Search

Background/Aims: Several pathological classification systems were commonly used in clinical practice to predict the prognosis of IgA nephropathy (IgAN). However, how prognostic value differs between these systems is unclear. The aim of this study was to compare the Lee grade, the Oxford classification, and the Haas classification and to find a simplified classification. Methods: We retrospectively analyzed IgAN cases diagnosed between January 2002 and December 2007. The endpoints were progression to end-stage renal disease (ESRD) or a ≥50% decline in estimated glomerular filtration rate (eGFR). The predictive capabilities were evaluated by comparing the ability of discrimination (continuous net reclassification) and calibration (Akaike information criterion [AIC]). Results: A total of 412 IgAN patients were included in the study. The average follow-up period was 80.62 ± 23.63 months. A total of 44 (10.68%) patients progressed to ESRD, and 70 (16.99%) patients showed a ≥50% decline in eGFR. All multivariate Cox regression models had limited power for high AIC values. The prognostic values of the Lee grade and the Oxford classification were higher than those of models containing only established baseline clinical indicators for progression to ESRD or a ≥50% decline in eGFR (Lee grade 0.50, 95% CI 0.21–0.74; Oxford classification 0.48, 95% CI 0.28–0.71). The prognostic value of the Haas classification was lower than that of the other pathological classification systems for progression to ESRD or a ≥50% decline in eGFR (Lee grade 0.53, 95% CI 0.23–0.92; Oxford classification 0.59, 95% CI 0.10–0.74). The prognostic value of hierarchical classification (Beijing classification) using M and T lesion was similar to the Oxford classification. Conclusions: Both the Lee grade and the Oxford classification showed incremental prognostic values beyond established baseline clinical indicators. The Haas classification was slightly inferior to the Lee grade and the Oxford classification. The hierarchical classification (Beijing classification) using less pathological parameters does not lose predictive efficiency. [ABSTRACT FROM AUTHOR]