Objective: To evaluate the effect of different methods (plateau or back extrapolation) of calculating total body water (TBW) from deuterium dilution in an environment characterised by high water turnover. The back extrapolation model is assumed to be more accurate when water turnover is high.
Design: Cross-sectional study with measurement of body composition by deuterium dilution and conventional anthropometry.
Setting: Rural Sarawak, Malaysia.
Subjects: 24 adults of the Iban population.
Results: TBW was significantly different by method of calculation (P < 0.0001), and mean fat free mass was lower by 1.3 kg using the back extrapolation technique, equivalent to a mean 3.1 (s.d. 0.8)% reduction. The same 1.3 kg difference was equivalent to a mean 16.6 (s.d. 12.3)% increase in fat mass using the back extrapolation technique. Back extrapolation values were used in subsequent analyses. Percentage fat correlated significantly with BMI and skinfold thicknesses in females, but only with BMI in males. Fat mass was significantly correlated with skinfolds and BMI in both sexes. Fat free mass was correlated with BMI in both sexes.
Conclusions: The back extrapolation method produced values for fat mass that differed substantially from those obtained by the plateau method. The difference between calculation methods could be lessened by using saliva samples in place of urine. Back extrapolation values for body fat correlated well with anthropometric measurements in females, less well in males. This difference is due to characteristics of the study population.