To understand differences in bone health between and within populations, it is crucial to characterise bone development during childhood and adolescence. Peak height velocity at age 16 and young adult height at age 23.5 years were recently reported in Gambian males accustomed to low calcium intake1. Our study aims to describe bone accrual after peak height velocity in the same population.
We used peripheral quantitative computed tomography to measure the 66% radius biennially, at mean (S.D., n) ages 19.4 (0.9, 60), 21.7 (1.0, 59) and 23.6 (0.9, 51) years. Outcomes were cortical volumetric bone mineral density (vBMD), bone mineral content (BMC), total and cortical cross-sectional area (CSA), and stress-strain index (SSI). We used random intercept and random slope models to assess the relationship between each outcome and changes in age, height and weight. Random intercept models assume that all individuals slopes share the same gradient, but allow their intercepts to vary. Random slope models allow for between-individual variation in gradients as well as in intercepts.
Random slope models were a significantly better fit than random intercept models for BMC, vBMD and total CSA against age (likelihood-ratio tests P<0.05), indicating individual differences in the rate at which these bone measures changed. For all other variables, random intercept models were sufficient. Annual change in SSI and cortical CSA was similar for all individuals. No bone measure showed significant differences between individuals gradients when modelled against height or weight. Slope coefficients in all models were significant and positive (P<5×10−5).
Individual differences in the rate of age-related change in BMC, vBMD and total CSA may reflect differences in deceleration of bone development in older individuals. Alternatively, differences in maturational stage may modify the effect of age on bone. Anthropometric measures are more directly associated with bone development than chronological age i.e. somatic and skeletal growth are co-ordinated but not determined by chronological age alone. This may reflect biomechanical responses in the skeleton, but could equally be due to joint control of growth and bone development by genetic, hormonal or nutritional factors. The age of cessation of bone accrual remains to be determined, and the implications for population bone health require further investigation.
Funding: UK Medical Research Council (Grant codes U105960371 and U123261351).
Reference: 1. Prentice et al. AJCN 2012 96 10421050.
22 - 25 Jun 2013