Excellent performances of Wells' score and of the modified Wells' score for the diagnosis of proximal or distal deep venous thrombosis in outpatients or inpatients at Toulouse University Hospital: TVP-PREDICT study

In order to validate a standardized strategy for the diagnosis of lower limb deep vein thrombosis (DVT) in the regional university hospital of Toulouse, we decided to study the performances of Wells' score and the modified Wells' score for the diagnosis of proximal and distal DVT. Inpatients or outpatients referred to the vascular medicine department from April 2006 to March 2007 with suspected DVT were included prospectively and consecutively. Wells' score was determined for each patient and compared with the duplex ultrasound result. Two hundred and ninety-seven patients were included. The prevalence of DVT was 13.5%. The negative predictive values of Wells' score and the modified Wells' score were 99 and 97% respectively. Similar results were found for proximal or distal thrombosis. The performances of the modified Wells' score were not statistically better than those of the original score. In 48% of patients, the determination of the D-dimers would not have been contributory. In the group with low probability (70% of patients), the incidence of thrombosis was 0.6%. Wells' score and Wells' modified score have shown excellent performances. The value of the modified Wells' score is not superior and our preference, for practical reasons, goes to the original score. The widespread use of duplex ultrasound, the large proportion of patients in which D-dimers would not have been contributory and the excellent results of Wells score for patients with a low probability of DVT are encouraging arguments in favor of the development of an alternative strategy for these patients.