The surgical treatment of osteogenesis imperfecta (OI) is negatively influenced by clinical features such as osteoporosis, limb deformities and bone changes caused by bisphosphonate therapy. Blood loss during femoral nailing surgeries in patients with OI is a serious problem. Platelet anomalies have been associated with an elevation of the serum pyrophosphate originating from the platelets during clotting, even if the causality with the platelet dysfunction has not yet been established. To identify predictive risk factors regarding intraoperative bleeding, a retrospective analysis was conducted on 23 patients aged between 6 and 13 years, affected by OI type III, who were treated to correct femoral deformities or to perform an osteosynthesis for femoral shaft fractures, using the Fassier-Duval telescopic nail. Osteotomies were performed in 14 cases of deformities and in two out of seven cases of fractures. A survey about the bleeding had been obtained by calculating the sum of the blood aspirated and that lost with the gauzes or present on the surgical drapes. To obtain an estimate of the intraoperative blood losses, one must resort to a calculation based on an algorithm that evaluates the ratio between the effective blood loss divided by the total blood volume expected as per age and weight (γ distribution). The average blood loss was 237.4 ml (0.12 γ). In seven cases, it was necessary to perform postoperative transfusions, owing to an average blood loss of 502.8 ml (0.27 γ). Patients aged less than 10 years had a minor blood loss. A greater number of osteotomies was associated with a significant increase of average bleeding (P=0.046). Patients who were never treated with bisphosphonates showed a significantly greater bleeding rate (P=0.048). Patients affected by OI type III have a high risk of severe blood loss during surgery, even caused by the platelet disfunction, which characterizes this OI type. In addition to this predisposing factor, there are other risk factors to consider in preoperative surgical planning. In patients who were never treated with bisphosphonates, the bleeding was higher than in the ones treated with bisphosphonates since at least 1 year. The effects of bisphosphonates on bone tissue (such as the medullar canal narrowing and the bone cortex thickening) could reduce the spongious bone amount and the bleeding. Inhibiting the farnesyl pyrophosphate synthase enzyme and reducing the prenylation of many plasma proteins, including the methylene tetrahydrofolate reductase, the bisphosphonates could lead to an alteration of the coagulation cascade. The correlation found with the intake of bisphosphonates, capable of inhibiting the action of the farnesyl pyrophosphate synthase enzyme, thus influencing coagulation, requires further prospective studies with research of the methylene tetrahydrofolate reductase mutation in patients with OI type III undergoing surgical procedures. The number of osteotomies, the patient's age and the intake of bisphosphonates for at least 1 year seem to be the best predictive factors for blood loss.