The effect of 5000 U of heparin given subcutaneously three times daily on the frequency of venous thrombosis detected by 125-I-fibrinogen scanning was evaluated in a prospective randomized study of 350 high-risk surgical and medical patients. Heparin treatment reduced the rate of venous thrombosis from 19 of 118 patients (16.1 per cent) to two of 108 (1.9 per cent) treated after elective surgery, from 11 of 23 untreated (48 per cent) to three of 23 (13 per cent) treated after hip fracture, and from nine of 40 (22.5 per cent) to one of 38 (2.6 per cent) treated patients admitted to the hospital for suspected myocardial infarction. Thrombosis occurred in 21.5 per cent of the total of 181 untreated and in 3.6 per cent of 169 heparin-treated patients. The activated partial thromboplastin time was moderately increased for five hours after subcutaneous heparin injections. Clinically important bleeding was not increased in heparin-treated surgical patients, though the blood requirement of transfused patients was moderately augmented, and treated patients had a slightly lower postoperative hematocrit.