POS-252 SERUM PHOSPHATE IS ASSOCIATED WITH FERRITIN IN NON-DIALYSIS CHRONIC KIDNEY DISEASE

Abstract

Since recent reports from the literature linked serum phosphate (PO4) and fibroblast growth factor 23 (FGF23) to anemia, and also FGF23 with low levels of ferritin and higher need for intravenous iron, the current cross-sectional study was aimed to investigate relationships between mineral metabolism and iron status parameters in chronic kidney disease (CKD) patients. Eigthy-seven adults with non-dialysis CKD [(61% males, 55% over 60 years-old, with CKD vintage 4 (95%CI 4.2 to 6) years, 46% in stage G3, 37% stage G4, and 17% stage G5] were prospectively selected from patients admitted in a single center nephrology unit. Patients with known causes of anemia, intravenous iron administration in the last month, malignancies, and active inflammatory and infectious diseases were excluded. None of the subjects were treated with erythropoiesis stimulating agents. Medical history (including data on medications), physical examination, plasma iron status (ferritin, transferrin, iron, transferrin saturation) and mineral metabolism (total calcium, PO4, intact parathyroid hormone, c-terminal FGF23, alkaline phosphatase) parameters, hemogram, serum albumin and C-reactive protein (CRP) were assessed. Data were expressed as median (95%CI) if not otherwise stated. Univariate correlations by Spearman method, and multivariate linear and logistic regression models were used for the statistical analysis. The median values of serum ferritin (Fer), transferrin saturation (TSAT), hemoglobin (Hb) and C-reactive protein were 223 (232 to 316) ng/mL, 21.3 (20.5 to 24.6) %, 12±2g/dL, and 3 (4.4 to 8.8) mg/L, respectively. The prevalence of anemia (defined as Hb<13g/dL in men and <12g/dL in women) was 51%, while absolute iron deficiency (ferritin<100ng/mL) was found in 15% and decreased iron availability (ferritin >100ng/mL and TSAT <20%) in 46%. Similar proportions of subjects treated with vitamin D or phosphate binders were encountered in subgroups defined by below/over median values for both Fer and TSAT. In bivariate analysis, Fer was negatively associated with glomerular filtration rate (rs=–0.27, p=0.01) and serum transferrin (rs=–0.34, p=0.001), while positively with serum phosphate (rs=0.35, p=0.001). The last two parameters remained independent predictors of Fer in a model of multiple linear regression which explained 15% of ferritin variation: Beta –0.002 (–0.002 to 0.000), p=0.03 for transferrin, and Beta 0.74 (0.16 to 1.33), p=0.01 for Log(PO4). In addition, Log(PO4) was independently associated with Fer above the median in the logistic regression model (Chi2 12, p=0.003). As regards TSAT, inverse univariate correlations were found with body mass index (BMI), CRP and transferrin (rs=–0.33, p=0.002, rs=–0.34, p=0.001, and rs=–0.33, p=0.002, respectively). In multivariate analysis (R2 21, p<0.001), the same variables were retained as independent predictors [Log(BMI) Beta –0.55 (–1.02 to –0.08), p=0.02, Log(CRP) Beta –0.034 (–0.063 to –0.004), p=0.02, and transferrin Beta –0.002 (–0.002 to –0.001), p<0.001]. Neither ferritin nor TSAT did correlate with other calcium-phosphate metabolism parameters. Mineral metabolism seemed to have limited correlations with iron status, assessed by currently used peripheral indices, as only serum phosphate independently predicted ferritin in this group of epoetin-näive, non-dialysis, CKD patients without overt inflammation.