We would like to commend Dr. Jackson (see page 1031) on both his historical review and critical appraisal of the use of etomidate as an anesthetic induction agent. This appraisal is a good summary of the debate that has occurred over the past quarter century in regard to the safety of etomidate as an anesthetic induction agent. The observations of Ledingham and Watt1– in the early 1980s indicated that etomidate should not be used for long-term sedation in the ICU due to the mortality cost incurred from long-term adrenal suppression. The effect of etomidate on adrenal function is both dose-dependent and cumulative. A single dose of etomidate will blunt the adrenocortical axis for up to 24 h. The effect of short-term suppression of adrenal synthesis on patient outcome is, however, less clear. This clinical question, although seemingly simple, is quite complex. The net effect of etomidate as an anesthetic induction agent is the sum of several factors. Etomidate has several properties, which makes it, at least in theory, a good first-line anesthetic induction agent. The dose required to achieve unconsciousness is relatively predictable. This hypnotic effect is much more predictable than that with benzodiazepines. The onset of action is fast, essentially in one arm to brain circulation. In addition, etomidate has a short duration of action. Etomidate does not cause histamine release, which is a factor contributing to its relative hemodynamic stability (the reader is referred to an in-depth clinical review of the pharmacology of etomidate2–). One would expect that these factors would result in a mortality benefit; however, the magnitude of this benefit is unknown. The major concern regarding the use of etomidate is transient adrenal suppression. The unpublished subgroup analysis data presented in the study by Annane et al3 may provide us with a glimpse of the cost resulting from the adrenal suppression by etomidate. The fact that 68 of the 72 patients (94%) who received etomidate for the induction of anesthesia did not respond to a high-dose cosyntropin stimulation test, is consistent with other published reports of adrenal insufficiency 12 to 24 h after the administration of etomidate. The data from the study by Annane et al3 seems to indicate a significant mortality cost for etomidate anesthetic induction in septic patients. The mortality rate in the placebo-treated group was 75.7% vs 54.8% for the corticosteroid-treated group. These data indicate that the adrenal insufficiency of sepsis should be treated with the administration of stress doses of corticosteroids. The continued use of etomidate for anesthesia induction would be a clinical conundrum if this mortality effect persisted despite corticosteroid administration. From the available data, we have a presumed, but have not yet measured, mortality benefit incurred from the beneficial effects of etomidate as an anesthetic induction agent. The mortality cost of adrenal suppression by etomidate anesthesia induction seems to be completely offset by corticosteroid administration in those patients who show evidence of adrenal insufficiency. As a result, we feel that the net effect still favors the use of etomidate as an anesthesia induction agent.