Assessing Pretest Clinical Risk of Pulmonary Thromboembolism in the Emergency Department: Proposal of a Simple Modification to the Wells' Score.

Background: Clinical scores have been proposed to stratify the risk of pulmonary thromboembolism (PTE), although this approach suffers a low specificity and the unavoidable need for computed tomography pulmonary angiography (CTPA) scans. Objective: Our study aimed to investigate a simple modification to the already validated Wells' score to improve its diagnostic accuracy in the emergency department (ED). Methods: We retrospectively reviewed all CTPA scans performed in the ED setting to rule out PTE over a 1-year (2017) period. Clinical variables potentially associated with PTE were assessed to improve diagnostic accuracy of the Wells' score, thus introducing a modified Wells' score (mWells). Results: Four thousand four hundred thirteen CTPAs were identified, of which 504 were for suspected PTE. The prevalence of PTE was 23.9%. Among clinical data, only peripheral capillary oxygen saturation was consistently correlated with PTE at univariate (odds ratio 2.75 [95% confidence interval 1.61-4.73]) and multivariate (odds ratio 3.78 [95% confidence interval 2.13-6.72]) logistic regression analysis. The mWells' score had a higher area under the receiver operating characteristic curve compared with the original Wells' score: 0.71 (95% confidence interval 0.67-0.75) vs. 0.65 (95% confidence interval 0.61-0.69) (p < .01) and improved diagnostic accuracy. Conclusions: Current clinical stratification tools for PTE are characterized by low specificity, leading to an overuse of CTPA. mWells', rather than Wells', score showed a better predictive performance of PTE detection. Our results suggest that current diagnostic pathway for PTE may be improved by simple adjustments (i.e., mWells') of clinical prediction scores. (C) 2019 Elsevier Inc. All rights reserved.