Abstract
Background: Venous thromboembolism (VTE) remains an important preventable complication among hospitalized patients, highlighting the need for effective risk assessment. This study evaluated D-dimer and routine coagulation parameters for predicting VTE and compared their performance with the Padua Prediction Score. Methods: In this prospective observational cohort study, 300 hospitalized adults at risk of VTE underwent admission assessment of D-dimer, coagulation parameters, and Padua score. Patients were followed for objectively confirmed VTE during hospitalization. Predictive performance was evaluated using receiver operating characteristic analysis and multivariable logistic regression. Results: VTE occurred in 44 patients (14.7%). Median D-dimer was higher in patients who developed VTE than in those who did not (1.76 versus 0.58 μg/mL; p<0.001). D-dimer showed the highest discrimination (AUC = 0.751), compared with the Padua score (AUC = 0.687) and conventional coagulation parameters. At 0.5 μg/mL, D-dimer demonstrated 86% sensitivity and 94% negative predictive value. D-dimer remained independently associated with VTE in multivariable analysis (OR = 2.29 per standard-deviation increase; 95% CI: 1.40-3.75; p<0.001). Conclusion: Admission D-dimer showed greater predictive performance than conventional coagulation parameters and the Padua score and may complement clinical VTE risk assessment in hospitalized patients.
Keywords: Venous thromboembolism, D-dimer, Coagulation, Padua prediction score, Risk assessment, Negative predictive value, Hospitalized patients