Analysis of Risk Factors and Development of a Prediction Model for Cognitive Dysfunction Associated with Pediatric Sepsis

Abstract Objective To investigate the risk factors for cognitive dysfunction in pediatric patients with sepsis, to develop a visualized risk prediction nomogram model, and to validate it. Methods A total of 206 pediatric patients with sepsis admitted to the PICU of Gansu Maternal and Child Health Hospital between January 2021 and December 2024 were enrolled as study subjects. Based on their Montreal Cognitive Assessment (MoCA) scores obtained 12 months after discharge, these patients were divided into a cognitive impairment group (n = 78) and a normal cognition group (n = 128). Using a stratified random sampling method at a ratio of 7:3, the patients were further subdivided into a training cohort (n = 144) and a validation cohort (n = 62). The training cohort was used to identify risk factors for cognitive dysfunction and to establish a predictive model, while the validation cohort was used to evaluate the predictive performance of the model. Within the training cohort, differences in general clinical data and laboratory parameters between the two groups were analyzed and compared; LASSO regression was employed for variable selection; multivariate logistic regression analysis was conducted to identify independent risk factors; and a nomogram prediction model was constructed. The discriminative ability of the model was evaluated using the receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC); the calibration performance was assessed using calibration curves and the Hosmer–Lemeshow test; and the Decision Curve Analysis (DCA) was used to evaluate the clinical utility of the model. Internal validation was subsequently performed on the validation cohort. Results Among the 206 pediatric patients, 78 cases (37.86%) developed cognitive dysfunction. In the training cohort, statistically significant differences were observed between the two groups of pediatric patients regarding duration of mechanical ventilation, admission PCIS score, proportion of patients with septic shock, NE%, CRP, PCT, D-dimer, Lac, CK, Cys-C, Cr, Glu, and Ca levels (P < 0.05). Multivariate logistic regression analysis revealed that septic shock (OR = 3.267, 95% CI: 1.611–6.625), CRP (OR = 1.006, 95% CI: 1.003–1.009), Lac (OR = 1.135, 95% CI: 1.001–1.287), Cys-C (OR = 2.006, 95% CI: 1.287–3.128), and Glu (OR = 1.234, 95% CI: 1.087–1.400) were independent risk factors for cognitive dysfunction in pediatric patients with sepsis. The nomogram model constructed based on these factors demonstrated AUC values of 0.810 in the training cohort and 0.795 in the validation cohort. The Hosmer–Lemeshow goodness-of-fit test results indicated good model performance. with P = 0.973 in the training cohort and P = 0.843 in the validation cohort. The DCA curve showed that the model exhibited favorable clinical net benefit within the probability threshold range of 0.1–0.9. Conclusion Septic shock, CRP, Lac, Cys-C and Glu are independent risk factors for cognitive dysfunction in pediatric patients with sepsis. The nomogram prediction model developed based on these factors demonstrates excellent discriminative ability, calibration, and clinical utility.