BACKGROUND: The relation between thrombosis and cancer is well documented, but the clinical evaluation appropriate for detecting underlying cancer in patients with deep venous thrombosis remains unknown. OBJECTIVES: To 1) determine the appropriate initial evaluation for cancer in patients with idiopathic deep venous thrombosis and 2) estimate the incidence of subsequently diagnosed cancer in patients who were not found to have cancer when deep venous thrombosis was diagnosed. DESIGN: Retrospective cohort study. SETTING: Urban, tertiary care teaching hospital. PATIENTS: 986 consecutive patients (637 women and 349 men; mean age +/- SD, 53 +/- 17 years) who had no risk factors for venous disease and had venous ultrasonography because idiopathic deep venous thrombosis was suspected. MEASUREMENTS: Initial clinical evaluation was assessed by using chart review. The incidence of cancer at a median of 34 months after diagnosis of deep venous thrombosis was obtained through hospital chart review (38%), computerized record extraction (54%), mailed questionnaires (6%), telephone interviews (1%), or a death registry (1%). RESULTS: Deep venous thrombosis was diagnosed in 142 patients (14%); 136 of the 142 were hospitalized. Cancer was diagnosed in 16 of these 136 patients (12%) during the index hospitalization. All 16 of these patients had one or more abnormalities on at least one of the four components of the clinical examination: history, physical examination, basic laboratory testing, or chest radiography. The probability of detecting cancer increased as the number of findings suggestive of cancer on the four components of the clinical evaluation increased. Cancer was diagnosed in none of the 56 patients with deep venous thrombosis who did not have findings on the clinical evaluation. The probability of cancer-free survival during follow-up (median, 34 months) was similar in patients with (3 of 122 [2.5%]) and without (23 of 844 [2.7%]) deep venous thrombosis and in the age- and sex-matched U.S. population. CONCLUSIONS: A clinical evaluation that includes a comprehensive medical history, physical examination, routine laboratory testing, and chest radiography seems to be appropriate for detecting cancer in these patients. Additional testing should be guided by any abnormalities detected by this clinical evaluation.