The use of increased dosages of glucocorticoids during periods of physiologic stress in allograft recipients represents a clinical dilemma in that the short-term exogenous therapy required may significantly impair wound healing and immunocompetence. To investigate whether "stress steroids" are actually necessary, a prospective study was conducted in 40 renal allograft recipients admitted with significant physiologic stress. Stress categories included sepsis, metabolic abnormalities, and surgery. These patients received only their baseline prednisone immunosuppression (5-10 mg/day) and no supraphysiologic or stress doses of glucocorticoids. The clinical course of the patients revealed no evidence of adrenal insufficiency. There was no mortality, increase in hospital stay, or eosinophilia. Five episodes of hyponatremia and seven instances of hypotension were attributed to primary disease processes and responded promptly to specific treatment without steroid supplementation. Biochemical evaluation during stress revealed suppression of ACTH levels in 74.5% of episodes, elevation of urinary free cortisol levels in 79.1% of episodes, and elevation of isolated serum cortisol levels in 55.9% of episodes. This suggested that these patients had physiologically adequate adrenal function. The cosyntropin stimulation test overestimated the incidence and degree of clinically significant adrenal dysfunction (63% of patients) and was not a useful indication of a requirement for additional glucocorticoids. We conclude that functional adrenal suppression is uncommon in renal allograft recipients receiving baseline prednisone immunosuppression (5-10 mg/day) and that the demands of physiologic stress are met by a combination of endogenous adrenal function plus exogenous, baseline, immunosuppressive doses of glucocorticoids. Supra-physiologic or high doses of so-called "stress steroids" are not required. The cosyntropin stimulation test has significant clinical limitations and did not serve to alter clinical care.