Developing an equation for estimating body height from linear body measurements of Ethiopian adults

Measurements of erect height in older people, hospitalized and bedridden patients, and people with skeletal deformity is difficult. As a result, using body mass index for assessing nutritional status is not valid. Height estimated from linear body measurements such as arm span, knee height, and half arm span was shown to be useful surrogate measures of stature. However, the relationship between linear body measurements and stature varies across populations implying the need for the development of population-specific prediction equation. The objective of this study was to develop a formula that predicts height from arm span, half arm span, and knee height for Ethiopian adults and assess its agreement with measured height.

A cross-sectional study was conducted from March 15 to April 21, 2016 in Jimma University among a total of 660 (330 females and 330 males) subjects aged 18-40 years. A two-stage sampling procedure was employed to select study participants. Data were collected using interviewer-administered questionnaire and measurement of anthropometric parameters. The data were edited and entered into Epi Data version 3.1 and exported to SPSS for windows version 20 for cleaning and analyses. Linear regression model was fitted to predict height from knee height, half arm span, and arm span. Bland-Altman analysis was employed to see the agreement between actual height and predicted heights. P values < 0.05 was used to declare as statistically significance.

On multivariable linear regression analyses after adjusting for age and sex, arm span (β = 0.63, p < 0.001, R² = 87%), half arm span (β = 1.05, p < 0.001, R² = 83%), and knee height (β = 1.62, p < 0.001, R² = 84%) predicted height significantly. The Bland-Altman analyses showed a good agreement between measured height and predicted height using all the three linear body measurements.

The findings imply that in the context where height cannot be measured, height predicted from arm span, half arm span, and knee height is a valid proxy indicator of height. Arm span was found to be the best predictor of height. The prediction equations can be used to assess the nutritional status of hospitalized and/or bedridden patients, people with skeletal deformity, and elderly population in Ethiopia.