Serum retinol-binding protein concentration and its association with components of the uremic metabolic syndrome in nondiabetic patients with chronic kidney disease stage 5

Introduction: Chronic kidney disease (CKD) is associated with insulin resistance also in the absence of overt diabetes mellitus. The liver-derived transport protein retinol-binding protein (RBP) has recently been proposed as a novel adipokine involved in the metabolism of glucose. Although RBP is elevated in type 2 diabetics with mild CKD, its role in advanced CKD is not well studied. We hypothesized that altered RBP levels in CKD could be one factor contributing to the uremic insulin resistance.

Patients and methods: In a cross-sectional study, we evaluated 141 nondiabetic stage 5 CKD patients (GFR 6.8 +/- 2.0 ml/min; 62% males, mean age 52 +/- 11 years) close to the start of renal replacement therapy. We studied circulating RBP (RIA), retinol and metabolic markers. Body composition was also assessed using DEXA and patients were divided according to truncal fat mass above (obese) or below (lean) the sex-specific median. A fasting plasma glucose > or =6.1 mM was defined as impaired glucose tolerance (IGT).

Results: Serum RBP levels were significantly elevated in CKD as compared to previous reports in non-renal patients. Whereas levels of RBP did not differ between lean and obese patients without IGT, they were lower in lean CKD patients with IGT (5.9 +/- 2.9 microM) than in obese CKD patients with IGT (7.0 +/- 2.9 microM; p < 0.05). While RBP did not correlate with truncal or total fat mass or biomarkers of inflammation, in univariate analysis, we found weak correlations with HbA1c% (rho = 0.17; p < 0.05), fasting serum triglycerides (rho = 0.20; p < 0.001) and fasting apolipoprotein (Apo) A1 (rho = 0.29; p < 0.001). RBP also correlated negatively with ApoB (rho = -0.29; p < 0.001). In multivariate analysis, RBP was a significant and independent predictor of both HbA1c% and ApoA1 levels. Finally, RBP was strongly correlated with serum retinol, and calculating a retinol/RBP index further strengthened the observed correlations with HOMA-IR and HbA1c%.

Conclusions: RBP is elevated in nondiabetic stage 5 CKD and correlates weakly with HbA1c and ApoA1. As RBP is thought to induce insulin resistance and directly affect lipoprotein metabolism in other disease states, these findings may support a role for RBP in contributing to the uremic metabolic syndrome, putatively by altering ApoA1 metabolism, but further studies are needed to test this hypothesis.