Utility of Columbia classification in focal segmental glomerulosclerosis: renal prognosis and treatment response among the pathological variants

Abstract Background The utility of the Columbia classification (Col-class) for focal segmental glomerulosclerosis (FSGS) has not yet been fully proven. Methods We extracted 201 FSGS patients from 10 nephrology centers in Japan and investigated the difference of a composite renal endpoint, defined as doubling of serum creatinine and/or development of end-stage renal disease, in pathological variants. Sensitivity analysis was used to prove the utility of the Col-class to predict renal outcomes. Additionally, the renal protective effects of steroids and/or immunosuppression (steroid/IS) were investigated in patients stratified according to the Col-class. Results The patients were classified into the following variants: not otherwise specified [NOS; n = 121 (60.1%)], perihilar [n = 31 (15.4%)], cellular [n = 19 (9.5%)], tip [n = 17 (8.5%)] and collapsing [n = 13 (6.5%)]. No tip variant patients reached the renal endpoint. The renal outcome in the collapsing variant was significantly poorer than that in the NOS [hazard ratio (HR) 3.71; P = 0.005]. In the sensitivity analysis, the area under the receiver operating characteristic curve for the renal endpoint was increased by adding Col-class to a model including common risk factors (P = 0.021). In a subgroup treated without steroid/IS, the outcome in the cellular variant was worse than that in the NOS (HR 5.10; P = 0.040) but the difference was not observed in the subgroup with steroid/IS (HR 0.54; P = 0.539). Conclusions The Col-class is useful to predict renal prognosis in Japanese patients with FSGS. In addition to good prognosis in the tip variant and poor in the collapsing variant, good clinical course in the cellular variant treated with steroid/IS was suggested.