Pharmacological thromboprophylaxis (TP) is known to reduce venous thromboembolism (VTE) in medical inpatients, but the criteria for risk-driven prescription, safety and impact on mortality are still debated. We analyze data on elderly patients with multimorbidities admitted in the year 2010 to the Italian internal medicine wards participating in the REPOSI registry to investigate the rate of TP during the hospital stay, and analyze the factors that are related to its prescription. Multivariate logistic regression, area under the ROC curve and CART analysis were performed to look for independent predictors of TP prescription. Association between TP and VTE, bleeding and death in hospital and during the 3-month post-discharge follow-up were explored by logistic regression and propensity score analysis. Among the 1,380 patients enrolled, 171 (15.2 %) were on TP during the hospital stay (162 on low molecular weight heparins, 9 on fondaparinux). The disability Barthel index was the main independent predictor of TP prescription. Rate of fatal and non-fatal VTE and bleeding during and after hospitalization did not differ between TP and non-TP patients. In-hospital and post-discharge mortality was significantly higher in patients on TP, that however was not an independent predictor of mortality. Among elderly medical patients there was a relatively low rate of TP, that was more frequently prescribed to patients with a higher degree of disability and who had an overall higher mortality.