Objective. To determine the cumulative rate of relapse of lupus nephritis that has been treated successfully with cyclophosphamide (CYC), and to estimate the association between time to relapse and demographic, clinical, laboratory, and treatment variables.Methods. This was an observational study of 48 systemic lupus erythematosus (SLE) patients who were treated successfully with CYC between 1979 and 1993 and followed up thereafter a t 3 university hospitals. Demographic and clinical variables, laboratory data during the first month of nephritis, and therapy-related variables were recorded from charts. Renal biopsy specimens were retrieved and analyzed by a pathologist. Relapse of nephritis was the outcome of interest. Descriptive analysis of patients who did and those who did not have a relapse was performed by chi-square test, Fisher's exact test, and Wilcoxon 2-sample test. The cumulative rate of relapse was computed using the actuarial method. Univariate comparisons of time to relapse were computed by log-rank test. Proportional hazards modeling was used to assess the combined effect of patient characteristics that have been hypothesized to be prognostic factors.ResuZts. Nephritis relapsed in 11 patients. Previous hematologic disorder, arthritis or arthralgia, and the use of CYC in oral form were more frequent in patients who had a relapse. The cumulative rate of relapse was 25% and 46% at 5 years and 10 years, respectively. A significant univariate difference in time to relapse was found when patients were stratified by Renal disease is a major cause of morbidity and mortality in systemic lupus erythematosus (SLE) (1-4). The use of immunosuppressive agents has been associated with a better prognosis (5-12). Combined cyclophosphamide (CYC) and glucocorticoid therapy is deemed to yield significant improvement in renal survival (13-15) and positive economic impact, compared with glucocorticoid therapy alone (16). However, several questions regarding the use of CYC in lupus nephritis remain unresolved. The optimum dosage, length of therapy, and method of administration have not been fully established. Toxicity and side effects are major limiting factors.The rate of relapse of nephritis following CYC therapy is largely unknown. This issue is important because recurrent episodes of nephritis may lead to progressive renal failure. In various studies, the rate of relapse of nephritis has ranged between 0 and 50%. However, followup of patients in these studies was short or was not mentioned in the reports (11,13,(17)(18)(19). Boumpas et a1 (20) and found that short courses of intravenous CYC pulses were associated with a higher rate of disease exacerbation than were longer courses. Information regarding the influence of demographic and clinical variables, renal histology, and methods of CYC administration is lack-