PADUA score as a predictor for pulmonary embolism: a potential strategy for reducing unnecessary imaging.
Kandagatla P, Goranta S, Antoine H, Marashi SM, Schmoekel N, Gupta AH. PADUA score as a predictor for pulmonary embolism: a potential strategy for reducing unnecessary imaging.. Journal of thrombosis and thrombolysis 2019; .
Journal of thrombosis and thrombolysis
An objective tool that is easy to integrate with an electronic medical record may help reduce unnecessary imaging for diagnosing a pulmonary embolism (PE). In this study, we assess the PADUA score in stratifying patients based on their risk of a PE. We reviewed charts of patients that underwent a computed tomography pulmonary angiogram (CT-PA) between January 2014 and September 2015 at our institution. Patient demographics including gender, age, race, and variables of the PADUA score were collected. The primary outcome was a positive CT-PA for a PE. Univariate and multivariate analysis was performed to derive predictors for a positive CT-PA. A receiver operator curve was calculated for the PADUA score and an optimal cutoff was calculated. Diagnostic test statistics were performed. Our study included 1067 patients. Of these, 185 (17.3%) had a PE. These patients tended to be older (64.3 SD 15.9 vs. 59.7 years SD 17.4, p < 0.01), have a higher proportion of Black patients (38.9% vs. 31.9%, p = 0.03), have a higher median [IQR] PADUA score (4.0 [3-6] vs. 3.0 [1-4], p < 0.01), and a higher rate of a DVT/PE history (30.3% vs. 5.2%, p < 0.01). Independent predictors included a DVT/PE history (OR: 7.65, 95% CI 4.89-12.0, p < 0.01), limited mobility (OR: 1.47, 95% CI 1.01-2.14, p = 0.046), and age 70 or greater (OR: 1.47, 95% CI 1.03-2.11, p = 0.03). The PADUA score had an AUC of 0.64 (95% CI 0.60-0.69, p = 0.046). The optimal cutoff was 4 and the sensitivity and specificity were 57.3% and 66.8%, respectively. The positive predictive and negative predictive values were 22.6% and 88.2%, respectively. The PADUA is a possible tool to stratify patients prior to performing a CT-PA. By using the score to guide management, we may be able to reduce unnecessary imaging through the implementation of the score in an EMR system. Further prospective research is warranted.
ePub ahead of print