OBJECTIVE--To compare the activity of plasminogen activator inhibitor (PAI-1) in diabetic and non-diabetic patients admitted with acute myocardial infarction and to determine whether PAI-1 activity influences reperfusion after thrombolytic therapy. DESIGN--Prospective study of patients admitted with acute myocardial infarction. SETTING--District general hospital. MAIN OUTCOME MEASURES--Reperfusion assessed by time to peak release of creatine kinase-MB isoenzyme. RESULTS--Baseline PAI-1 activity and antigen concentrations were significantly higher in diabetic patients (n = 45) than in non-diabetic patients (n = 110) (24.6 (6.9) v 18.6 (7.9) AU/ml (AU = arbitrary units) (p = 0.0001) and 58.8 (13.1-328.8) v 41.0 (10.9-125.4) ng/ml (p = 0.004). Time to peak release of creatine kinase-MB was calculated in 123 (80%) patients. In 98 who received thrombolytic therapy the median time to peak enzyme release was 15.5 h (7.5-24 h) in diabetic patients (n = 26) and 12 h (5-26 h) in non-diabetic patients (n = 72) (p = 0.005). In those with a time to peak release of < or = 12 h, indicating likely successful reperfusion, PAI-1 activity was 17.5 (7.8) AU/ml compared with 22.8 (7.7) AU/ml in those with a time to peak release of > 12 h (p = 0.001). In multiple regression analysis both diabetes (p = 0.0001) and PAI-1 activity at admission (p = 0.029) were independently related to successful reperfusion. In 13 patients with evidence of reinfarction in hospital PAI-1 activity on day 3 was 26.7 (6.4) AU/ml compared with 21.7 (6.3) AU/ml in those without evidence of reinfarction (p = 0.032). CONCLUSION--Both raised PAI-1 activity on admission and diabetes were associated with a reduced likelihood of enzymatic evidence of reperfusion after thrombolytic therapy. Increased PAI-1 activity on day 3 was associated with an increased risk of reinfarction. Diabetic patients had higher PAI-1 activity on admission. This may partly explain their reduced likelihood of reperfusion.