External validation of the international IgA nephropathy prediction tool in a Thai cohort

To externally validate the performance of the International IgA Nephropathy Prediction Tool (IIgANPT) in a Thai cohort with biopsy-proven IgA nephropathy and to compare the predictive accuracy of model versions with and without the race variable.

We conducted a retrospective cohort study to externally validate the IIgANPT in Thai adults with biopsy-proven IgAN at a tertiary-care center. Both model versions (with and without race) were evaluated. The primary outcome was a composite of a sustained 50% decline in estimated glomerular filtration rate (eGFR) or progression to end-stage renal disease (ESKD). Model performance was assessed using discrimination (C-statistic, AUC), calibration (slope and plots), and risk stratification with Kaplan-Meier analysis.

A total of 133 patients were included, with a median follow-up of 3.4 years. During follow-up, 45 patients (33.8%) reached the renal endpoint. The model demonstrated good discrimination, with C-statistics of 0.753 (95% CI 0.68-0.82) and 0.754 (95% CI 0.69-0.82) for models with and without race, respectively. Five-year AUC values were similar (0.790 vs. 0.787). Calibration was broadly preserved, although confidence intervals around calibration slopes were wide. Risk stratification showed clear separation of kidney survival across predefined risk groups (log-rank p < 0.0001).

The IIgANPT demonstrated good discrimination and effective risk stratification in this Thai cohort of patients with biopsy-proven IgAN. Calibration point estimates were close to the ideal range, although precision was limited by wide confidence intervals. Both model versions performed similarly, suggesting that exclusion of the race variable may have limited impact on prognostic performance in this relatively homogeneous population.