RISKS OF PERIMENOPAUSAL FRACTURES - A PROSPECTIVE POPULATION-BASED STUDY

Objective. To examine the associations between potential risk factors and fractures in perimenopausal women. Subjects. A total of 3140 women (mean age 53.4+/-2.8 (s.d.) years) were followed-up for 2.4 years aft er axial bone densitometry (lumbar spine and femoral neck) with regard to the occurrence of fractures. Results. In all, 5.6% of the women su stained a fracture. There were 169 low energy fractures (falling on a level surface) in 157 women after the exclusion of 18 fractures caused by a high energy trauma. The wrist was the most frequent site of frac ture (n=46). Lumbar bone mineral density was 5.8% lower and femoral bo ne mineral density 4.6% lower among fracture cases compared with non-f racture cases (p<0.0001). History of a fracture during 1980-1989 eleva ted the risk of all fractures 2.83-fold (95% confidence interval (CI) 1.95-4.10) and the risk of a first wrist fracture 2.25-fold (95% CI 1. 10-4.62). The amount of weekly alcohol intake was higher among fractur e cases than among non-fracture cases yielding an age-adjusted odds ra tio (OR) of 1.45 (95% CI 1.05-2.02). Past or present use of hormone re placement therapy was protective against fractures (age-adjusted OR 0. 70, 95% CI 0.50-0.96). If bilateral oophorectomy had been carried out under the age of 45 years, the risk of fracture was 3.64-fold (95% CI 1.01-13.04) compared with women operated upon after the age of 45 year s. Age at menarche, parity, lactation and smoking history did not diff er between the fracture and non-fracture groups. Conclusions. A former history of fractures, low baseline bone mineral density (BMD) and use of alcohol are predisposing factors associated with perimenopausal fr actures, while hormone replacement therapy is protective in this respe ct.