Nomogram for anemia risk prediction and validation in sepsis patients

Anemia is a common sepsis complication with poor long-term outcomes, and accurate risk stratification tools are critical for clinical management. We conducted a multicenter retrospective observational study to develop and validate a nomogram for anemia in Chinese Han sepsis patients, enrolling 723 eligible patients from two tertiary hospitals. Lasso regression for parameter screening and binary logistic regression identified age, gender, APACHE Ⅱ scores, albumin (Alb), procalcitonin (PCT), and cholinesterase (ChE) as independent predictors. Based on multifactorial results of the binary logistic regression analysis in the training cohort, a nomogram was established to predict the anemia risk in sepsis patients. The nomogram exhibited excellent discriminative performance (AUC: 0.911 [95% CI, 0.880-0.936] in the training set, 0.896 [95% CI, 0.843-0.936] internally, 0.876 [95% CI, 0.801-0.931] externally) and good calibration. These findings provide a highly accurate nomogram for anemia risk prediction in this population, offering a reliable tool for clinical decision-making.