DESCRIPTION (provided by applicant): Fractures are a common and debilitating consequence of osteoporosis, and bisphosphonates (BPs) are the first-line pharmacologic treatment worldwide. Unfortunately, accumulating evidence from observational studies points towards a relationship between BP use and the development of radiographically distinct atypical femoral fractures (AFF). Although the total number of suspected or confirmed AFFs is quite small, there is increasing patient and clinician concern that the use of BPs should account for AFF risk. In fact, despite the aging population the use of BPs since has declined by a third since 2007, likely reflecting uncertainty about the perceived risks vs. benefits. Specifically, there is conflicting evidence about the overall incidence of AFF and its relationship to BP use as evidenced by the huge variability in the reported relative risks (ranging from 1 to over 130). Now that a specific case definition and objective radiographic criteria have been established, future clinical studies must include expert radiograph review of suspected cases so that only confirmed AFF are included. Critical areas of uncertainty include relationships with BP dosing and duration of treatment, and importantly, resolution of risk after discontinuation of BP treatment. Additional potential risk factors for AFF, including certain medical conditions and medication use either alone or in combination with BPs, are poorly studies but might be clinically important and allow more rational treatment decisions. We therefore propose a longitudinal study using comprehensive Danish national health care data, including complete pharmacy records dating back to 1995, plus collection and review of digital radiographs of all Danish subtrochanteric and femoral shaft fractures from 2010-14. This study will fill significant gaps in our knowledge of AFF incidence and key risk factors by addressing 3 clinically important aims:1) To describe the population-based rates of radiographically confirmed AFF according to strata defined by age, gender and bisphosphonate use, 2) To determine the independent effect of bisphosphonate use as a risk factor for AFF by duration of use and cumulative dose, and to determine if other factors, such as severity of osteoporosis, co-morbid conditions and use of other medications, are associated with AFF, and 3) To determine if AFF risk declines after discontinuation of bisphosphonates, and to determine patient and treatment characteristics that influence AFF risk after discontinuation. The proposed study will provide clinicians and patients with important information about AFF incidence and risk factors, allowing an evidence-based comparison of both risks and benefits, and will have a substantial impact on the use of BPs in clinical practice.