NORMAL RANGES FOR BONE LOSS RATES

We reported previously that the variability in bone loss rates among p ostmenopausal women decreases dramatically during the first few years of followup. In this paper, we have examined the distributions of bone loss rates measured at the calcaneus, distal radius and proximal radi us. The incidence of physical impairment was five times greater among women with bone loss rates faster than 2 S.D. below the mean. Because the rate of change in bone density was skewed at the lower end of the distribution (representing rapid bone loss), the influence of values a t the extreme ends of the distribution were statistically removed in o rder to estimate the normal distribution of bone loss rates. For the c onvenience of clinicians, the upper and lower limits of the 90 and 70% normal ranges are presented. Because average bone loss rates vary wit h age, normal ranges are provided separately by age group. The width o f each normal range decreased by at least half after 3 or 4 years of f ollowup, compared to less than 1 year. Consequently, measured loss rat es which were well within the normal range at 1 year were sometimes fa r outside the normal range for longer followup times. We conclude that followup duration has a profound effect on estimates of the normal ra nge, and must be considered when interpreting the clinical significanc e of measured loss rates.