Applicability of the Padua scale for Chinese rheumatic in-patients with venous thromboembolism

OBJECTIVE: To investigate the risk factors for venous thromboembolism (VTE) in hospitalized patients with rheumatic diseases in China. The efficacy of the Padua scale was evaluated and an improved model for predicting VTE in hospitalized patients with rheumatic diseases was developed. METHODS: Records of 2282 patients hospitalized in the department of rheumatology of the Sichuan Provincial People’s Hospital were retrospectively reviewed. The risk factors for VTE were analyzed. The efficacy of the Padua scale was evaluated, Padua-combined prediction model and the independent risk factor-combined prediction model for predicting VTE were assessed using the receiver operating curve (ROC). RESULTS: A total of 50 patients in the VTE group and 2232 in the non-VTE group were included. Antiphospholipid syndrome (APS), VTE history, a hospital stay of over 3 days, high D-dimer (D-D), and decreased serum albumin were independent risk factors for VTE. APS was very closely associated with VTE (OR = 19.446). Padua scores in the VTE group and the non-VTE group were 3 (2, 6) and 2 (1, 2) points, respectively (p < 0.05), and the proportion of high-risk patients were 48.0% and 7.4%, respectively (p < 0.05). The incidence of VTE in the high-risk (Padua score ≥4) and low-risk (Padua score <4) groups was 12.7% and 1.2%, respectively (p < 0.05). The area under curve (AUC) of the Padua scale, Padua combined prediction model (Padua scale along with D-D and serum albumin), and the independent risk factor-combined prediction model was 0.771, 0.836, and 0.873, respectively. CONCLUSION: The Padua scale has limited predictive efficacy of VTE in hospitalized rheumatic patients. The independent risk factor-combination prediction model was superior in predicting VTE compared to Padua scale and Padua-combined prediction model.