Abstract: Objective To develop and validate a predictive model for renal anemia hyporesponsive to erythropoiesis- stimulating agents (ESA) in patients with chronic kidney disease not dependent on dialysis (NDD-CKD). Methods A retrospective study was conducted on 208 NDD-CKD patients with renal anemia in our hospital from January 2020 to January 2024. They were treated with ESA, and the responsiveness to ESA was assessed. Their clinical data were collected. Multivariate logistic regression was applied to analyze the influencing factors for low ESA reactivity in these patients. A nomogram prediction model for low ESA reactivity was constructed based on the regression results, and value of the nomogram was validated. Results Low ESA reactivity was observed in 45 cases (21.63%). Multivariate logistic regression showed that high neutrophil/lymphocyte ratio (NLR) (OR=1.614, 95% CI:1.063～2.451, P＜0.001) and high parathyroid hormone (PTH) (OR=1.457, 95%CI: 1.074～1.979, P＜0.001) were the risk factors, and higher estimated glomerular filtration rate (eGFR) (OR=0.547, 95% CI: 0.314～0.953, P=0.002) and higher hemoglobin (OR=0.598, 95% CI: 0.365～0.977, P=0.006) were the protective factors for low ESA reactivity in NDD-CKD patients with renal anemia. The nomogram prediction model for low ESA reactivity exhibited higher calibration (average absolute error=0.006), prediction accuracy (area under the curve=0.922) and clinical practicability. Conclusion The levels of eGFR, hemoglobin, NLR, and PTH are correlated with low ESA reactivity in NDD-CKD with renal anemia. The nomogram prediction model constructed based on the risk factors can effectively predict risk of low ESA reactivity in these patients.

Key words: Non-dialysis-dependent chronic kidney disease, Erythropoiesis-stimulating agents, Hyporesponsiveness, Nomogram prediction model