OBJECTIVE: To determine which parameters, including femur morphology, proximal femur bone mineral density, or patient characteristics, are associated with bisphosphonate-related atypical femur fractures (AFFs) and to investigate the relationships between AFF location and these parameters. MATERIALS AND METHODS: Sixteen females with a history of bisphosphonate use who presented with AFFs and 38 females without AFFs, even those with long-term bisphosphonate use of >
5 years, were compared. Patient characteristics
physique, gait ability, and history of pain and medication, were recorded. In terms of femur morphology, the mechanical lateral distal femoral articular angle (mLDFA), femoral bowing angle, femoral neck - shaft angle, and ratio of the femoral lateral cortex:neck width were measured via anteroposterior femur radiographs. Bone mineral density values of the femur neck, trochanter, inter, and Ward's triangle were used. Logistic regression analysis was conducted to determine independent factors for the AFF. In the 16 AFF patients, AFF locations were divided into 1: subtrochanteric, 2: proximal, 3: middle, and 4: distal femur diaphysis. The Spearman correlation coefficients (rs) between the AFF locations and the parameters were calculated. RESULTS: Logistic regression analysis revealed that the mLDFA and Ward's triangle values were significantly independently associated with AFF patients (odds ratios = 1.4 and 0.0, respectively). A more distal AFF location was significantly correlated with a shorter body height and greater BMI (rs = -0.69 and 0.67, respectively). CONCLUSION: An evaluation combining the mLDFA and Ward's triangle value could be used to predict the risk of bisphosphonate-related AFFs. Body height or BMI may be helpful for predicting AFF location.