Improved prediction equations for estimating height in adults from ethnically diverse backgrounds

Background and aims: When body height cannot be measured, it can be predicted from ulna length (UL). However, commonly used published prediction equations may not provide useful estimates in adults from all ethnicities. This study aimed to evaluate the relationship between UL and height in adults from diverse ethnic groups and to consider whether this can be used to provide useful prediction equations for height in practice.Methods: Standing height and UL were measured in 542 adults at seven UK locations. Ethnicity was self-defined using UK Census 2011 categories. Data were modelled to give two groups of height prediction equations based on UL, sex and ethnicity and these were tested against an independent dataset (n=180). Results: UL and height were significantly associated overall and in all groups except one with few participants (P=0.059). The new equations yielded predicted height (Hp) that was closer to measured height in the Asian and Black subgroups of the independent population than the Malnutrition Universal Screening Tool (MUST) equations. For Asian men, (Hp (cm) = 3.26 UL (cm) + 83.58), mean difference from measured (95% confidence intervals) was -0.6 (-2.4, +1.2); Asian women, (Hp = 3.26 UL + 77.62), mean difference +0.5 (-1.4, 2.4) cm. For Black men, Hp = 3.14 UL + 85.80, -0.4 (-2.4, 1.7); Black women, Hp = 3.14 UL + 79.55, -0.8 (-2.8, 1.2). These differences were not statistically significant while predictions from MUST equations were significantly different from measured height.Conclusions: The new prediction equations provide an alternative for estimating height in adults from Asian and Black groups and give mean predicted values that are closer to measured height than MUST equations.