Total body water reference values and prediction equations for adults

Background. The clinical interpretation of total body water (TBW) necessita tes the availability of timely comparative reference data. The prediction o f TBW volume in renal disease is critical in order to prescribe and monitor the dose of dialysis in the determination of Kt/V. In clinical practice, u rea distribution (V) is commonly predicted from anthropometric equations th at are several decades old and for white patients only. This article presen ts new reference values and prediction equations for TBW from anthropometry for white and black adults. Methods. The study sample included four data sets, two from Ohio and one ea ch from New Mexico and New York, for a total of 604 white men, 128 black me n, 772 white women, and 191 black women who were 18 to 90 years of age. The TBW concentration was measured by the deuterium or tritium oxide dilution method, and body composition was measured with a Lunar DXA machine. An all- possible-subsets of regression was used to predict TBW. The accuracy of the selected equations was confirmed by cross-validation. Results. Blacks had larger TBW means than whites at all age groups. The 75t h TBW percentile for whites approximated the TBW median for blacks at most ages. The white men and black men and women had the largest TBW means ever reported for healthy individuals. The race- and sex-specific TBW prediction equations included age, weight, and stature, with body mass index (BMI) su bstituted for weight in the white men. The root mean square errors (RMSEs) and standard errors for the individual (SEIs) ranged from approximately 3.8 to 5.0 L for the men and from 3.3 to 3.6 L for the women. In both men and women, high values of TBW were associated with high levels of total body fa t (TBF) and fat-free mass (FFM). Conclusion. TBW in these healthy adults is relatively stable through a larg e portion of adulthood. There are significant race and sex differences in T BW. These accurate and precise equations for TBW provide a useful tool for the clinical prediction of TBW in renal disease for white and black adults. These are the first TBW prediction equations that are specific for blacks.