Remission, relapse, and re-remission of proliferative lupus nephritis treated with cyclophosphamide

Background. Long-term intravenous cyclophosphamide (IVC) in combination with corticosteroids is standard therapy for proliferative lupus nephritis, but it has limitations. There are few data on long-term remission rates, predictors of relapse, and the ability to achieve a second remission with currently recommended IVC regimens. Methods. A cohort of 85 patients with proliferative lupus glomerulonephritis (focal N = 33, diffuse N = 52) treated with IVC was assembled in three institutions. Timing and predictors of remission, relapse, and re-remission were evaluated with Kaplan-Meier analyses and Cox models. Results. The median time to remission was 10 months, whereas an estimated 22% of patients had not remitted after 2 years. The median time to relapse among 63 patients who had achieved remission was 79 months. In multivariate models, adverse predictors of remission were a delay in the initiation of therapy from the time nephritis was clinically diagnosed [hazard ratio (HR) 0.58, P = 0.063] and a higher amount of proteinuria (HR 0.86 per 1 g/24 hours, P = 0.014). Predictors of earlier relapse for patients entering remission included a longer time to remission (HR 1.029 per month, P = 0.025), a history of central nervous system involvement (HR 8.41, P = 0.002), and World Health Organization histology (P = 0.01). Among the 23 patients who relapsed during follow-up, the median time to re-remission was 32 months, and with three exceptions, all patients took substantially longer time to remit the second time compared with their first remission (P = 0.01). The time to re-remission was longer in patients who had taken longer to remit the first time (HR 0.979 per month, P = 0.16), in patients who had relapsed earlier after the first remission (HR 1.071 per month, P = 0.002), and in those with evidence of chronicity in the original kidney biopsy (P = 0.015). Conclusions. Prolonged courses with a cumulative risk of toxicity are needed to achieve remission in many first-treated patients and in most patients treated for a second time. The optimal management of patients with identified adverse predictors of response needs further study.