Peritoneal dialysis (PD) is an established dialysis modality in patients with end stage renal disease (ESRD). However, there is growing awareness of the deleterious effect of high glucose content in PD solutions on the peritoneal membrane over time (1). Accordingly, development of new solutions to minimize glucose-induced toxicity and/or containing an alternative osmotic agent to glucose such as icodextrin and amino-acid were developed. Icodextrin is a mixture of high molecular weight, water soluble glucose polymers isolated by fractionation of hydrolyzed cornstarch (2). Unlike glucose which is absorbed from the peritoneal cavity primarily by diffusion across the peritoneal capillary endothelium, its absorption occurs mainly due to convective fluid movement out of the peritoneal cavity via the lymphatics (2). As a result, relatively constant osmotic pressure is created by icodextrin, thus it can provide sustained ultrafiltration during the long dwell.

A number of studies have reported that icodextrin-based solution provides various clinical benefits compared with conventional glucose-based solutions (3-7). In particular, icodextrin has been successfully used in the fluid management of PD patients (4-5, 7). However, excessive ultrafiltration may induce underhydration, resulting in faster decline in residual renal function. This concern was first raised by Konings et al (8). In this study, a greater fall in residual glomerular filtration rate (GFR) was observed in patients using icodextrin compared to those using 1.36% glucose solution. In contrast, contradictory findings were also reported from the two studies indicating that residual renal function can be preserved by icodextrin solution (4, 9). Although the mechanisms are not clear, possible explanation includes the presence of high-molecular-weight icodextrin metabolites in plasma, which in turn may increase plasma oncotic pressure and hence preserve plasma volume and renal perfusion as suggested by Davies et al (10). Such discrepant findings may be explained by differences in study design, baseline fluid status, and other factors affecting residual renal function during the study. In the study by Konings et al (8), the comparative solution was 1.36% glucose, whereas 2.27% glucose was used in the study by Davies et al (4). Therefore, it can be speculated that volume status might differ depending on different concentration of glucose solution, thus leading to conflicting results. Also, these two prior studies are limited by residual renal function as secondary outcome, a short follow-up duration (4 mo vs. 6 mo), and small number of patients (32 vs. 50).

To further explore the effects of icodextrin solution on residual renal function, the investigators will conduct a multicenter prospective randomized controlled open-label trial. Briefly, incident or prevalent adult CAPD patients with residual urine volume > 750 ml will be included. Patients on APD will be excluded. After a 4-week screening period, patients will be randomly assigned to icodextrin or 2.5% glucose solution for the long dwell. Residual GFR and fluid status will be assessed at baseline, 6, and 12 months. Residual GFR will be calculated as an average of urea and creatinine clearance from a 24-hour urine collection. To assess fluid status, the investigators will use three different assessment tools; 1) echocardiography for measuring intra vena cava (IVC) diameter and left ventricular end diastolic diameter, 2) measurement of plasma atrial natriuretic peptide, 3) bioimpedence analysis. Primary outcome is residual GFR change at 1-year and secondary outcome is change of fluid status during the study period. Also, biochemical laboratory data such as hemoglobin, hsCRP, plasma osmolality, and lipid profile, peritoneal equilibration test, dialysis adequacy, and daily peritoneal glucose exposure will be monitored. At least 50 subjects (a total of 100) would be required for each group to detect 50% difference of residual GFR between the two groups if type I error rate is 5% and type II error is 20% given 30% of drop-out rate during the study period.

Preservation of residual renal function is of paramount importance because it is an independent risk factor of mortality in PD patients. In addition, achievement of adequate ultrafiltration is another crucial therapeutic goal to improve clinical outcomes in these patients. In this regard, if these two goals can be accomplished by icodextrin, it would be an ideal dialysis solution in PD practice. The investigators study will address this issue to answer the unresolved question on the effect of icodextrin on residual renal function."

Inclusion Criteria:

1. Patients who were maintained on PD for over 3 months.
2. Patients with residual renal function > 750 ml/day.
3. Patients who give informed consent.

Exclusion Criteria:

1. patients less than 20 years of age,
2. uncontrolled volume status requiring the repeated use of 4.25% glucose PD solutions in addition to 2.5% glucose PD solution or icodextrin,
3. volume depletion or hypotension (systolic blood pressure < 90 mmHg) caused by 2.5% glucose PD solution or icodextrin solution,
4. allergic to starch,
5. life expectancy less than 12 months,
6. prior history of kidney transplantation or hemodialysis,
7. patients on automated PD.