Abstract
BACKGROUND: Sepsis is a severe complication in hospitalized patients with diabetic foot (DF), often associated with high morbidity and mortality. Despite its clinical significance, limited tools exist for early risk prediction. AIM: To identify key risk factors and evaluate the predictive value of a nomogram model for sepsis in this population. METHODS: This retrospective study included 216 patients with DF admitted from January 2022 to June 2024. Patients were classified into sepsis (n = 31) and non-sepsis (n = 185) groups. Baseline characteristics, clinical parameters, and laboratory data were analyzed. Independent risk factors were identified through multivariable logistic regression, and a nomogram model was developed and validated. The model's performance was assessed by its discrimination (AUC), calibration (Hosmer-Lemeshow test, calibration plots), and clinical utility [decision curve analysis (DCA)]. RESULTS: The multivariable analysis identified six independent predictors of sepsis: Diabetes duration, DF Texas grade, white blood cell count, glycated hemoglobin, C-reactive protein, and albumin. A nomogram integrating these factors achieved excellent diagnostic performance, with an AUC of 0.908 (95%CI: 0.865-0.956) and robust internal validation (AUC: 0.906). Calibration results showed strong agreement between predicted and observed probabilities (Hosmer-Lemeshow P = 0.926). DCA demonstrated superior net benefit compared to extreme intervention scenarios, highlighting its clinical utility. CONCLUSION: The nomogram prediction model, based on six key risk factors, demonstrates strong predictive value, calibration, and clinical utility for sepsis in patients with DF. This tool offers a practical approach for early risk stratification, enabling timely interventions and improved clinical management in this high-risk population.