PLoS ONE · 2013 · 56 citations · 32 references
AgricultureNutritionNutritional SciencesChild NutritionHuman NutritionPediatric NutritionAnthropometric VariationBody CompositionChildhood ObesityNutrition AssessmentNutritional AssessmentBioelectrical Impedance MonitoringBody Composition MeasurementBody Fat PercentageEthnic Differences In ObesityFat-free MassHuman Body MeasurementsRacial Differences In ObesityWhole-body Composition
BACKGROUND: Bioelectrical impedance analysis (BIA) is a potentially valuable method for assessing lean mass and body fat levels in children from different ethnic groups. We examined the need for ethnic- and gender-specific equations for estimating fat free mass (FFM) from BIA in children from different ethnic groups and examined their effects on the assessment of ethnic differences in body fat. METHODS: Cross-sectional study of children aged 8-10 years in London Primary schools including 325 South Asians, 250 black African-Caribbeans and 289 white Europeans with measurements of height, weight and arm-leg impedance (Z; Bodystat 1500). Total body water was estimated from deuterium dilution and converted to FFM. Multilevel models were used to derive three types of equation {A: FFM = linear combination(height+weight+Z); B: FFM = linear combination(height(2)/Z); C: FFM = linear combination(height(2)/Z+weight)}. RESULTS: Ethnicity and gender were important predictors of FFM and improved model fit in all equations. The models of best fit were ethnicity and gender specific versions of equation A, followed by equation C; these provided accurate assessments of ethnic differences in FFM and FM. In contrast, the use of generic equations led to underestimation of both the negative South Asian-white European FFM difference and the positive black African-Caribbean-white European FFM difference (by 0.53 kg and by 0.73 kg respectively for equation A). The use of generic equations underestimated the positive South Asian-white European difference in fat mass (FM) and overestimated the positive black African-Caribbean-white European difference in FM (by 4.7% and 10.1% respectively for equation A). Consistent results were observed when the equations were applied to a large external data set. CONCLUSIONS: Ethnic- and gender-specific equations for predicting FFM from BIA provide better estimates of ethnic differences in FFM and FM in children, while generic equations can misrepresent these ethnic differences.
32
Multiple regression: testing and interpreting interactions
Choice Reviews Online · 1992 · 37.2K citations
A H BECKETT, M ROWLAND, P TURNER · The Lancet · 1965 · 16.9K citations
Establishing a standard definition for child overweight and obesity worldwide: international survey
Tim Cole · BMJ · 2000 · 16.5K citations · Full text
Medicine, Child Health, Pediatrics +16