Assessment of total body water in paediatric patients on dialysis
Journal Articles
Overview
Research
Identity
Additional Document Info
View All
Overview
abstract
BACKGROUND: Various anthropometric techniques are used to assess total body water in children on dialysis; however, their predictive accuracy and precision has not been validated. METHODS: We compared total body water measurements obtained by deuterium oxide (D2O) dilution with predictions of total body water from (1) height and weight, (2) skinfold measurements, and (3) bioelectrical impedance analysis, using previously published formulae for healthy children. Measurements were performed in 14 patients on peritoneal and in nine patients on haemodialysis, aged 4-22 years. RESULTS: In the total population of dialysed patients, weight was the strongest single predictor of total body water (R2 = 0.93), followed by the resistance index (RI = height2/impedance; R2 = 0.85) and height (R2 = 0.93). A prediction formula based on height and weight predicted total body water with a residual mean square error (RMSE) of 1.97 1 (coefficient of variation (CV) = 10.0%) and with a systematic overestimation of true total body water by 0.4%. A prediction equation based on skinfold measurements yielded a total body water estimate with an RMSE of 2.15 1 (CV = 10.5%) and overpredicted true total body water by an average of 2.2%. Using three published prediction equations incorporating RI, RMSEs of 2.78 1 (CV = 14.1%) with a mean under- or overestimation of true total body water by 6.9, 7.1, and 0.8% respectively, were achieved. The prediction of total body water was optimized by linear combinations of RI or the log-transformed sum of four skinfolds (logsum) with weight by the following equations: total body water (1) = 9.97 - 3.13 x logsum (1) +0.59 x weight (kg) (R2 = 0.951; RMSE = 1.67 1; CV = 8.17%). total body water (1) = 1.99 + 0.144 x RI (Ohm/cm2) (2) + 0.40 x weight (kg) (R2 = 0.949; RMSE = 1.67 1; CV = 8.53%). The fit of these prediction formulae, which were derived from the total population, did not differ significantly between haemo- and peritoneal dialysis patients or between boys and girls. CONCLUSIONS: Both skinfold measurements and bioelectrical impedance analysis can be used to improve the height- and weight-based prediction of total body water in children on dialysis.
