Although cyclophosphamide and prednisolone are effective in treating systemic vasculitis, the optimum treatment regimes and duration of treatment are unknown. We randomized 54 patients aged 15-70 years (median 57.5 years) with systemic vasculitis (classical polyarteritis n = 8, microscopic polyarteritis n = 17, Wegener's granulomatosis n = 29) to treatment with either pulse cyclophosphamide and prednisolone (PCYP) (n = 24) or continuous oral and prednisolone and cyclophosphamide, with the latter followed after a median of 3 months (range 1.5-10 months) by azathioprine (CCAZP) (n = 30). Patients on CCAZP were more likely to develop leucopenia (13/30) than patients on PCYP, (7/24) although the difference was not significant. The numbers of infective episodes during follow up were comparable in the two groups at 1.7/patient for PCYP and 1.66/patient for CCAZP. Overall, 26/30 patients (87%) treated with CCAZP developed treatment-related toxicity, as did 17/24 patients (71%) treated with PCYP. After a median follow-up of 40.4 months (range 0.7-64.8), there was no difference in the frequency of deaths (PCYP 5, CCAZP 4), relapses (PCCYP 7, CCAZP 8), treatment failures (PCYP 4, CCAZP 4), improvement in disease activity scores or renal function. Survival at three years was 77% in patients treated with PCYP, and 90% in patients on CCAZP (p = 0.38). There was a tendency towards increased toxicity in patients treated with the continuous regimen.