Development and validation of a risk prediction model for hepatorenal syndrome in hepatic failure patients based on...

This study aimed to develop and validate a novel risk prediction model for hepatorenal syndrome (HRS) in hepatic failure (HF) patients by integrating glucose-6-phosphate dehydrogenase (G6PD) activity with conventional hepatic and renal function biochemical parameters, thereby enhancing early HRS detection beyond the limitations of traditional indicators.

We performed a retrospective analysis of 264 HF patients (82 with HRS, 182 without HRS) hospitalized between July 2020 and July 2022. G6PD levels and standard hepatic/renal function biochemical parameters (ALT, AST, TBil, GGT, BUN, Scr, UA, and CysC) were assessed. Key predictors were identified via Least Absolute Shrinkage and Selection Operator (LASSO) regression, and a multivariate logistic regression model was developed. Model performance was evaluated using receiver operating characteristic (ROC) analysis, with internal validation conducted through a 70:30 training-validation split.

HRS patients exhibited significantly lower G6PD activity than non-HRS HF controls (P < 0.05). While G6PD alone showed moderate predictive value (AUC = 0.742; sensitivity 59.76%, specificity 79.12%), the composite model integrating G6PD, GGT, UA, Scr, and CysC demonstrated markedly improved discrimination, achieving AUCs of 0.960 (95%CI: 0.931-0.990) in the training cohort and 0.957 (95%CI: 0.913-1.000) in the validation cohort with both sensitivity and specificity outperforming individual indicators. The derived risk equation was Combined testing Youden = -17.038 + -0.116 x G6PD + 0.102 x GGT + 0.016 x UA + 0.040 x Scr + 3.760 x CysC.

The integration of G6PD with hepatic and renal function biochemical parameters significantly enhances HRS risk stratification in HF patients. This validated tool offers superior sensitivity and specificity for the early identification of HRS.