Longitudinal relationships between fluid status, inflammation, urine volume and plasma metabolites of icodextrin in patients randomized to glucose or icodextrin for the long exchange.

Background: Randomized trials have shown that icodextrin reduces the volume of extra-cellular fluid (ECFv) with variable effects on residual renal function. To explore this fluid shift and its possible mechanisms in more detail, prospectively collected data from one such trial, including measures of inflammation (C-reactive protein, tumour necrosis factor-alpha, albumin and low and high molecular weight hyaluronan) ANP (atrial naturetic peptide), an indirect marker of intra-vascular volume, plasma concentrations of icodextrin metabolites and alpha-amylase activity were analysed.

Methods: 50 patients were randomized to either 2.27% glucose or icodextrin (n = 28) for a long exchange following a month run in. Blood samples were obtained at -1, 0, 3 and 6 months, coincident with measurements of urine volume and fluid status.

Results: In both randomized groups, a significant correlation between the fall in ECFv and the decline in urine volume was observed (P = 0.001), although the relative drop in urine volume for patients randomized to icodextrin tended to be less. At baseline, ANP was higher in patients with proportionately more ECFv for a given body water or height. Icodextrin patients had non-significantly higher ANP levels at baseline, whereas by 3 (P = 0.026) and 6 months (P = 0.016) these differed between groups due to divergence. There was a correlation between increasing ANP and reduced ECF at 3 months, r = -0.46, P = 0.007, in patients randomized to icodextrin, but not glucose. There were no relationships between fluid status and any inflammatory markers at any point of the study, with the exception of albumin at baseline, r = -0.39, P = 0.007. Amylase activities at -1 month and baseline were highly correlated, r = 0.89, P < 0.0001. Within patients, concentrations of icodextrin metabolites were highly correlated; the only predictor of between-patient variability on multivariate analysis was body weight. There was no relationship between plasma concentrations of icodextrin metabolites and any of the other clinical parameters, including change in daily ultrafiltration, urine volume, fluid or inflammatory status.

Conclusions: This analysis supports observational data that changes in fluid status are associated with changes in urine volume. Icodextrin was not associated with a greater fall in urine output despite its larger effect on ECFv. Changes in fluid status could not be explained or did not appear to influence systemic inflammation. Nor can they be explained by individual variability in plasma concentrations of icodextrin that are in turn inversely proportional to the volume of distribution.