Body composition in Ethiopian infants: Development of accurate and reliable methods for future assessment

Start date: 1 September, 2011 End date: 6 May, 2012 Project type: Master's Thesis (prior to 2018) Project code: A13471 Countries: Ethiopia Institutions: University of Copenhagen (UCPH), Denmark Grant recipient: Rasmus Wibæk Christensen Total grant: 10,000 DKK



A continuously high prevalence of low birth weight, underweight and stunting combined with increasing rates of obesity intensified by a rapid nutritional transition make sub-Saharan African countries like Ethiopia particularly prone to nutrition related diseases. Reliable and accurate assessment of infant body composition is a crucial tool to better understand the consequences of a suboptimal nutritional status and post-natal growth, for short and long-term health, and thus imperative to meet many of the present and future public health challenges in sub-Saharan Africa.


To evaluate the accuracy of air-displacement plethysmography for assessment of infant body composition, and develop new population specific reference values of the density of fat-free mass, from a three-component model, in a large sample in a sub-Saharan African population. Furthermore, to develop age-, sex- and population specific prediction equations for estimating body composition in infancy based on the more feasible method, bio-impedance analysis.


A total of 193 healthy Ethiopian infants aged 0-6 months from Jimma Town were enrolled in a cross sectional validation study using applied body composition methods. Body composition was assessed by air-displacement plethysmography, deuterium dilution and bio-impedance analysis.


Air-displacement plethysmography consistently underestimated fat-free mass by 0.373 kg and overestimated fat percentage by 6.1 %-points, both with wide limits of agreement. Compared to previously published reference values, the densities of fat-free mass were found to be considerably lower for both sexes and in all age groups. All prediction equations improved when body weight and sex together with the impedance index were included in the model. A multiple regression equation based on impedance index, body weight and sex predicted fat-free mass, estimated by a three- component model, with an r2 and SEE of 0.90 and 0.245 kg, respectively. Furthermore, bio-impedance analysis had a high reliability based on two subsequent measurements, r2=0.94.


The findings question the accuracy of the present air-displacement plethysmography setup when validated against deuterium dilution in a sub-Saharan African population. Both methods have methodological constraints impairing their accuracy, and body composition estimates from deuterium dilution was not obtained in newborns. Bio-impedance analysis is a feasible, inexpensive and reliable tool for estimating total body water and fat-free mass in infants, provided they are calm or sleeping. The final multivariate regression equations had a good predictive accuracy, and can be used in future epidemiological studies in sub-Saharan Africa assessing trends and comparing groups in terms of body composition.