Enoxaparin 40 mg/day (E) or fondaparinux 2.5 mg/day (F) are recommended to prevent venous thromboembolism (VTE) in medical and surgical patients at risk. Over the two years after switching from E to F in our 35-bed department of pulmonology and thoracic surgery, an increase in the number of transfusions was observed. A retrospective explanatory investigation was undertaken. Hospitalised patients in the two years before and after switching from E to F were compared. The files of all transfused patients were reviewed. A blinded independent committee adjudicated major bleeding events. In the investigated time period, the overall transfusion rate increased from 1.8% of 2,989 patients to 3.1% of 3,085 patients (p=0.002). Mean ages (58.4 vs. 59.1 years), proportions of surgical patients (63.6% vs. 58.4%), cancer patients (72.1% vs. 69.5%), and treated patients (≥ 1 dose of E or F: 51.8% vs. 52.5%) were similar. The number of medical patients transfused while receiving E or F did not increase significantly (0.9% vs. 1.3%, RR=1.45 [0.66-3.17], p=0.35). The number of surgical patients transfused postoperatively while receiving E or F increased significantly (0.7% vs. 1.9% of all surgical patients, relative risk [RR]=2.75 [1.45-5.23], p=0.001), including a significant increase in transfusions for major bleeding (0.2% vs. 0.9%, RR=5.97 [1.74-20.4], p<0.001). A multivariate analysis did not find confounding factors. The incidence of symptomatic postoperative pulmonary embolism remained very low (0.05% vs. 0.17%). In conclusion, in thoracic surgery patients, switching from enoxaparin to fondaparinux to prevent VTE was associated with a significant increase in the risk of postoperative major bleeding. A causal relationship appears plausible.