Five patients who had received corticosteroids for periods of years experienced steroid withdrawal symptoms when attempts were made to reduce or discontinue the drugs. Summarized herein are studies of hypothalmic-pituitary-adrenocortical (HPA) function in these five people during corticosteroid withdrawal. Analysis of these data and of data in previous reports discloses four subgroups of corticosteroid withdrawal syndrome: Type I, symptomatic and biochemical evidence of HPA suppression; type II, recrudescence of the disease for which the drug was originally described; type III, dependence upon corticosteroids, either physical or psychological, with demonstrably normal HPA function and no recrudescence of underlying disease; and type IV, biochemical evidence of HPA suppression without symptoms and without recurrence of underlying disease. Any combination of types I, II and III may exist. The major conclusions are these five. (1) Some syndromes that clinically suggest HPA suppression are not. (2) Some syndromes that resemble drug-dependence are not. (3) The rapid ACTH test is a clinically useful way to assess HPA function; this test should govern the rate of corticosteroid withdrawal in the absence of steroid-treatable disease. (4) If disease is present, the rate and degree of corticosteroid withdrawal are governed by the status of the disease. (5) Patients have an unpredictable tendency to abuse corticosteroids; physicians should guard against inattentively permitting long-term, unnecessary overdosage to continue.