Objective: To compare the accuracy of the predictive equations derived from older Chinese people without vertebral deformity with those derived from unselected older people, in estimating height from armspan.
Methodology: Height and armspan were measured in 789 older Chinese and 885 young Chinese in Hong Kong. All the older subjects had spinal X-ray (T4 to L4). Vertebral deformity was determined by a trained technician using a translucent digital analyser. Definite deformity was defined by a greater than 3 standard deviation increase in one of the three vertebral height ratios (anterior/posterior, middle/posterior and posterior/posterior). Eighty of the older subjects without vertebral deformity were randomly selected from the studied population for internal validation of predictive equations. The remaining subjects were used to derive sex-specific predictive equations for height by linear regression against armspan. The accuracy of these equations was analysed by the Bland and Altman method.
Results: Vertebral deformity was present in 30.7% and 15.6% of older female and male subjects, respectively. Older women without vertebral deformity had significantly greater armspan: height ratio when compared with the young (1.04 versus 1.02), but the corresponding ratios were the same in older and young men (both 1.04). The predictive equation derived from older females without vertebral deformity had lower standard error of estimation and gave estimates closer to measured height than those derived from unselected older women. But this was not the case in older men.
Conclusion: In deriving predictive equation to estimate height by armspan in older Chinese women, selecting out older women without vertebral deformity increased accuracy of the equation. This was, however, not the case in older men.