In previously thrombolysed patients, we analysed residual myocardial viability using the PET-FDG technique and correlated its presence and extent to the angiographic appearance of the infarct-related vessel and left ventricular function. Thirty-six patients who had undergone intravenous thromboloysis for acute myocardial infarction 4.8 +/- 7.2 months previously were studied. Coronary angiography, left ventriculography, and assessment of myocardial perfusion and metabolism were all performed within 1 week. All patients exhibited perfusion defects consistent with the clinically identified myocardial infarction site. Residual viability, as assessed by the PET-FDG technique, was present in 53% of cases. The infarct-related coronary artery was patent in 19 (53%) patients (TIMI grade 3, 79%); of the remaining 17 with occluded infarct-related arteries, 11 had collaterals to the infarct area. Significant FDG uptake was observed in 63% of patients with a patent infarct-related artery and in 41% of those with an occluded infarct-related artery. The same study protocol was adopted in a control group of 30 patients with myocardial infarction who did not receive thrombolysis. The number of infarct-related patent vessels was significantly lower in these patients (30 vs 53%) (TIMI grade 3, 56%), but the overall percentage of PET viability was again 53%. Qualitative analysis of the regional perfusion pattern showed that the magnitude and severity of the perfusion defect was similar in the two groups, regardless of the presence or absence of FDG uptake. Global left ventricular function was also similar in the two groups. However, regional wall motion was significantly better in the thrombolysed patients with a patent infarct-related artery than in those who had not received thrombolysis and whose culprit vessel was also patent. In conclusion, the results of our study support the notion that early recanalization of the infarct-related artery is critical for preserving left ventricular function. Although the number of patent infarct-related coronary arteries is greater and left ventricular function is better in successfully thrombolysed patients, the regional metabolic pattern does not apparently correlate with the patency of the infarct-related artery. This suggests that, in "chronic' myocardial infarction, residual tissue viability as assessed by fluorodeoxyglucose uptake does not necessarily correlate with coronary recanalization.