ASSOCIATIONS OF EARLY PREMENOPAUSAL FRACTURES WITH SUBSEQUENT FRACTURES VARY BY SITES AND MECHANISMS OF FRACTURES

In a retrospective population-based study we assessed whether and how self-reported former fractures sustained at the ages of 20-34 are asso ciated with subsequent fractures sustained at the ages of 35-57. The 1 2,162 women who responded to fracture questions of the baseline postal enquiry (in 1989) of the Kuopio Osteoporosis Study, Finland formed th e study population. They reported 589 former and 2092 subsequent fract ures. The hazard ratio (HR), with 95% confidence interval (CI), of a s ubsequent fracture was 1.9 (1.6-2.3) in women with the history of a fo rmer fracture compared with women without such a history. A former low -energy wrist fracture was related to subsequent low-energy wrist [HR = 3.7 (2.0-6.8)] and high-energy nonwrist [HR = 2.4 (1.3-4.4)] fractur es, whereas former high-energy nonwrist fractures were related only to subsequent high-energy nonwrist [HR = 2.8 (1.9-4.1)] but not to low-e nergy wrist [HR = 0.7 (0.3-1.8)] fractures. The analysis of bone miner al density (BMD) data of a subsample of premenopausal women who underw ent dual x-ray absorptiometry (DXA) during 1989-91 revealed that those with a wrist fracture due to a fall on the same level at the age of 2 0-34 recorded 6.5% lower spinal (P = 0.140) and 10.5% lower femoral (P = 0.026) BMD than nonfractured women, whereas the corresponding diffe rences for women with a former nonwrist fracture due to high-energy tr auma were -1.8% (P = 0.721) and -2.4% (P = 0.616), respectively. Our r esults suggest that an early premenopausal, low-energy wrist fracture is an indicator of low peak BMD which predisposes to subsequent fractu res in general, whereas early high-energy fractures are mainly indicat ors of other and more specific extraskeletal factors which mainly pred ispose to same types of subsequent fractures only.