Relationships between risk factors and fractures differ by type of fracture: a population-based study of 12,192 perimenopausal women.

Relationships between selected factors and fractures according to type of fracture were retrospectively examined in 12,192 women aged 47-56 years responding to the baseline postal enquiry of the Kuopio Osteoporosis Study, Finland, in 1989. A total of 1358 women reported fractures sustained during the previous 9.4 years, i.e. at ages 38-57 years. The incidence of fractures per 1000 person-years was 17.2 after menopause and 9.5 before (p < 0.0001). The adjusted fracture risk was elevated in smokers versus non-smokers (OR = 1.5; (95% CI = 1.3-1.9) and in those with chronic health disorders versus the healthy (OR = 1.3; 95% CI 1.1-1.5). Long-term work disability was associated with fractures independently of health disorders (OR = 1.3; 95% CI 1.1-1.6). Anthropometric measures were not associated with the overall fracture risk. Menopause was strongly and linearly related to wrist fracture but not to ankle fracture. A 1 SD increase in body mass index decreased the risk of wrist fracture by 21% (p = 0.0001) but increased that of ankle fracture by 24% (p = 0.002). Smoking was related to ankle fracture (OR = 2.2; 95% CI 1.6-3.2) but not to wrist fracture (OR = 0.9; 95% CI 0.6-1.4). Health disorders were more markedly associated with fractures other than those of the wrist or ankle. Relationships between several risk factors and pre- and perimenopausal fractures vary by type of fracture. This may affect, for example, the comparability of studies with varying fracture profiles.