The aim of our study was to compare the efficacy of 3 different therapeutic protocols in the treatment of patients with WHO class IV lupus nephritis and normal renal function. We carried out a randomized prospective trial. The treatment programs consisted of a standard therapy regimen alone (protocol A), plus plasmapheresis (protocol B) or pulse methylprednisolone (protocol C), followed by a slow (protocols A and B) or fast (protocol C) prednisone tapering schedule. Statistical analysis was performed, using univariate survival analysis according to Kaplan Meier and Breslow's test to compare survival curves. Eighteen patients entered the study: 6 protocol A, 5 protocol B and 7 protocol C. No patients developed renal insufficiency. Moreover, no statistical differences in the probability of inducing partial or complete disease remission and in reducing 24-hour urinary protein excretion to < or = 2 g per day were observed among the groups. Protocols A and B were more effective in comparison with protocol C in decreasing 24-hour urinary protein excretion to < or = 0.5 g and < or = 0.2 g per day. In conclusion, a slow prednisone tapering schedule is more effective in reducing 24-hour urinary protein excretion to < or = 0.5 and < or = 0.2 g per day as compared with a fast prednisone tapering schedule, even if it is preceded by methylprednisolone pulse therapy.