Abnormalities of body composition in peritoneal dialysis patients.

OBJECTIVES: Body composition changes occur in peritoneal dialysis (PD) due to abnormalities in nutrition and hydration. We investigated abnormalities of nutrition and hydration in PD patients compared with healthy controls by measurement of total body potassium (TBK) and body water compartments. DESIGN: Cross-sectional comparison study. METHODS: We measured TBK--an indicator of body cell mass--by whole body counting, total body water (TBW) by deuterium oxide dilution, and extracellular water (ECW) by bromide dilution in 29 PD patients and 32 controls. RESULTS: The absolute mean value of TBK for PD patients was not significantly lower than in controls. The ratios of observed TBK to predicted TBK from prediction formulas were compared. Equations used were those of Boddy, Bruce, Burkinshaw, and Ellis and our own equation derived from a local control database (Leeds). Observed/predicted ratios of TBK were significantly less in PD than in control subjects for all equations. Water volumes did not differ between PD and control groups. Observed/predicted ratios for TBK in PD patients correlated with serum potassium (Boddy r = 0.355, p = 0.06; Bruce r = 0.411, p < 0.05; Burkinshaw r = 0.457, p < 0.01; Leeds r = 0.412, p < or = 0.05; Ellis r = 0.356, p = 0.06) and tended to correlate with serum albumin (Bruce r = 0.343, p = 0.07; Burkinshaw r = 0.421, p < 0.05; Leeds r = 0.357, p = 0.06; Ellis r = 0.310, p = NS). There was no relationship with serum potassium in controls. Serum albumin in PD correlated with TBK (r = 0.445, p < 0.02), TBK/height (r = 0.419, p < 0.05), TBK/weight (r = 0.554, p = 0.002), and TBK/TBW (r = 0.586, p = 0.0001). Extracellular water/intracellular water (ECW/ICW) was inversely related to TBK (r = -0.455, p < 0.02 in PD; r = -0.387, p < 0.05 in controls) and to TBK/height (r = -0.446, p < 0.02 in PD; r = -0.411, p = 0.02 in controls). TBK/weight reduced with age in PD (r = -0.445, p < 0.02), as did TBK/TBW in PD (r = -0.463, p < 0.02). ECW/ICW tended to increase with age in PD (r = 0.351, p = 0.06). CONCLUSIONS: Observed/predicted ratio of TBK is reduced in PD patients relative to healthy controls, indicating reduced body cell mass. Serum albumin and potassium reflect TBK indices in PD. Body water volumes did not differ between PD and controls, implying no overall abnormality in hydration in the PD group. However, ECW is relatively increased compared to ICW with decreasing TBK indices, suggesting relative ECW expansion with reduction in body cell mass.

Comparison of anthropometric equations for estimation of total body water in peritoneal dialysis patients.

BACKGROUND: Several formulae exist for estimating total body water (TBW). We aimed to assess their validity in peritoneal dialysis patients by comparison with TBW estimated by deuterium oxide dilution (TBW(D)). METHODS: We compared the equations of Chertow (TBW(Cher)), Chumlea (TBW(Chum)), Hume and Weyers (TBW(HW)), Johansson (TBW(J)), Lee (TBW(L)), Watson (TBW(W)) and TBW as 58% of body weight (TBW(0.58Wt)) with TBW(D) in 31 peritoneal dialysis (PD) patients and 32 controls. Estimates were compared with TBW(D) using Bland and Altman comparison. Extracellular water (ECW) was also estimated by sodium bromide dilution. RESULTS: In PD patients, mean TBW(D) was 35.04 (SD 7.84) l. Estimates were greater for TBW(Cher), TBW(Chum), TBW(HW), TBW(J) and TBW(0.58Wt). Mean TBW(L) and TBW(W) did not differ from TBW(D). Ninety-five percent limits of agreement (LOA) compared with TBW(D) (as a percentage of the mean) were similar for all of the different equations in PD patients (between +/-15.4 and +/-17.3%) except TBW(0.58Wt), which was far greater (+/-26.4%). In controls, mean TBW(D) was 37.03 (SD 6.63) l. Estimates were greater for TBW(Cher), TBW(Chum), TBW(HW), TBW(J) and TBW(0.58Wt). Mean TBW(L) and TBW(W) did not differ from TBW(D). Ninety-five percent LOA compared with TBW(D) (as a percentage of the mean) were similar for all equations in the controls, and closer than in PD patients (between +/-9.1 and +/-11.5%) except TBW(0.58Wt), which was again far greater than the other equations (+/-28.1%). TBW(HW) - TBW(D) correlated with mean TBW (r=-0.412, P<0.05 in PD and r=-0.383, P<0.05 in controls). TBW(W) - TBW(D) (r=-0.539, P<0.005) correlated with mean TBW in PD. TBW(0.58Wt) - TBW(D) correlated with body mass index (BMI) (r=0.624, P<0.0001 in PD and r=0.829, P<0.0001 in controls) and ECW/TBW (r=0.406, P<0.05 in PD and r=0.411, P<0.02 in controls). CONCLUSIONS: Predictive equations were less accurate in PD than controls. TBW(0.58Wt) was most inaccurate, with systematic overestimation of TBW with increasing BMI and ECW/TBW. There were no differences in LOA with TBW(D) for the other equations within each group.