Objective: To establish the accuracy of an eight-polar tactile-electrode impedance method in the assessment of total body water (TBW). Design: Transversal study. Setting: University department. Subjects: Fifty healthy subjects (25 men and 25 women) with a mean (s.d.) age of 40 (12) y. Methods: TBW measured by deuterium oxide dilution; resistance (R) of arms, trunk and legs measured at frequencies of 5, 50, 250 and 500 kHz with an eight-polar tactile-electrode impedance-meter (InBody 3.0, Biospace, Seoul, Korea). Results: An algorithm for the prediction of TBW from the whole-body resistance index at 500 kHz (height 2 =R 500 where R is the sum of the segmental resistances of arms, trunk and legs) was developed in a randomly chosen subsample of 35 subjects. This algorithm had an adjusted coefficient of determination (r 2 adj ) of 0.81 (P < 0.0001) and a root mean square error (RMSE) of 3.6 l (9%). Cross-validation of the predictive algorithm in the remaining 15 subjects gave an r 2 adj of 0.87 (P < 0.0001) and an RMSE of 3.0 l (8%). The precision of eight-polar BIA, determined by measuring R three times a day for five consecutive days in a fasting subject, was 2.8% for all segments and frequencies. Conclusion: Eight-polar BIA is a precise method that offers accurate estimates of TBW in healthy subjects. This promising method should undergo further studies of precision and its accuracy in assessing extracellular water and appendicular body composition should be determined. Sponsorship: Modena and Reggio Emilia University.

No abstract

Total body water (TBW) was measured by deuterium oxide (DzO) dilution and predicted from bioelectrical impedance (Z) gave a more accurate prediction of TBW by bioelectrical impedance analysis on the study subjects, with biases of -0.1 (SD 1.8) and 0.5 (SD 1.7) litres in controls and patients respectively (NS). However, the individual bias was sometimes high. It is concluded that bioelectrical impedance analysis can be used to predict TBW in anorexic women at a population level, but the predictions are less good than those based on body weight alone.

This preliminary communication reports data regarding the distribution between intracellular (ICW) and extracellular (ECW) water compartments in a group of 21 prepubertal young obese children of both sexes in comparison with a group of 18 normal children weight matched for age. Our data indicate that obesity is associated with a highly significant relative expansion of extracellular water (ECW/ICW = 0.61 +/- 0.19 and 0.76 +/- 0.09 in control and obese subjects, respectively; P < 0.0015). This observation, which has been already reported in adult women, suggests that some disturbances of water homeostasis have an early onset and stress the need for an early control of energy imbalance in children. These findings are of great concern also in the field of human body composition, suggesting the opportunity for a critical reevaluation of the assumed constancy of some human body characteristics. Body composition methodologies developed for "normal" populations would require adjustment for use in the obese population, since a considerable error would be introduced.

Objectives: To assess the reliability of bioelectric impedance analysis (BIA) for predicting total body water (TBW) and extracellular water (ECW) in obese children. Design: Comparison of ®ve prediction models based on: (i) body weight (Wt), (ii) the impedance (Z) index (ZI height 2 /Z), (iii) the association of Wt and ZI, (iv) the body surface area (SA) to impedance ratio (SA:Z) and, (v) the body volume (V) to impedance ratio (V:Z). Subjects: Thirty obese and 25 control children of 11.2 AE 1.8 y of age. Measurements: TBW and ECW were assessed by deuterium and bromide dilution; Z was measured at frequencies of 5, 50 and 100 kHz. Results: In controls, Wt explained 11% more variance of TBW than ZI (r 2 0.977, SEE 0.9 I, CV 3.8%) and the association of Wt and ZI improved the prediction of TBW only slightly (r 2 0.982, SEE 0.8 I, CV 3.5%). The SA:Z and V:Z indexes explained 6 and 33% less variance of TBW respectively as compared to Wt alone. In obese subjects, ZI explained 4% more variance of TBW than Wt (r 2 0.914, SEE 1.8 I, CV 6.4%) and the SA:Z ratio was the most accurate predictor of TBW (r 2 0.959, SEE 1.2 l, CV 4.4%). However, the increase in the explained variance of TBW associated to the use of the SA:Z ratio was of only 1% as compared to the association of ZI and Wt. The V:Z ratio explained 9% less of variance of TBW as compared to ZI. In both control and obese subjects, the association of Wt and ZI offered the best prediction of ECW (r 2 0.807, SEE 1.564 I and r 2 0.826, SEE 1.035 I, respectively). However, the values of CV were much higher in controls than in obese children (17.5% vs 8.4%) owing to their lower ECW and greater variability in ECW%. ZI was the most accurate predictor of TBW on the pooled sample (n 55; r 2 0.910, SEE 1.932 I, CV 7.4%). However, it was a poor predictor of ECW on the same sample owing to its high CV (n 55; r 2 0.866, SEE 1.806 I, CV 17.0%). Conclusions: The body surface area to impedance ratio is the most accurate predictor of TBW in obese children but the association of ZI and Wt may be of more interest when BIA is used to estimate both TBW and ECW. The impedance index offers a good prediction of TBW but not of ECW in children with different levels of fatness.