Abstract Background and Aims Focal segmental glomerulosclerosis (FSGS) is one of the leading causes of glomerulonephritis in Korea. FSGS can be caused by heterogenous aetiologies with subdivided by primary, secondary, and genetical forms. The proposed clinicopathologic classification is dependent on amount of proteinuria, laboratory findings of possible secondary causes, and genetic profiles, however, that may not provide definite subtypes of FSGS. Therefore, it is important to search for prognostic markers of any form of FSGS. Method We enrolled a total of 20,404 adult patients who underwent native renal biopsy from 1979 to 2018 in 18 hospitals in Korea. Clinical, demographic, and laboratory data were collected in conjunction with pathologic findings. Among diagnosis of FSGS by pathologists of each hospital, we selected patients with definite segmental sclerosis by light microscopic examination including ≥10 glomeruli and without electron dense deposits by electron microscopic examination. We excluded patients with other forms of pathologic diagnosis except FSGS, with diabetes mellitus, or with follow-up period < 6 months after renal biopsy. We evaluated prognostic factors to estimate the risk of ESRD which event was found from ESRD registry of the Korean Society of Nephrology or the data from each hospital. Results Among 511 selected FSGS patients, 314 (61.4%) men were included. Mean age was 45.0 ± 17.9 years at biopsy. The laboratory finding of serum creatinine, estimated GFR, and urine protein to creatinine ratio (UPCR) was 1.38 ± 0.85 mg/dl, 71.8 ± 36.6 ml/min/1.73 m2, and 4.32 ±4.84 g/g creatinine, respectively (there were no missing data). GFR <60 ml/min/1.73 m2 was found in 41.5 % of patients and UPCR ≥ 3.0 g/g creatinine in 41.1 % of patients. During 117.5 ± 100.1 months of follow-up period, there were 97 (19.0%) patients with incident ESRD. Clinical prognostic factors to incident ESRD were estimated GFR, diastolic blood pressure, and year of renal biopsy. Pathologic prognostic factors were glomerular size, amount of global sclerosis and segmental sclerosis, severity of increase in glomerular matrix, interstitial fibrosis, interstitial inflammation, and tubular atrophy, and presence of artherosclerosis of intrarenal vessels. Adjusted with clinicopathologic factors related to incident ESRD, Cox's proportional hazard models revealed only pathologic findings of amount of glomerular global sclerosis and segmental sclerosis and presence of tubular atrophy were the independent prognostic factors (Model with 3.5 % of missing cases and -2 log likelihood ratio is 956.9, p < 0.001). GFR and UPCR showed significant relationship with incidence of ESRD, however, the significance was gone in the model adjusted with pathologic findings. AUC by ROC to estimate the event of ESRD by percentage of global sclerosis (GS) was 0.601 (range; 0.538-0.663, p = 0.002) and, by percentage of segmental sclerosis (SS), 0.679 (range; 0.616-0.743, p < 0.001). We grouped patients with three pathology categorical variables of a criterion of GS 30.9%, SS 10.5 %, and moderate degree in severity of tubular atrophy (TA), resulted in 8 groups, and summarized into 4 groups according to the relative risk to develop ESRD (Fig. 1). UPCR was a significant risk factor to estimate ESRD only in Group 4 with the criterion of 2 g/g creatinine by the Kaplan-Meier test (p = 0.034). GFR was not a prognostic factor in any pathology subgroup. Conclusion Important prognostic factors to estimate incident ESRD were pathologic findings of global sclerosis, segmental sclerosis, and tubular atrophy in FSGS patients. Significance of UPCR and GFR as prognostic factors was varied according to pathologic findings. We need to reconsider the criterion of UPCR to initiate immunosuppressive treatment and to break down the effect of immunosuppressive treatments for FSGS according to pathologic findings at renal biopsy.