1. ¹Nutrition Unit, UR 106 (WHO Collaborating Centre for Nutrition), IRD (Institut de Recherche pour le Développement), Montpellier, France
2. ²Equipe de Nutrition, Laboratoire de Physiologie, Département de Biologie Animale, Faculté des Sciences et Techniques, Université Cheikh Anta Diop de Dakar, Sénégal, West Africa
3. ³Institute of Physiology, Faculty of Medicine, University of Lausanne, Switzerland

Correspondence: A Gartner, Nutrition Unit, UR 106, IRD, B.P. 64501, 911 Avenue Agropolis, 34394 Montpellier Cedex 5 France. E-mail: gartner@.ird.sn

Guarantors: A Gartner, Y Schutz.

Contributors: F Delpeuch and B Maire initiated the project comprising this study and contributed to writing the manuscript. A Gartner was in charge of the study in Senegal and was the principal investigator. A Dioum was in charge of recruitment of the subjects and data collection from ADP. A Gartner and Y Schutz were responsible for data management and analysis, and were responsible for writing the manuscript.

Received 10 March 2003; Revised 28 May 2003; Accepted 14 June 2003.

Abstract

Objective: To test the validity of a simple, rapid, field-adapted, portable hand-held impedancemeter (HHI) for the estimation of lean body mass (LBM) and percentage body fat (%BF) in African women, and to develop specific predictive equations.

Design: Cross-sectional observational study.

Setting: Dakar, the capital city of Senegal, West Africa.

Subjects: A total sample of 146 women volunteered. Their mean age was of 31.0 y (s.d. 9.1), weight 60.9 kg (s.d. 13.1) and BMI 22.6 kg/m² (s.d. 4.5).

Methods: Body composition values estimated by HHI were compared to those measured by whole body densitometry performed by air displacement plethysmography (ADP). The specific density of LBM in black subjects was taken into account for the calculation of %BF from body density.

Results: Estimations from HHI showed a large bias (mean difference) of 5.6 kg LBM (P<10^-4) and -8.8 %BF (P<10^-4) and errors (s.d. of the bias) of 2.6 kg LBM and 3.7 %BF. In order to correct for the bias, specific predictive equations were developed. With the HHI result as a single predictor, error values were of 1.9 kg LBM and 3.7 %BF in the prediction group (n=100), and of 2.2 kg LBM and 3.6 %BF in the cross-validation group (n=46). Addition of anthropometrical predictors was not necessary.

Conclusions: The HHI analyser significantly overestimated LBM and underestimated %BF in African women. After correction for the bias, the body compartments could easily be estimated in African women by using the HHI result in an appropriate prediction equation with a good precision. It remains to be seen whether a combination of arm and leg impedancemetry in order to take into account lower limbs would further improve the prediction of body composition in Africans.