We tested the hypothesis that lesion rethrombosis after streptokinase reperfusion is related to luminal size of the residual stenosis. Two independent techniques of analyzing coronary angiograms, quantitative coronary angiography and computer-based videodensitometry, were used to estimate the size of the residual lumen immediately after discontinuation of streptokinase. These techniques were selected because they provide independent estimates of cross-sectional area of a lesion with high degrees of reproducibility and minimal observer variability. Twenty-four patients who had undergone successful reperfusion with streptokinase were studied. Seven patients had lesion rethrombosis documented either on a repeat angiogram, at autopsy, or, in one case, by the fact that the patient had an acute transmural infarction resulting in death. Vessel patency was documented by repeat coronary angiography 8 to 14'days after initial streptokinase reperfusion in the other 17 patients. As assessed by quantitative coronary angiography, seven of 13 patients (54%) with minimal luminal crosssectional areas of less than 0.4 mm2 had rethrombosis. None of the 1 1 patients with lumens greater than 0.4 mm2 had rethrombosis. In the 17 patients with vessels that remained patent the size of the residual lesion at repeat catheterization was compared with its size immediately after reperfusion with streptokinase. Over the intervening 8 to 14 day interval, an average percentage increase in minimal crosssectional area of 116 + 34% was observed. In seven patients minimal luminal cross-sectional area more than doubled. Integrated optical density, an index of the severity of coronary stenosis derived from computer-based videodensitometry, was also useful in identifying a subgroup of patients at high risk for rethrombosis of lesion. Sixteen patients were identified as having integrated optical densities less than 2.5, and seven (44%) of these had rethrombosis of their lesions. Among the eight patients with integrated optical densities greater than 2.5, none had rethrombosis. These results show that rethrombosis of the vessel is in part related to the size of the residual lesion immediately after reperfusion with streptokinase. Vessels with residual stenotic cross-sectional areas less than 0.4 mm2 are at high risk for rethrombosis whereas vessels with minimal cross-sectional areas of greater than 0.4 mm2 are unlikely to develop rethrombosis. Furthermore, residual size of the lumen may change significantly during the 8 to 14 days after reperfusion. These changes may be due to remodeling of a ruptured atherosclerotic plaque, resolution of persistent coronary spasm, or lysis of persistent thrombi.Circulation 69, No. 5, 991-999, 1984. AFTER REPERFUSION WITH STREPTOKINASE residual high-grade stenoses are frequently present at the site of the previous obstruction. ' Decisions regarding the definitive care of these patients are based in part on the appearance of these lesions. In particular, treat-