Three patients admitted to the intensive care unit after multiple injury were observed to suffer episodes of adrenocortical insufficiency suggested by clinical manifestations and confirmed by appropriately low cortisol concentrations. This prompted a prospective study of pituitary-adrenocortical function in six multiply injured patients, three of whom showed evidence of adrenocortical suppression. The only factor common to the six patients with abnormally low adrenocortical function was an association between periods of adrenocortical suppression and intravenous infusion of etomidate; when the drug was stopped adrenocortical function was restored, and renewed administration of the drug caused further inhibition. Etomidate infusions lasting only six hours were found to cause low, flat responses to short tetracosactrin tests and grossly raised plasma concentrations of adrenocorticotrophic hormone, suggesting direct suppression of the adrenal cortex. Median plasma cortisol concentrations measured at 0900 were significantly lower and median plasma concentrations of adrenocorticotrophic hormone measured at 0900 were significantly higher in the three patients studied prospectively who were receiving etomidate infusions compared with the three patients who did not receive etomidate (p = 0.05).