BACKGROUND: Controlled trials in lupus nephritis have demonstrated that cyclophosphamide therapy is superior to corticosteroid therapy alone. The long-term effectiveness and side-effect profiles of pulse immunosuppressive regimens warrant further study. OBJECTIVE: To define the long-term risk and benefit of monthly treatment with boluses of methylprednisolone, cyclophosphamide, or both. DESIGN: Extended follow-up (median, 11 years) of a randomized, controlled trial. SETTING: U.S. government research hospital. PATIENTS: 82 patients with proliferative lupus nephritis. MEASUREMENTS: Rates of treatment failure (defined as need for supplemental immunosuppressive therapy or doubling of serum creatinine concentration, or death) and adverse events. RESULTS: In an intention-to-treat survival analysis, the likelihood of treatment failure was significantly lower in the cyclophosphamide (P = 0.04) and combination therapy (P = 0.002) groups than in the methylprednisolone group. Combination therapy and cyclophosphamide therapy alone did not differ statistically in terms of effectiveness or adverse events. Of patients who completed the protocol (n = 65), the proportion of patients who had doubling of serum creatinine concentration was significantly lower in the combination group than in the cyclophosphamide group (relative risk, 0.095 [95% CI, 0.01 to 0.842]). CONCLUSION: With extended follow-up, pulse cyclophosphamide continued to show superior efficacy over pulse methylprednisolone alone for treatment of lupus nephritis. The combination of pulse cyclophosphamide and methylprednisolone appears to provide additional benefit over pulse cyclophosphamide alone and does not confer additional risk for adverse events.