Serial venous duplex scans (VDS) were done in 507 trauma patients with at least one risk factor (RF) for venous thromboembolism (VTE) during a 2-year study period. Deep vein thrombosis (DVT) was detected in 31 (6.1%) patients. This incidence was 3.1 per cent in low (1-2 RFs), 3.4 per cent in moderate (3-5 RFs), and 7.7 per cent in high (> or =6 RFs) VTE scores (P = 0.172). Incidence was statistically different (3% vs. 7.2%, P = 0.048) on reanalyzing patients in two risk categories, low-risk (1-4 RFs) and high-risk (> or =5 RFs). Only 4 of 16 RFs had statistically higher incidence of DVT in patients with or without RFs: previous VTE (27.3% vs. 5.6%, odds ratio (OR) 6.628, P = 0.024), spinal cord injury (22.6% vs. 5%, OR 5.493, P = 0.001), pelvic fractures (11.4% vs. 5.1%, OR 2.373, P = 0.042), and head injury with a greater than two Abbreviated Injury Score (10.5% vs. 4.2%, OR 2.639, P = 0.014). On reanalyzing patients with > or =5 RFs vs. <5RFs, obesity (14.3 vs. 6.1%, P = 0.007), malignancy (5.6% vs. 0.6%, P = 0.006), coagulopathy (10.8% vs. 1.8%, P = 0.000), and previous VTE (3.2% vs. 0%, P = 0.019) were significant on univariate analysis. Patients with DVT had 3.70 +/- 1.75 RFs and a 9.61 +/- 4.93 VTE score, whereas, patients without DVT had 2.66 +/- 1.50 RFs and a 6.83 +/- 3.91 VTE score (P = 0.000). DVTs had a direct positive relationship with higher VTE scores, length of stay, and number of VDS (>1 r, P < or = 0.001). Increasing age was a weak risk factor (0.03 r, P = 0.5). First two VDS diagnosed 77 per cent of DVTs. Patients with injury severity score of > or =15 and 25 had higher DVTs compared with the ones with lower injury severity score levels (P < or = 0.05). Pulmonary embolism was silent in 63 per cent and DVTs were asymptomatic in 68 per cent.