Abstract Background and Aims Hepcidin-25 (Hep25) is a key known regulator of iron metabolism and its interactions with inflammation, iron stores and erythropoietic activity were involved in the pathogenesis of chronic kidney disease (CKD)-associated anemia. Therefore, our aim was to assess the determinants of serum Hep25 level in non-dialysis CKD patients. Method In this cross-sectional, single-center study, 52 subjects (56% men, 65±13 years) with CKD [estimated glomerular filtration rate, eGFR 14.5 (95%CI 16 to 25) mL/min] and anemia [hemoglobin, Hb 9.8 (95%CI 9.2 to 9.9) g/dL], not treated with erythropoiesis-stimulating agents (ESA) or iron in the previous 6 months, were enrolled. Patients with anemia of other causes than CKD, active infectious and inflammatory diseases, malignancy, severe hyperparathyroidism, transfusions during the last 3 months, and immunosuppressive therapy were excluded. The iron status was evaluated using both peripheral and central parameters. The iron stores were assessed by serum ferritin (Fer) and iron content in bone marrow macrophages (iMf, measured quantitively on a scale from 0 to 6). The iron available for erythropoiesis was assessed by transferrin saturation (TSAT) and the percentage of sideroblasts (%Sb). Anemia was further evaluated by a peripheral blood smear, erythrocytes indices and reticulocyte index. Serum Hep25 and erythropoietin (Epo) were assessed by ELISA (Bachem®, and Abcam® 119522, respectively). C-reactive protein (CRP), albumin, and parameters of kidney disease (eGFR, proteinuria) were also measured. Mann-Whitney, Kruskal-Wallis, Chi2 tests, Spearman bivariate correlation and multiple linear regression were used for statistical analysis. Results The median serum Hep25 of the whole cohort was 82.1 (95%CI 68.7 to 92.1) ng/mL. According to median Hep25, subjects were clustered in Group 1 (below median - G1) and Group 2 (above median - G2). %Sb and reticulocyte index had higher levels in G2 than in G1 [9 (95%CI 5 to 14) vs. 5 (95%CI 4 to 7) %, p=0.003 and 0.55 (95%CI 0.39 to 0.77) vs. 0.41 (95%CI 0.32 to 0.58), p=0.05, respectively], while the proportions of subjects with iMf suggestive for iron deficiency or iron overload were similar in G2 and G1 (38% vs. 50%, p=0.40, and 26% vs. 23%, p= 0.75, respectively). Peripheral blood smear, peripheral iron indices and all the other studied parameters were also alike. In bivariate analysis, Hep25 was positively associated both with indices of iron stores, i.e. Fer (rs = 0.30, p=0.03) and iMf (rs = 0.34, p=0.01) and indices of iron available for erythropoiesis, i.e. %Sb (rs = 0.55, p<0.001) and (marginally) with TSAT (rs = 0.26, p=0.06). Meanwhile, Hep25 was not related to serum Epo, CKD parameters or inflammation markers. In a multivariate linear regression model that explained 28% of Hep25 variation, the percentage of bone marrow sideroblasts, i.e. the tissue iron available for erythropoiesis, was the only independent determinant of Hep25: Variables entered in the first step: reticulocyte index, percentage of medullary sideroblasts (%Sb), iron content in the bone marrow macrophages (iMf), serum ferritin, and transferrin saturation Conclusion In stable patients with advanced CKD, not treated with ESA or iron, with low to moderate inflammation, serum hepcidin was related only to bone marrow iron available for erythropoiesis, suggesting that in this clinical setting the need of iron for erythropoiesis prevails over inflammation in regulation of hepcidin synthesis.