OBJECTIVES--To evaluate the role of Holter ST monitoring for identifying patients at risk of recurrent ischaemic events after acute myocardial infarction treated by thrombolysis. BACKGROUND--The natural history of myocardial infarction has changed with the introduction of thrombolytic treatment. There is now a lower mortality but a higher incidence of recurrent thrombotic events (reinfarction, unstable angina). Preliminary evidence indicates that Holter ST monitoring may be of prognostic value in patients with acute myocardial infarction, but there are limited data available in patients treated by thrombolysis. METHODS--Prospective observational study of 256 consecutive patients who presented with acute myocardial infarction treated by thrombolysis. All underwent 48 hour Holter ST monitoring early after thrombolysis (mean 83, range 48-180 hours) and were followed up for eight (range three to 12) months. RESULTS--Recurrent ischaemic events occurred in 45 patients (fatal reinfarction 17, non-fatal reinfarction 12, unstable angina 16). Also four patients died as a result of progressive heart failure, and a further 15 patients required revascularisation. Analysis of the Holter data showed that 32% of patients had at least one episode of isolated ST depression (> or = 0.1 mV) and 41% either ST depression or elevation (> or = 0.2 mV). Ischaemic episodes were silent in 95% of cases. Event free survival analysis showed a significant association between Holter findings and recurrent ischaemic events (ST depression: p = 0.009; ST depression or elevation: p = 0.002). The association remained significant when the end point was restricted to fatal or non-fatal reinfarction (ST depression: p = 0.005; ST depression or elevation p = 0.001), the period of greatest risk for patients with an abnormal recording occurred early after investigation. An abnormal Holter recording identified patients at risk of early (within 30 days) reinfarction with 79% sensitivity and 60% specificity. Although positive predictive accuracy was low (11%), a normal Holter recording was associated with 98% negative predictive accuracy. CONCLUSIONS--In patients treated by thrombolysis, ST change on Holter monitoring may be useful for identifying patients at increased risk of recurrent ischaemic events, and in particular early reinfarction.