Relationship of demographic, dietary, and clinical factors to the hydration status of patients on peritoneal dialysis

Objectives: To establish which clinical factors are associated with an increased proportion of extracellular fluid (ECF) in peritoneal dialysis (PD) patients. Design: A single-center, cross-sectional analysis of 68 stable PD patients. Method: Bioelectrical impedance measurements (RJL, single frequency; RJL Systems, Clinton, Michigan, USA) of resistance and reactance were used to determine the proportion of ECF comprising total body water (TBW) in 68 stable PD patients attending for routine clearance and membrane studies. All patients underwent detailed dietetic, adequacy, and membrane function tests. Blood pressure and antihypertensive requirements were also documented. Results: Significant gender differences in body composition were observed, such that women had lower absolute TBW and fat-free mass per kilogram body weight, but proportionately more ECIF for a given TBW, mean ECF:TBW 0.5 +/- 0.03 versus 0.44 +/- 0.05, p < 0.005. In view of this, patients were split into two groups, defined as "over-" or "normally" hydrated, either by using the single discriminator (median ECF:TBW = 0.47) for the whole population, which resulted in groups distorted by gender, or by using different discriminators according to gender (women: 0.49, men 0.45). In both analyses, overhydrated patients were older, had significantly lower plasma albumin, less total fluid removal per kilogram body weight, and higher peritoneal solute transport. When split by a single discriminator, the overhydrated patients had lower sodium removal and significantly less intracellular fluid volume due to an excess of women in this group who also had less residual function and had been on dialysis longer. Using gender-specific discrimination, overhydrated patients were heavier due to expansion of the ECIF volume: 20 +/- 4.1 L versus 16 +/- 3.3 L, p < 0.001. Stepwise multivariate analysis found age (p = 0.001), albumin (p = 0.009), and fluid losses per kilogram body weight (p = 0.025) to be independent predictors of gender-adjusted hydration status. Sodium intake did not vary according to hydration status. Conclusion: Gender influences the assessment of hydration status of 1313 patients when employing bioimpedance, such that women tend to have more ECF. Taking this into account, age, albumin, and achieved fluid removal appear to be independently associated with hydration status, whereas peritoneal solute transport is not. Advice on dietary sodium should take account of hydration status and achievable losses.