Bone density was determined in six areas (head, arms, thorax, spine, l
egs, and pelvis) using total body dual photon X-rav absorptiometry. Va
lues were compared with the conventional anterior-posterior lumbar spi
ne (L2-L4) absorptiometry measurement in 20 young females (25 +/- 4.2
years), 41 perimenopausal females (51 +/- 2 years), and 39 postmenopau
sal females (time since menopause 2.7 +/- 1.9 years). The two older gr
oups were of similar age and physical characteristics. Total body bone
mineral density and leg bone mineral density were significantly lower
in the perimenopausal women than in the younger women, but this diffe
rence was no longer apparent after adjustment for height. Postmenopaus
al women exhibited diffuse bone loss with a particularly marked decrea
se in the lumbar spine measurement. Lumbar bone mineral density was si
gnificantly correlated with bone mineral densities at other sites; cor
relation coefficients ranged from r=0.56 for the head to r=0.83 for th
e total skeleton. However, one third of females with a lumbar measurem
ent predictive of a high fracture risk had normal bone mineral density
at another site, and vice versa.'Our findings show that postmenopausa
l bone loss affects the entire skeleton and that measurements should b
e performed at several sites in order to identify high-risk women at c
essation of menses.