Interrelationships among bone densitometry sites in normal young women.

Interrelations among skeletal sites were examined in a population of up to 300 normal young white women aged 20-40 years. Measurements were done on the radius shaft, ultradistal radius, and ultradistal ulna using single-photon absorptiometry (SPA) and on the lumbar spine, proximal humerus, and proximal femur using dual-photon-absorptiometry (DPA). Because of the narrow range of intrapopulation variance in these normal young women, the intercorrelations among skeletal sites were not very high; the average correlation (r) was 0.43 for bone mineral content (BMC) and 0.45 for bone mineral density (BMD). The results at any one site predicted values at other sites with a standard error of estimate (SEE) of approximately 11% for BMD and 17% for BMC. Even in this relatively homogeneous sample, measurements at a single skeletal site could not be extrapolated to indicate skeletal status at other sites. BMC, and to a lesser extent BMD, were influenced by body size much as has been observed in both men and postmenopausal women. Bone results are preferably expressed as BMD, because (a) intrapopulation variability is lower (10% versus 15% for the lumbar spine), (b) intercorrelations among skeletal sites, and even among adjacent vertebrae, are higher for BMD, and (c) the influence of body size on BMD is lower (the average r for weight and BMD was 0.26 versus 0.33 for BMC).