The concept of relative adrenal insufficiency in patients with severe sepsis continues to be controversial. This arises in part from the lack of an accepted "gold standard" for the diagnosis of adrenal insufficiency in the critically ill. Historically, assessment of adrenal function in this population has relied on measurement of plasma total cortisol level, in a blood sample taken either at random or as part of a corticotropin stimulation test. However, an alternative is to focus on the site of glucocorticoid activity within the tissues as a potentially more useful index of functional adrenal status. We review the mechanisms known to affect tissue glucocorticoid activity and examine how they may be modified by critical illness. These include both free and interstitial cortisol concentrations, intracellular cortisol generation, and glucocorticoid-receptor activity and density. Changes in these factors are not reflected in plasma total cortisol concentrations, and more sophisticated techniques, including genetic transcriptional surveys, may be required to reveal the role of glucocorticoid insufficiency in critical illness.