The relationship between past fractures and current bone density (BMD)
was analyzed in a population sample of 3222 women aged 48-58. BMD was
determined with dual X-ray absorptiometry (DXA) at the spine and femo
ral neck. 702 women reported fractures. Wrist and ankle were the most
common fracture sites. Fracture history increased the risk [OR (95% CI
)] of low spinal BMD (of more than 1 SD below the study population mea
n) by 1.75 (1.41; 2.18). The sensitivity and specificity of fracture h
istory to detect a low spinal BMD were 31 and 80%, respectively. One S
D decreases of spinal and femoral BMD equalled to respective overall f
racture risks (adjusted ORs) of 1.36 (1.24; 1.50) and 1.38 (1.25; 1.51
). Both BMDs related more strongly to wrist fracture [1.73 (1.47; 2.05
)/1.69 (1.43; 1.99)] than to all nonwrist fractures combined [1.24 (1.
11; 1.37)/1.27 (1.14; 1.42)]. Ankle and rib fractures related only to
spinal [1.21 (1.00; 1.46)/1.45 (1.12; 1.87)] but tibia and foot bone f
ractures only to femoral [2.04 (1.37; 3.04)/2.20 (1.42; 3.41)] BMD. Sp
inal BMD related more strongly to fractures due to falls on same level
than to fractures due to all other trauma combined. Fracture history
poorly screens out low perimenopausal BMD. The results suggest that pr
e- and perimenopausal fractures relate to low axial bone density and t
hat the magnitude of this relation depends on the sites of fracture an
d densitometry as well as on the type of trauma.