We randomly assigned 46 patients (mean age, 11.7 years; range, 4.5 to 32.8) with newly diagnosed insulin-dependent diabetes mellitus within two weeks of beginning insulin to receive either corticosteroids for 10 weeks plus daily azathioprine for one year or no immunosuppressive therapy. Half the 20 immunosuppressed patients completing the one-year trial had satisfactory metabolic outcomes (hemoglobin A1c less than 6.8 percent; stimulated peak C peptide greater than 0.5 nmol per liter; insulin dose less than 0.4 U per kilogram of body weight per day) as compared with only 15 percent of the controls. Three of 20 immunosuppressed patients, but no controls, were insulin independent at one year. Two of these continue to receive azathioprine without insulin after more than 27 months of follow-up. The response to immunosuppression correlated with older age, better initial metabolic status, and lymphopenia (less than 1800 lymphocytes per cubic millimeter) resulting from immunosuppression. The side effects of azathioprine included vomiting in one patient and mild hair loss in several others. Prednisone use resulted in a transient cushingoid appearance, weight gain, and hyperglycemia. The growth rate remained normal in all patients. We conclude that early immunosuppression with short-term use of corticosteroids plus daily azathioprine can improve metabolic control in some patients with insulin-dependent diabetes mellitus, but results from this unblinded study are preliminary and require further confirmation and long-term follow-up.