Diabetic patients under multiple injection insulin therapy (i.e., intensified insulin therapy, IIT) usually start this treatment during hospitalization. We report here on the logistics, efficacy, and safety of IIT, started in outpatients. Over 8 months, 52 type I and type II diabetics were followed up whose insulin regimens consecutively had been changed from conventional therapy to IIT. Two different IIT strategies were compared: free mixtures of regular and intermediate (12 hrs)-acting insulin versus the basal and prandial insulin treatment with preprandial injections of regular insulin, and ultralente (24 hrs-acting) or intermediate insulin for the basal demand. After 8 months HbA1 levels had decreased from 10.6% +/- 2.4% to 8.0% +/- 1.3% (mean +/- SD). There was no difference between the two regimens with respect to metabolic control; but type II patients maintained the lowered HbA1 levels better than type I patients. Only two patients were hospitalized during the follow-up time because of severe hypoglycemia. An increase of body weight due to the diet liberalization during IIT became a problem in one-third of the patients. Our results suggest that outpatient initiation of IIT is safe and efficacious with respect to near-normoglycemic control. Weight control may become a problem in IIT patients.