The incidence of deep venous thrombosis (DVT) and the efficacy of prophylactic measures were prospectively evaluated in all patients admitted to a level I trauma center during 1991. Patients with Injury Severity Scores (ISS) > 9 who survived a minimum of 48 hours (n = 395) were monitored using venous Doppler and ultrasound studies during hospitalization (total, 1308 studies). Two hundred eighty-one patients (71%) were randomly assigned to low-dose heparin or sequential compression devices. There were 18 cases of lower extremity DVT (4.6%) and four cases (1.0%) of pulmonary emboli (PE), three of which were fatal. Eight patients (2.9%) on prophylaxis and 10 (8.8%) without prophylaxis developed DVT (p < 0.02 by Chi-square). There were two PEs in each group. Fourteen of these 18 patients sustained blunt trauma and included seven spinal fractures or subluxations (four paraplegic) and four severe head injuries. This represented 14.0% of 50 patients admitted with spinal injuries and 4.3% of 92 patients with severe head injuries. Compared with those with no neurologic injury (7 of 253 or 2.7%), the risk of DVT is significantly higher in the spinal injury patients (p < 0.001, Chi-square) and twice as high as in the head injury group, although not statistically significant (p = 0.4, Chi-square). Three of the four patients with penetrating trauma and DVT had venous injuries. We conclude that DVT prophylaxis can significantly reduce the incidence of DVT in trauma patients with ISS > 9. Patients with severe neurologic injuries (particularly spinal cord) are at high risk for DVT and PE and may be considered for a prophylactic Greenfield filter.