Risk factors and a nomogram for predicting severe acute kidney injury in pediatric sepsis: a retrospective study.
Xiang Yiru Y, Zang Ping P, Chen Runfang R, Yan Haipeng H et al.
To develop and validate a nomogram for predicting the risk of severe acute kidney injury (AKI) in children with sepsis. A total of 987 children with sepsis admitted to the pediatric intensive care unit (PICU) of Hunan Children's Hospital between July 2018 and January 2021 were enrolled. Patients were stratified into a severe AKI group (n = 228) and a no clinically significant AKI group (n = 759) according to the severity of AKI during hospitalization. Based on independent risk factors identified by multivariate logistic regression, a predictive nomogram was constructed. Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). Severe AKI occurred in 228 patients (23.1%). The mortality rate in the severe AKI group was 2.57-fold higher than that in the no clinically significant AKI group (31.1% vs. 12.5%, P < 0.05). Independent predictors of severe AKI included elevated phosphate (P5+; per 1 mmol/L increase: OR = 2.789, 95% CI = 1.693-4.592, P < 0.001), decreased albumin (ALB; per 1 g/L increase: OR = 0.930, 95% CI = 0.879-0.984, P = 0.012), elevated uric acid (UA; per 1 μmol/L increase: OR = 1.004, 95% CI = 1.003-1.005, P < 0.001), and reduced antithrombin III (AT3; per 1% increase: OR = 0.990, 95% CI = 0.980-0.999, P = 0.048). The logistic regression model showed moderate discrimination with an area under the ROC curve (AUC) of 0.782 (95% CI = 0.744-0.819). The corresponding nomogram achieved an AUC of 0.761 (95% CI = 0.724-0.797). At their optimal cutoff values, the model and nomogram yielded sensitivities of 60.10% and 62.70%, specificities of 85.40% and 80.80%, and Youden indices of 0.455 and 0.435, respectively. Severe AKI is strongly associated with increased mortality in pediatric sepsis. The nomogram incorporating P5+, ALB, UA, and AT3 shows moderate discrimination, acceptable calibration, and potential clinical utility for predicting severe AKI in children with sepsis.