Background/Aims: An accurate non-invasive method is required to measure body composition of infants. Ultrasound (US) and bioelectrical impedance analysis (BIA) are inexpensive methods that could fulfill this need in predicting percentage fat mass (%FM). Both have been well validated in adults and children but not in infants. This study compared these methods with reference to international normative data.
Methods: %FM was calculated from whole-body BIA and US skin-fold (SF) measurements (biceps, triceps, subscapular, suprailiac) of 2, 5, 9 and 12 m-old healthy, breast-fed term infants (n=44). %FM values were determined using 6 age-appropriate BIA equations and 3 US-derived SF equations. Data were analysed using ANOVA and probability analysis.
Results: %FM was significantly (P<0.001) lower at 2 than at 5, 9, or 12 months of ageirrespective of measurement method. Overall predictions based on US were not significantly different from BIA-based predictors (P=0.076) but there were significant differences between individual predictions. No predictor equation, US or BIA, performed well at 2 m, with <30% of predicted values falling within the 95% confidence range for the reference data. At 5, 9, 12 m, BIA-based predictors performed better than US-based predictors but with most predicting only between 40 and 68% of subjects within the 95% confidence reference range compared to 0 to 50% for US predictors.
Conclusions: Simple techniques for the assessment of body composition do not predict accurately percentage body fat in infants under 1 year of age when compared with reference data.
Funding source(s): Australian Postgraduate Award (ZG); Medela AG (Switzerland).