In a randomized pilot study we compared the efficacy of temporary anticoagulation with intravenous heparin sodium to the efficacy of aspirin in preventing cerebral infarction in hospitalized patients with recent (less than 7 days) transient ischemic attacks (TIAs). Fifty-five patients (33 men, 22 women) aged 36-81 (mean 62.7) years met entry criteria and agreed to participate. Symptoms prompting hospitalization were referable to the carotid distribution in 43 patients (34 hemispheric, nine retinal); 12 patients had vertebrobasilar distribution TIAs. Twenty-seven patients received heparin and 28 received aspirin. Patients were treated until surgery or until long-term medical therapy was instituted, 3-9 (mean 5.5) days in the heparin group and 3-15 (mean 5.8) days in the aspirin group. Recurrent TIAs occurred in eight patients given heparin and in seven treated with aspirin. Infarction occurred in one patient in the heparin group and in four patients in the aspirin group (three brain, one retinal infarction). Initial symptoms in these five patients were referable to the carotid distribution in two and to the vertebrobasilar distribution in three. All patients but the one with a retinal infarction had recurrent TIAs prior to stroke. Our pilot study suggests that hospitalized patients with recent TIAs are at high risk for recurrent TIAs (15 of 55, 27%) and brain infarction (five of 55, 9%) and that patients with recent vertebrobasilar distribution TIAs have a marginally significantly higher risk (odds ratio 6.83, 95% confidence interval 0.65-88.66) of infarction than patients with recent carotid distribution TIAs.