OBJECTIVE--To determine the frequency of deep venous thrombosis (DVT) in medical intensive care unit (MICU) patients. DESIGN--Prospective ultrasound case series. SETTING--An MICU in a large tertiary care hospital in Boston, Mass. SUBJECTS--Patients older than 18 years of age admitted to the MICU with an anticipated stay of more than 48 hours. MAIN OUTCOME MEASURE--Deep venous thrombosis as detected by ultrasonography with color Doppler imaging performed twice weekly in the MICU and once within 1 week of discharge from the MICU. RESULTS--Deep venous thrombosis was detected in 33% (95% confidence interval, 24% to 43%) of 100 eligible patients during the 8-month study period. Forty-eight percent (16/33) were proximal lower extremity DVT, and 15% (5/33) were upper extremity DVT associated with central venous catheters, with one patient having both upper and proximal lower extremity DVT. Ultrasound examination results led to inferior vena cava filter placement in three patients, initiation of full-dose anticoagulation in four patients, initiation or continuation of low-dose subcutaneous heparin in 10 patients, follow-up ultrasound studies in three patients, central line removal in one patient, and no intervention in 10 patients due to active bleeding, prior filter, or heparin-induced thrombocytopenia. Two patients remained anticoagulated for other reasons. In this series, there was no difference in age, gender, body mass index, diagnosis of cancer, recent surgery, duration of hospitalization prior to DVT detection, and DVT prophylaxis between patients with DVT and those without. CONCLUSIONS--An unexpectedly high rate of DVT was detected by ultrasound in these MICU patients despite prophylaxis in 61%. Traditionally recognized DVT risk factors failed to identify patients who developed DVT. Routine ultrasound surveillance or more intensive prophylaxis regimens may be warranted in this patient population if these DVT rates are confirmed in other settings.