Clinical question: How useful is the Wells score for risk-stratifying hospitalized patients with suspected deep vein thrombosis?
Bottom line: The Wells score is not helpful in the inpatient setting to predict the presence or absence of deep vein thrombosis (DVT). Based on this study, if a hospitalized patient has a low Wells score, the risk of having DVT is still relatively high (6%). If a patient has a moderate or high score, however, the risk of having DVT is fairly low (10% to 16%). In all 3 categories, a patient would need further testing with ultrasound to evaluate for DVT. (LOE = 2b)
Reference: Silveira PC, Ip IK, Goldhaber SZ, Piazza G, Benson CB, Khorasani R. Performance of Wells score for deep vein thrombosis in the inpatient setting. JAMA Intern Med. 2015;175(7):1112-1117.
Study design: Diagnostic test evaluation
Funding source: Unknown/not stated
Allocation: Uncertain
Setting: Inpatient (any location)
Synopsis: The Wells score has been previously validated to risk-stratify outpatients with suspected DVT but its utility in the inpatient setting is unknown. These investigators evaluated 1135 hospitalized patients with suspected DVT who underwent a lower extremity ultrasound study in the hospital. When ordering these studies, clinicians were required to enter information regarding clinical predictors in order to calculate an individual patient’s Wells score. The patients were divided into 3 Wells score categories that determined their pre-test probability for DVT (low risk = 0 or lower, moderate risk = 1 or 2, high risk = 3 or higher). Baseline characteristics for the patients in the study showed that 71% were recently bedridden or had recent major surgery and almost 40% had active cancer.
Overall, 12% of patients in the study had proximal DVT confirmed by a lower extremity ultrasound study. When classified by Wells score categories, the incidence of proximal DVT was 5.9%, 9.5%, and 16.4% in low, moderate, and high pre-test probability groups, respectively. The area under the receiving operating characteristics curve for the Wells score as a diagnostic test was 0.60.
This indicates that the ability of the Wells score to discriminate between the presence and absence of DVT in hospitalized patients was only slightly better than chance. The authors postulate that the reason for this is that hospitalized patients are inherently different from outpatients: they have a higher prevalence of immobilization and/or have active cancer; they receive routine DVT prophylaxis; and they are more likely to have other comorbidities that increase DVT risk, such as heart failure and chronic obstructive pulmonary disease — risk factors that are not accounted for in the Wells score calculation. As such, the Wells score is less meaningful in this population.
Dr. Kulkarni is an assistant professor of hospital medicine at Northwestern University in Chicago.