Development and validation of a routine urinalysis-based risk stratification model for IgA nephropathy in a...

IgA nephropathy (IgAN) is the most common primary glomerular disease worldwide. Routine urinalysis is universally accessible but underutilized for IgAN risk stratification. This study aimed to develop and validate a urinalysis-based risk stratification model for IgAN in a multicenter real-world hospital population.

This retrospective multicenter study was based on routine urinalysis records from 8 hospitals. A total of 354 patients had biopsy-confirmed primary IgAN, and 18,002 controls had no documented IgAN diagnosis after electronic medical record (EMR) review. The dataset was randomly divided into a training cohort (70%) and a hold-out validation cohort (30%). LASSO penalized logistic regression was used for variable selection. Model performance was evaluated by discrimination, calibration, and decision curve analysis (DCA). Bootstrap validation with 1000 resamplings was performed. An 8-variable simplified model was additionally developed.

LASSO regression selected 13 variables for the primary full model. The primary model achieved an AUC of 0.849 (95% CI: 0.824-0.873) in the training cohort and 0.861 (95% CI: 0.833-0.890) in the validation cohort. Calibration was good in the training cohort (Brier score = 0.017), while moderate calibration deviation was observed in the validation cohort (Brier score = 0.020). The model yielded a high negative predictive value (>0.99). The 8-variable simplified model retained favorable discrimination with a validation AUC of 0.840.

This routine urinalysis-based model may serve as an accessible laboratory-based risk stratification tool for identifying patients with relatively low or elevated predicted probability of IgAN in real-world hospital settings. It should be interpreted as an adjunct to clinical evaluation rather than as a stand-alone diagnostic test or replacement for renal biopsy.