OBJECTIVE To evaluate the predictors of mortality, mortality rate, and causes of death in patients with lupus nephritis (LN) depending on final renal function. METHODS The cohort included 401 Korean patients diagnosed with LN between 1985 and 2019. We retrospectively analyzed the clinical and laboratory indices, treatment response, and the final renal function. The final renal function was defined by the last stable level of eGFR measured in an out-patient department more than 3 times before death occurred and was categorized into five groups depending on CKD stage. RESULTS The median follow-up time after the diagnosis of LN was 131 months. No difference in baseline demographic characteristics and laboratory findings was found except for the proportion of Hb less than 10 mg/dl and baseline eGFR (p = 0.011 and 0.037). We found no significant differences in therapeutic parameters, but all the response parameters including treatment response at 6 months (p = 0.004) and 12 months (p = 0.004), time to remission (p < 0.001), final renal response (p < 0.001), and the final renal function (p < 0.001) differed significantly between the two groups. In multivariate Cox proportional hazards analysis, the final renal function was an independent risk factor predicting mortality. The main causes of death were infection and SLE flare. Contrary to existing knowledge, SLE flare also triggered mortality in a few patients with LN progressed to end-stage renal disease (ESRD). Only two cases of mortality occurred in the kidney transplantation (KT) group (n = 25) with a median follow-up period of 224 months. The overall mortality rates calculated using the Kaplan-Meier method were 6.8%, 10.3%, 19.7%, and 28.0% at 5, 10, 20, and 30 years, respectively. CONCLUSION Renal function deterioration was an independent determinant of mortality in Korean patients with LN. SLE flare also caused mortality in patients with LN who required maintenance dialysis, suggesting the benefit of KT on lupus activity and survival.