Icodextrin Versus Glucose Solutions for the Once-Daily Long Dwell in Peritoneal Dialysis: An Enriched Systematic Review and Meta-analysis of Randomized Controlled Trials

Rationale & Objective The efficacy and safety of icodextrin versus glucose-only peritoneal dialysis (PD) regimens is unclear. The aim of this study was to compare once-daily long-dwell icodextrin versus glucose among patients with kidney failure undergoing PD. Study Design Systematic review of randomized controlled trials (RCTs), enriched with unpublished data from investigator-initiated and industry-sponsored studies. Setting & Study Populations Individuals with kidney failure receiving regular PD treatment enrolled in clinical trials of dialysate composition. Selection Criteria for Studies Medline, Embase, CENTRAL, Ichushi Web, 10 Chinese databases, clinical trials registries, conference proceedings, and citation lists from inception to November 2018. Further data were obtained from principal investigators and industry clinical study reports. Data Extraction 2 independent reviewers selected studies and extracted data using a prespecified extraction instrument. Analytic Approach Qualitative synthesis of demographics, measurement scales, and outcomes. Quantitative synthesis with Mantel-Haenszel risk ratios (RRs), Peto odds ratios (ORs), or (standardized) mean differences (MDs). Risk of bias of included studies at the outcome level was assessed using the Cochrane risk-of-bias tool for RCTs. Results 19 RCTs that enrolled 1,693 participants were meta-analyzed. Ultrafiltration was improved with icodextrin (medium-term MD, 208.92 [95% CI, 99.69-318.14] mL/24h; high certainty of evidence), reflected also by fewer episodes of fluid overload (RR, 0.43 [95% CI, 0.24-0.78]; high certainty). Icodextrin-containing PD probably decreased mortality risk compared to glucose-only PD (Peto OR, 0.49 [95% CI, 0.24-1.00]; moderate certainty). Despite evidence of lower peritoneal glucose absorption with icodextrin-containing PD (medium-term MD, −40.84 [95% CI, −48.09 to−33.59] g/long dwell; high certainty), this did not directly translate to changes in fasting plasma glucose (−0.50 [95% CI, −1.19 to 0.18] mmol/L; low certainty) and hemoglobin A1c levels (−0.14% [95% CI, −0.34% to 0.05%]; high certainty). Safety outcomes and residual kidney function were similar in both groups; health-related quality-of-life and pain scores were inconclusive. Limitations Trial quality was variable. The follow-up period was heterogeneous, with a paucity of assessments over the long term. Mortality results are based on just 32 events and were not corroborated using time-to-event analysis of individual patient data. Conclusions Icodextrin for once-daily long-dwell PD has clinical benefit for some patients, including those not meeting ultrafiltration targets and at risk for fluid overload. Future research into patient-centered outcomes and cost-effectiveness associated with icodextrin is needed.