Detection of prefracture spinal osteoporosis using bone mineral absorptiometry.

Bone mineral measurements have been criticized for their inability to clearly distinguish fracture and "nonfracture" populations. However, this failure is not unexpected, since some individuals in the "nonfracture" group have low bone mass and are at increased risk but have not yet experienced fractures. Although standard radiographs are not sensitive indicators of vertebral demineralization, they do identify some of the "prefracture" osteoporotic subpopulation within the nonfracture group. Prospective follow-up of 536 Japanese-American women demonstrated that 14 new spine fractures occurred in the prefracture osteoporosis group, whereas none occurred in the nonosteoporotic group (p less than or equal to 0.03). However, bone mineral content (BMC) measurements using photon absorptiometry were much more accurate than radiographs as indicators of spine fracture risk. BMC values were somewhat higher in the prefracture group than in those with existing fractures, but values for both groups were significantly lower than in nonosteoporotic patients even after adjusting for age, height, and weight (p less than 0.0001). The magnitude of the difference was proportional to the trabecular bone content of the measurement site; the differences were greatest for the os calcis and lumbar spine, smaller for the distal radius, and least for the proximal radius. The prevalence of spinal osteoporosis (including both fracture and prefracture cases) was inversely proportional to BMC (p less than 0.0001). Again, the relations were strongest for the os calcis and lumbar spine. These results indicate that BMC measurements are valid indicators of osteoporosis status, particularly when osteoporosis is defined to include both patients with existing fractures and those at increased risk for fractures. However, dual-photon spine BMC was adversely influenced by the presence of aortic calcification, arthritis, and other disease processes (p less than or equal to 0.0001).