Objectives: To determine the accuracy of colour flow Duplex scanning (CFDS) in the diagnosis of deep vein thrombosis (DVT) and subsequently to investigate its diagnostic value in patients who have normal deep veins despite symptoms.
Design: Prospective open clinical study.
Setting: Vascular laboratory and radiology departments of University Hospital.
Materials and methods: In the first part 112 limbs in 103 patients, 94 with symptoms of acute DVT and nine with pulmonary embolism (PE) were examined prospectively with CFDS and venography. Subsequently, in the second part, 328 legs in 304 patients were examined by CFDS alone for acute symptoms of DVT or PE.
Main results: DVT was detected in 55 limbs by venography: proximal DVT was seen in 23 limbs, distal DVT in 25 limbs and both proximal and distal in seven limbs. CFDS was 100% sensitive and 98.8% specific in detecting proximal DVT whereas its sensitivity and specificity was 87.5% and 98.7% for distal DVT. Positive and negative predictive values were over 95% for both limb segments. The overall accuracy for the proximal DVT was 99.4% and for the distal 93.1%. In the second part, CFDS alone detected DVT in 156 limbs (47.6%); DVT was in the proximal segment in 82, distal segment in 61 and both in 13. In 172 limbs other causes of symptoms were identified in 34 (20%).
Conclusions: We have demonstrated that CFDS is as accurate as venography when used by experienced operators. The average time of examination is 15-20 minutes and compares favourably with venography. Other causes of leg symptoms can also be diagnosed by CFDS in around 20% of patients who are found to have normal veins.