Wells score for venous thromboembolism. Basic diagnostic algorithm for venous thromboembolism.

V thromboembolism (VTE) affects 1-2 per 1000 people in the general population each year, usually as deep-vein thrombosis (DVT) of the leg or pulmonary embolism (PE).1 Venous thromboembolism is a common, yet challenging diagnostic problem among both inpatients and outpatients. Clinical pre-test probability assessment is a cornerstone of the algorithms for the exclusion, or diagnosis of VTE.2,3 For patients suspected of VTE, the Wells score appears to be the most useful and well-validated clinical pre-test probability assessment.3 The Wells score, also called Canada score, including Wells DVT score and Wells PE score, has been built by Philip S. Wells in University of Ottawa, Canada on the basis of a series of investigations. In this article, we summarize the derivation, and the recent investigations of the Wells score for VTE. Wells DVT score. In 1995 Wells et al4 developed a clinical model to stratify pretest probability for DVT into high, moderate, and low categories. Items included in the clinical model were assembled from information obtained by a literature review, and from the collective experience of the participating investigators. These items were devided into 3 groups: signs and symptoms of DVT, risk factors for DVT, and potential alternative diagnosis. The clinical model was composed of specific items, designated as either major or minor that included proven risk factors, and pertinent symptoms, and physical signs at patient presentation. A probability score was derived, which categorized the patients into low, moderate, or high probability groups. The clinical model was prospectively tested to stratify symptomatic outpatients with suspected DVT, who had symptoms for less than 60 days. Finally, the clinical model predicted prevalence of DVT in 3 categories: 85% in the high, 33% in the moderate, and 5% in the low category. The weighted Kappa value for the assessment of interobserver reliability, for the clinical model, was 0.85 which represents an excellent level of agreement. However, the clinical model, criticized as being cumbersome, was not convenient for ordinary physicians, so Wells et al5 simplified it to a score by univariate, and stepwise logistic regression analysis of 529 patients’ clinical data. After retrospective analysis, Wells DVT score including 9 significant variables was shown in Table 1. According to the score, 529 patients were divided into 3 categories. In the high probability category the prevalence of DVT was 73%, in the moderate probability category the prevalence was 28%, and in the low probability the prevalence was 6%. The original model and score model were compared with respect to the prevalence of DVT in each of the 3 categories, and no significant difference was demonstrated (p=0.694, p=0.419, p=0.086). Wells et al6 used prospectively Wells DVT score in combination with ultrasound to guide management of patients with suspected DVT. Five hundred and ninetythree patients with suspected DVT were categorized as being at low, moderate, or high clinical probability for DVT by the Wells score, then all patients underwent deep venous ultrasound imaging of lower limb. Patients at low clinical probability underwent a single ultrasound test. A negative ultrasound excluded the diagnosis of DVT, whereas a positive ultrasound was confirmed by venography. Patients at moderate probability with a positive ultrasound were treated for DVT, whereas patients with an initial negative ultrasound had a single follow-up ultrasound one week later. Patients at high