Wells clinical prediction criteria in patients suspected of having deep vein thrombosis: evaluation of components and use in the emergency department

Objectives.

To determine the applicability of the Wells clinical prediction criteria for deep vein thrombosis (DVT) in patients in hospital emergency departments and to evaluate the relevance of the score’s components.

Methods.

Prospective multicenter cohort study in consecutive hospital emergency department patients suspected of having DVT. Full-leg Doppler compression ultrasound imaging was performed on all patients. We recorded information

on variables related to risk for DVT and the components of clinical prediction scales. Wells and Oudega clinical prediction scores were calculated.

Results.

We studied 362 patients in 23 hospital emergency departments; the mean (SD) age was 65 (18) years and 52.8% were women. DVT was diagnosed in 254 patients (70.16%); 171 (47.2%) had proximal DVT. The clinical probability of DVT according to the Wells scale and the prevalence of proximal DVT were as follows: low probability, 57 patients (14 with DVT, 24.6%); intermediate probability, 124 (43 with DVT, 34.7%), and high probability, 181 (114 with DVT, 63%). Only 5 of the components of the Wells scale were associated with the presence of proximal

DVT.

Conclusions.

The prevalence of DVT is very high in the 3 categories of clinical probability indicated by the Wells score. The prevalences do not correspond to those of the cohort used to validate the scale. It appears to be necessary to develop scales adjusted for use in hospital emergency departments when DVT is suspected.