Relationship between height and ulna lengths in girls aged 4-17 years from Uttarakhand, Northern India

Background: Assessing nutritional status of children living in the Northern Indian state of Uttarakhand is important because the prevalence of under-nutrition is higher there than in most other parts of India (Luthra et al., 2012). Assessment usually includes height measurement. This is not always possible so alternative methods of assessing height are required and predicting this from ulna length has been suggested. This has been studied in adults and children in different locations and evidence suggests that specific equations are required due to anthropometric variation in different populations (Madden et al., 2012). To date, no studies of ulna length in Uttarakhand have been published. The aim of this study was to investigate the relationship between height and ulna length in girls in this area and to evaluate whether data could be used to derive an equation to predict height.
Methods: Sixty-nine girls aged 4-17 years from an orphanage in Uttarakhand consented to participate. Height and ulna length were measured in each participant using standard procedures. The distribution of data was tested for normality. The relationship between the two variables was assessed using Spearman’s rank for the whole group and Pearson’s correlation coefficient for a subgroup aged ≥11 years (n=42). Linear regression was used to develop a prediction equation to estimate height from ulna length in the older girls. The mean difference between measured height and values predicted using the derived equation was examined using a paired sample T test. Ethical permission was obtained.
Results: A significant positive correlation was identified between ulna length and height in all participants (r=0.935, p=0.001) and in the older girls (r=0.803, p=0.001). The following regression equation was developed to predict height in girls in the older girls:
Height [cm] = 59.425 + (3.976 x ulna length [cm]).
The mean difference between measured and predicted height was <0.1 cm (p=0.998).
Discussion: The results demonstrate a close relationship between height and ulna length in this population which agrees with previous findings from studies of adults in India and in Asian children in the UK (Gauld et al., 2004; Shah et al., 2012) but differs from those of Asian women in the UK where a weak relationship was found (Madden et al., 2012). The equation developed needs to be tested in another population to assess predictive validity before it can be considered a useful method for estimating height. The study limitations include its small sample size, uncertainty over the age of some girls and the potential for confounding due to earlier faltering growth in some participants.
Conclusion: Ulna length is closely associated with height in girls from Uttarakhand which has allowed an equation to be derived to predict height in those aged ≥ 11 years.