Predictors of fractures in elderly women

In a prospective study of 348 apparently healthy women, aged 70 years and o ver (mean 80.3 years), we examined bone mineral density (BMD), biochemical markers of bone metabolism, and some easily measurable predictors in relati on to hip and osteoporotic fractures. In addition, we constructed risk prof iles for hip and osteoporotic fractures. At baseline, BMD at both hips, usi ng dual-energy X-ray absorptiometry, body height and body weight were measu red. At the same rime, serum and urine samples were obtained for biochemica l analysis. Serum samples were analyzed for vitamin D metabolites, sex horm one binding globulin, serum intact parathyroid hormone, osteocalcin, alkali ne phosphatase, phosphate, albumin, calcium and creatinine. In 2 h fasting urine, hydroxyproline, type I collagen crosslinked N-telopeptide (NTx) and calcium excretion were measured. Furthermore, easily measurable predictors, such as previous fracture, body mass index (BMI) and mobility were assesse d. During the follow-up period (mean duration 5.0 years), hip and any osteo porotic fracture (wrist, humerus or hip fracture) occurred in 16 and 33 par ticipants, respectively. Data were analyzed using Cox regression analysis. BMD of the trochanter (per 1 SD decrease) and previous fracture were most s trongly associated with hip fractures (adjusted relative risk (RR) = 3.0, 9 5% confidence interval (CI): 1.4-6.6; RR = 4.2, 95% CI: 1.5-11.6, respectiv ely) and osteoporotic fractures (RR = 1.8, 95% CI: 1.1-2.8; RR = 2.9, 95% C I: 1.5-5.7, respectively). Previous fracture, BMI and mobility were identif ied as easily measurable predictors for hip fractures, whereas previous fra cture, use of loop diuretics and age were predictors for osteoporotic fract ures in the risk profile model. The risk of fractures can be predicted with three easily measurable predictors. This study confirms the importance of previous fracture as a predictor for hip fractures and other fractures. It also shows that the use of loop diuretics is a predictor for osteoporotic f ractures.