A simple prediction of total body water to aid quality control in isotope dilution studies in subjects 3-87 years of age

The principal method of measuring total body water (TBW) is by isotope dilution. Also, the doubly labelled water method, which is the method of choice for measuring total energy expenditure (TEE) in free-living individuals, includes calculation of TBW as the dilution space of the tracer. TBW was measured in 261 subjects (135 males and 126 females), aged 3-87, including healthy children, children with HIV and adults with non-insulin dependent diabetes mellitus (type 2 diabetes), mild hypertension, pancreatic cancer and lung cancer, either in studies of body composition or TEE. A linear relationship was found between TBW and height in all subjects. When TBW is plotted against height cubed (Ht3, m3) the regression line can be forced through the origin. Considering only adults with 18.5>body mass index <29.9 and all children (n=220), this yielded TBW (l)=7.40 x Ht3, R2=0.95. This simple linear relationship between measured TBW and Ht3 compared favourably with other prediction methods, assuming TBW is a constant proportion (55%) of body weight and TBW predicted from height and weight (mean difference between measured and predicted TBW 0.55 l compared with -1.95 and -1.20 l, respectively). Absolute errors were greater at higher TBW, but use of a log10 transformation reduced this effect. This simple relationship of TBW with Ht3 is too crude to be used as a body composition predictor in individual subjects as it ignores, for instance, body shape. However, it can be used as a quality control tool. Here, use of a log10 transformation and residual plot can serve to identify outliers, which can be checked for gross errors in data input and if necessary samples are re-analysed.