Is the bioelectrical impedance method suitable for epidemiological field studies?

Body composition was measured in a group of 100 pubertal children, aged 11-16 years, by means of densitometry and bioelectrical impedance. Bioelectrical impedance was measured twice within a period of 2 weeks, once at the Department immediately before the measurement of the body density, and once during a home visit. A population-specific prediction equation for fat-free mass from body impedance was developed from the impedance value measured at the Department. The developed prediction formula had a high correlation coefficient and a relatively low standard error of estimate, and was used to predict body composition from the measurements made during the home visit. Body impedance measured at the subjects' home was significantly higher (19 +/- 25 Ohm). Individual differences were up to 70 Ohm. These intra-individual differences gave rise to large over- and underestimates of the body composition in individuals, but at a group level body composition could be precisely predicted. When using (non-population-specific) prediction formulas from the literature, the predicted mean body composition deviated from the measured body composition. The estimation error in the predicted body composition from anthropometric variables alone was slightly larger, but due to the better reproducibility of anthropometric measurements the predicted individual body composition was quite as good or even better compared to the assessment of the body composition from body impedance. It is concluded that the bioelectrical impedance may be an adequate method to assess body composition in epidemiological studies only as far as population-specific prediction formulas are used.(ABSTRACT TRUNCATED AT 250 WORDS)