Normalization of spine densitometry.

We investigated several transformations of bone mineral content (BMC) and area density (BMD), in particular volumetric density (BMAD), to ascertain the influence on (1) body size dependence, (2) diagnostic sensitivity, and (3) precision. These transformations were examined in a group of 657 normal postmenopausal women and 327 women with osteoporotic fracture. First, expression of results as BMAD removed some of the slight dependence on body size; 21% of the variation in BMC and 15% of the variation i BMD were associated with body weight, but only 8% with BMAD. Second, the Z scores compared with those for age-matched controls for BMD and BMC were -1.85 and -1.71, respectively; the Z score for BMAD was -1.64. Third, the precision error for BMC was reduced by expressing results as BMD (1.1 versus 0.5%); BMAD degraded precision slightly (0.7%). BMD appeared to be the optimal expression for bone densitometry because it provided the best diagnostic sensitivity and lowest precision error; there was a minimal influence of body size on BMD results. This study also showed that osteoporotic women, even in the first postmenopausal decade, had low spine BMD, small vertebral area, and low body weight. Such women may be particularly at risk of crush fracture.