BACKGROUND: The predictive value of coagulation markers for venous thromboembolism (VTE) in COVID-19 patients has been investigated with conflicting results. OBJECTIVE: Our aim was to investigate the correlation between biomarkers and VTE and the predictive value of D-dimer for VTE in hospitalized COVID-19 patients. METHODS: Complete blood count, inflammatory and coagulation biomarkers at admission were collected. VTE was defined as diagnosed pulmonary embolism or deep vein thrombosis. Events were defined as in-hospital death or ICU admission. Predictors of VTE were identified with Pearson prediction models. A ROC curve was constructed to assess the predictive value of D-dimer. RESULTS: 1651 participants were included, 111 VTE were identified. Events incidence was higher in the VTE group (49.5% vs 28.2%, p < 0.001). Neutrophil-lymphocyte ratio (NLR, 0.001; 95% CI 0.000-0.002; p 0.019) and D-dimer (0.00005; 95% CI 0.00002-0.00008; p < 0.001), Geneva score (0.026; 95% CI 0.012-0.040; p < 0.001) and Wells score (0.047; 95% CI 0.033-0.061; p < 0.001) were associated with VTE. D-dimer had a goor predictive value for VTE (ROC area 0.85, 95% CI 0.816-0.893), with an optimal cut-off value of 2677μg/L (Youden index of 0,602). CONCLUSIONS: Among coagulation biomarkers D-dimer had the best predictive value for VTE, but higher cut-off values should be used in COVID-19.