Background: The combined use of bone mineral density, fracture history and other risk markers for fracture is advocated for identifying subjects with high fracture risk. An incident fracture is suggested as an accurate indication for osteoporosis investigation, but there are still insufficient data for grading the priority between ages and types of fractures. We therefore decided to examine a consecutive series of 55-75 year old women with an incident fracture for evaluating a standardized clinical routine program and for studying the covariance between fracture history, bone mineral density and other risk markers.
Materila and methods: We invited 600 consecutive women 55-75 years old with an incident newly diagnosed fracture in distal radius forearm, proximal humerus, vertebra or hip. External drop-out was 33%. Of the 400 responders 31 had a high-energy trauma, 62 were on treatment against osteoporosis and 4 were living in other counties and were therefore excluded. The remaining 303 subjects entered the study. A questionnaire on previous fractures and risk factors was enclosed with the invitation to the osteoporosis unit. At a single visit a short history was assessed and physical examination performed as well as a few laboratory investigations. Bone mineral density was measured at the hip, lumbar spine and forearm by DXA (Hologic QDR 4500A).
Results: The fracture spectrum was: distal radius 56.4 %, proximal humerus 12.2%, vertebra 18.2% and hip 13.2%. 49% had had at least one previous fracture, 19% at least two previous and 6.3% three or more previous fractures before the recent one. As few spine X-rays were performed, the true prevalence of vertebral fracture is unknown. Patients with fracture in vertebra or hip had lower BMD and more previous fractures than patients with forearm or humerus fracture. The number of previous fractures was inversely correlated to BMD of the hip and forearm, while BMD of the spine had a biphasic relationship.
The Odds ratio of having either osteopenia and osteoporosis were >20 for patients with hip fracture and 75 for the spine (mean values), whereas the OR of the forearm fracture group was slightly above 10, table 6. The OR were as expected dependent of cut-off limit used. Mean value for the OR was in the hip fracture group 8.2 and 9.2 for !-score -2.5 and -2.0 respectively at the lower end of the confidence interval. For the spine and the forearm, the corresponding odds ratios were 16-17 and 7-9 respectively.
Conclussion: Vertebral fracture was the strongest and distal radius the weakest predictor of low BMD. The number of previous fractures is a helpful information for finding the most osteoporotic patients. Only 15 % had been treated for osteoporosis before the index fracture. Osteoporosis investigation therefore seems warranted in every woman 55-75 years old with a recent low-energy fracture in distal radius, proximal humerus, spine or hip, with highest priority to those in spine or hip and those with multiple previous fractures.