Predictors of outcome for severe IgA nephropathy in a multi-ethnic US cohort

Background: Although IgA nephropathy (IgAN) is the leading cause of glomerulonephritis worldwide, there are few large cohorts representative of U.S. populations. Prognosis remains challenging, particularly as more patients are treated with RAAS blockade and immunosuppression. Methods: We analyzed a retrospective cohort of IgAN patients followed at Columbia University Medical Center from 1980 to 2010. We evaluated two outcomes - halving of eGFR and ESRD - using three proportional hazards models: 1) a model with only clinical parameters, 2) a model with only histopathologic parameters, and 3) a model combining clinical and histopathologic parameters. Results: Of 154 patients with biopsy-proven IgAN, 126 had follow-up data available and 93 had biopsy slides re-read. Median follow-up was 47 months. The cohort was 64% male, 60% white, and the average age was 34 years at diagnosis. Median (IQR) eGFR and proteinuria at diagnosis were 64.1 (38.0 - 88.7) mL/min/1.73 m(2) and 2.7 (1.3 - 4.5) g/day. Over 90% of subjects were treated with RAAS blockade, and over 66% received immunosuppression. In the clinical parameters-only model, baseline eGFR and African-American race predicted both halving of eGFR and ESRD. In the histopathologic parameters-only model, no parameter significantly predicted outcome. In the combined model, baseline eGFR remained the strongest predictor of both halving of eGFR (p = 0.03) and ESRD (p = 0.001), while the presence of IgG by immunofluorescence microscopy also predicted progression to ESRD. Conclusion: In this diverse U.S. IgAN cohort in which the majority of patients received RAAS blockade and immunosuppression, baseline eGFR, African-American race, and co-staining of IgG predicted poor outcome.