertion. This mechanism may be predominant in patients with minor coronary artery disease.