A prospective, randomized drug trial compared prednisone (60 mg per day initially) to azathioprine (3 to 4 mg/kg of body weight - day initially) plus prednisone in 24 patients with life-threatening systemic lupus erythematosus. Each group contained patients matched for age, sex, disease duration, previous therapy, and clinical and laboratory features of lupus erythematosus. During a mean follow-up period of 18 to 24 months, there were no significant differences between the two groups in number of deaths, renal or extrarenal manifestations of disease, serum complement levels, DNA antibodies, antinuclear antibody titers, lupus erythematosus cells, or Coombs' antibodies. There was no convincing evidence of a steroid-sparing effect of azathioprine. Side effects attributable to steroids were equally common in both groups; infections were not increased in the combination therapy group. Azathioprine was hepatotoxic in doses of 200 mg daily or more. Azathioprine was not a useful adjunct to corticosterolds in short-term therapy of a small number of patients with severe systemic lupus.