OBJECTIVE: To evaluate the feasibility of prehospital thrombolysis in Sweden in terms of safety and to examine the various components of the delay between onset of symptoms and start of treatment. SETTING: A total of 16 hospitals in Sweden in both urban and less populated areas and the associated ambulance organisations. DESIGN: Prospective evaluation of patients with an ST-elevation infarction treated with reteplase. An ECG was recorded and transmitted to hospital. The ambulances were staffed by a physician in 1% of cases, a nurse in 67% and a staff nurse in 32%. RESULTS: Of the 148 patients who received treatment prior to hospital admission, six (4%) had a cardiac arrest prior to hospital admission and two (1%) died prior to arrival at hospital. One patient was given treatment despite an exclusion criterion (previous stroke) and died on the 1st day in hospital due to a cerebral haemorrhage. The overall 30-day mortality was 7.1% and 1-year mortality 9.8%. Treatment was initiated within 2 h after the onset of symptoms in 53% of patients and within 1 h in 17% of patients. The median interval between the arrival of the ambulance and sending an ECG was 13 min and the median interval between sending an ECG and the start of thrombolysis was 18 min. The delay was similar regardless of ambulance staff. CONCLUSION: Implementation of prehospital thrombolysis on a national basis in Sweden appears to be safe. More than half the patients can be given treatment less than 2 h after the onset of symptoms. There is potential for reducing this time still further.