Prediction of height using ulna length in African-Caribbean children

Standing height is an essential anthropometric measurement in pediatrics. In children unable to stand upright, measurement of ulna to predict standing height is recommended, but height prediction equations based on the ulna have not been established in children of African ancestry. We hypothesized that such equations would result in lower predicted height compared to using equations derived from non-African children. We measured prospectively standing height and both ulna in 358 African-Caribbean children without bone deformity or muscle disease, referred to two pulmonary function test laboratories. Interobserver variability was low for standing height (n = 54) and ulna measurement (n = 51) (mean biases [95%CI]: -0.02 [-0.99; 0.95] and 0.05 [-0.91; 1.01] cm, respectively), as well as inter-ulna variability (n = 352; mean bias 0.03 95%CI [-0.66; 0.73] cm). We used the mean value of 247 bilateral ulna measurements to calculate prediction equations using a generalized linear model including age, sex, ulna length, and geographic origin group, the latter showing no influence on the model. In the validation population of 107 children, the median difference [inter-quartile range] between standing height and ulna-predicted height was -0.4 [-2.7; 1.0] cm. Of 260 reliable baseline spirometry, there was a strong concordance between bronchial obstruction diagnosis established using standing height or ulna-predicted height (kappa coefficient: 0.85 [0.77; 0.94]) with only 11 (4.3%) children misclassified. The ulna predicted height calculated from African-Caribbean prediction equations resulted in a smaller height than the height calculated using equations derived from non-African children.