Abstract
Background and Aims
Both the relative erythropoietin (Epo) deficiency and its relationship with serum hemoglobin (Hb) are widely postulated in chronic kidney disease (CKD), but the influence of chronic inflammation and iron status on serum Epo levels is still a matter of debate, with yet divergent reported results. Therefore, we aimed to assess the determinants of serum Epo in non-dialysis CKD patients.
Method
Fifty-two adults with CKD and anemia (defined as Hb <12g/dL), in stable clinical condition, never treated with erythropoiesis-stimulating agents (ESA) entered this cross-sectional, single-center study. Diabetes mellitus, active infectious and inflammatory diseases, malignancy, anemia of other causes than CKD, current immunosuppressive therapy, iron supplementation and blood transfusions in the previous six months were exclusion criteria. The subjects were mostly men (56%), elderly (two thirds over 60 years), with advanced CKD [71% in CKD stages G4-G5, median estimated glomerular filtration rate – eGFR 14.5 (95%CI 16 to 25) mL/min], moderate anemia [Hb 9.8 (95%CI 9.2 to 9.9) g/dL], and mild to moderate inflammation [C-reactive protein 6 (95%CI 9.2 to 18.4) mg/L].
Serum Epo was assessed by ELISA (Abcam® 119522). Complete blood count, reticulocyte index, peripheral blood smear, bone marrow aspiration (Perls’ stain), serum ferritin, and transferrin saturation, were used to investigate anemia and iron metabolism. Parameters of kidney disease (CKD etiology, eGFR and proteinuria), demographic data (age, gender), C-reactive protein, serum albumin, and serum hepcidin-25 (Hep-25, Bachem® commercial ELISA kit) were also analyzed.
Results
The median serum Epo of the whole cohort was 4.8 (95%CI 5.1 to 9.9) mU/mL. According to median Epo, subjects were clustered in Group 1 (below median, G1) and Group 2 (above median, G2). Estimated GFR and serum Hep-25 were lower in G1 than in G2 [10.6 (95%CI 9.7 to 20.8) vs. 26 (95%CI 19.1 to 32.8) mL/min, p=0.004, and 62.6 (95%CI 51.0 to 85.1) vs. 95.4 (95%CI 77.0 to 108.5) ng/mL, p=0.03, respectively]. All the other investigated parameters were similar in the two groups.
In bivariate analysis (Spearman rank correlation), serum Epo was positively associated only with eGFR (rs=0.40, p=0.003). Marginal associations with the percentage of bone marrow sideroblasts, as marker of the iron available for erythropoiesis (rs=0.25, p=0.08), erythrocyte mean corpuscular hemoglobin concentration (rs=−0.26, p=0.07), and reticulocyte index (rs=0.24, p=0.09) were observed. Conversely, serum Epo was not related to hemoglobin, indices of iron stores (e.g. serum ferritin and iron content in bone marrow macrophages), inflammation and nutritional status (e.g. C-reactive protein and serum albumin).
In a model of multiple linear regression which explained 14% of serum Epo variation, eGFR was the only determinant: Beta 0.14 (95%CI 0.05 to 0.23), p=0.004. Also, a binary logistic multiple regression model predicting serum Epo lower or higher than the median retained the eGFR as an independent predictor, while serum hepcidin showed only borderline significance:
Conclusion
Kidney function is the main determinant of endogenous erythropoietin level in moderately anemic patients with advanced CKD, ESA or iron naive, while serum hepcidin-25 seems to exert a limited influence.
Serum Albumin and High-Sensitivity C-reactive Protein are Independent Risk Factors of Chronic Kidney Disease in Middle-Aged Japanese Individuals: the Circulatory Risk in Communities Study
