HPA suppression is a common consequence of glucocorticoid therapy, whereas overt secondary adrenal insufficiency is a rare but life-threatening condition. Prolonged hypotension and a response to adequate doses of a glucocorticoid agent are not reliable ways to assess adrenocortical function. One must also demonstrate plasma cortisol levels that are inappropriately low for the clinical situation. Hypotension in patients previously treated with glucocorticoids is caused by loss of the permissive effect of glucocorticoids on vascular tone, which may be related in turn to enhanced PGI2 production in the absence of glucocorticoids. It is not caused by mineralocorticoid deficiency. Recurrent problems of study design and interpretation have plagued this area of investigation. Any patient who has received a glucocorticoid in doses equivalent to at least 20 mg a day of prednisone for more than 5 days is at risk for HPA suppression. If the doses are closer to but above the physiologic range, 1 month is probably the minimal interval. Recovery from prolonged exposure to high doses of glucocorticoids may take up to 1 year. Pituitary function returns before adrenocortical function. Recovery from short courses of treatment (e.g., 5 days) occurs more rapidly, in about 5 days. Recovery is time-dependent and spontaneous. The rate of recovery is a function of the dose and duration of therapy before tapering is started and while the dose is being reduced. ACTH therapy does not cause adrenocortical suppression but offers no advantage over glucocorticoids, has several disadvantages, and should no longer be used. Patients on alternate day glucocorticoid therapy have some suppression of basal cortisol levels but have normal or nearly normal responses to provocative tests of adrenocortical function. The standard short ACTH stimulation test is a reliable means of assessing adrenocortical function preoperatively. The low dose (1 microgram) short ACTH test is promising but has not been sufficiently well characterized, requires serial dilutions and cannot be recommended at this time. Studies of the physiologic adrenocortical response to surgical stress provide a basis for revised dose recommendations for perioperative coverage in the patient with known or suspected HPA suppression. Recommendations of a multidisciplinary group are presented.