Population-based epidemiological studies on osteoporosis are few. Our study evaluated the effects of menopause and certain putative behavioural risk factors on bone mineral density (BMD). Spinal and femoral neck BMD were measured with dual X-ray absorptiometry (DXA) from 1600 perimenopausal women aged 48-59 years (mean 53.2 years) with no diseases or medications known to affect bone metabolism. These women were a selected sample of the Kuopio Osteoporosis Risk Factor and Prevention Study population (n = 14,220). There was a wide variation of BMD among perimenopausal women. Menopause had a major effect on BMD. Postmenopausal women had significantly lower BMD in both spine (-6.2%) and femoral neck (-3.9%) as compared with premenopausal women. Multiple regression analysis showed that weight, menopausal status, age, and grip strength were significant independent predictors of both spinal and femoral BMD. Additionally, physical activity was found to be a significant predictor of femoral BMD, and alcohol consumption was a significant predictor of spinal BMD. However, current anthropometric and lifestyle factors explained only 18.7-25.4% of the variability of BMD. Therefore, the estimation of the risk factor status at menopause is not an adequate substitute for bone densitometry. However, our results may in part help clinicians to identify the risk groups at which to direct bone density measurements.