The ideal non-invasive method for detecting coronary reperfusion has not yet been established. In 63 patients with acute myocardial infarction, serum myoglobin and creatine kinase-MB were measured every 15 min. Thrombolytic treatment was given (n = 52) and acute coronary angiography showed a patent infarct-related artery in 49 patients while 14 patients had no coronary reperfusion. Median time to peak serum myoglobin was shorter (reperfusion group 178 min vs no reperfusion group 480 min, P < 0.0001) than time to peak serum creatine kinase-MB (reperfusion group 550 min vs no reperfusion group 1080 min, P < 0.0001), P < 0.0001. Myoglobin appearance rate, calculated as the concentration at 2 h divided by baseline values (Mb2/Mb0) was highest in the reperfusion group (4.0 vs 1.6), P < 0.001. An earlier proposed index, Mb2/Mb0 > 2.4 for identification of reperfusion 2 h after thrombolytic therapy, showed predictive values of positive and negative tests of 0.94 and 0.44, respectively. Combining this index with signs of medium to larger infarct size (Mb2 > 200 micrograms.l-1) increased the predictive value of the negative test to 1.00. In patients with signs of minor infarcts (Mb2 < 200 micrograms.l-1) the predictive values of positive and negative tests were 0.94 and 0.79, respectively, 5 h after onset of thrombolytic therapy. An early rise and a peak in serum myoglobin values seems to be a reliable and simple non-invasive indicator of successful and unsuccessful reperfusion therapy.