The risk of administering thrombolysis to patients with suspected myocardial infarction who subsequently do not sustain an infarct, but develop complications associated with thrombolysis is of concern to all physicians. The objective of this study was to ascertain the effect of altering the criteria for the administration of thrombolysis on the number of patients who received thrombolysis in the absence of infarction. During 1990 and 1992 details of all admissions with chest pain were recorded. During 1991 the policy for the administration of thrombolysis was altered so that only patients with ST elevation were eligible. A total of 1473 patients were admitted with chest pain in 1990 and 1967 in 1992. Of the patients admitted in 1990, 663 (45%) had confirmed infarction of whom 378 (57.0%) received thrombolysis. In 1992, 855 (43%) were admitted with infarction and of these 450 (52.6%) had thrombolytic therapy. 118 patients had no evidence of myocardial infarction, but received thrombolysis. 91 (77.1%) were admitted in 1990 and 27 (22.9%) in 1992 (P < 0.01). Of these only 24 (20%) subjects had ST elevation or bundle branch block on the admission electrocardiograph and 41 (35%) had normal tracings. Four (3%) subjects had serious complications of whom one (0.8%) died. The implementation of ECG criteria resulted in a significant reduction in the number of patients without infarction who received thrombolysis, but did not significantly alter the rate of thrombolysis in those with definite myocardial infarction.