Changes in metacarpal bone mineral density with age and menopause using computed X-ray densitometry in Japanese...

Bone mineral density (BMD) loss with age and menopause is widely accepted in elderly women. However, only a few studies have utilized a multiple regression model that includes physical characteristics to assess comprehensive lifetime changes in BMD.

A prospective study was conducted to characterize the normal patterns in metacarpal BMD changes in Japanese women, and to assess the applicability of a fitting model using cross-sectional data compared with longitudinal variability.

The study consisted of 5422 healthy women in cross-sectional data and a 1-year follow-up of 359 women. The metacarpal BMD was measured by computed X-ray densitometry. Multiple linear and nonlinear regression analyses were performed in cross-sectional subjects. Nonparametric analysis was used to compare percentage rates of BMD changes between actual and estimated values.

The cross-sectional data showed that the best-fit equation was a nonlinear change model using the variables of age and height in premenopausal women, and years since menopause (YSM), age and height in postmenopausal women. The results of longitudinal data indicated the following. In premenopausal women, the actual BMD changes were greater in the 30-39 age group than the 20-29 age group and were less in the 50-59 group than the 40-49 group. The rates of annual change in BMD between the actual value and estimated value by change model were very similar. In postmenopausal women, the actual changes in BMD indicated that the rapid rate of reduction observed was over 3% at 0-5 YSM and 1.5% at 6-10 YSM, and thereafter showed a slower rate of decline at 11 YSM. The change model represented the trend of actual change in BMD for postmenopausal women, whereas the rates of estimated BMD loss underestimated the actual changes at 1-10 YSM.

The change model for premenopausal women using cross-sectional data is beneficial in evaluating the actual metacarpal BMD variability, whereas that for postmenopausal women is insufficient in estimating the longitudinal BMD variability.