A 5-YEAR LONGITUDINAL-STUDY OF FOREARM BONE MASS IN 307 POSTMENOPAUSAL WOMEN

We measured forearm bone mineral content at the beginning and end of a 5 year period in 307 untreated postmenopausal volunteers. We also mea sured height, weight, and a number of biochemical variables in plasma and urine after an overnight fast. The initial mean age of the subject s was 59.0 years (range 39-72), and the mean years since menopause was 10.0 (range 1-37). The mean forearm BMC fell from 1034 +/- 9.6 (SEM) to 982 +/- 9.3 mg/cm (P < 0.001). The coefficient of correlation betwe en the first and second measurements was 0.96. The mean rate of change was -1.0% per annum (with a 99% range of -4 to 1% per annum), which a greed well with previous estimates from cross-sectional data. There wa s a significant negative correlation between rate of change in bone ma ss and initial value (r = -0.23; P < 0.001), which was eliminated by e xpressing change as a percentage of initial bone mass. Of the other va riables measured, the one that was most significantly related to the p ercentage change in bone mass was the urinary hydroxyproline/creatinin e ratio (r = -0.35; P < 0.001), which we regard as a marker only. By s tepwise regression, the only significant determinants of the rate of c hange in bone mass were body weight (positive, P < 0.001), years since menopause (positive, P < 0.005), urine calcium (negative, P < 0.01), and serum estrone (positive, P < 0.05). The rate of change in BMC was not significantly related to dietary calcium, plasma alkaline phosphat ase, serum DHEAS, or any other variable measured at the outset or to s erum PTH or serum osteocalcin measured at the conclusion.