The past few years have seen the incidence of adrenal insufficiency in medical and surgical intensive care units (ICUs) rising. In certain groups of ICU patients, the number of therapeutic interventions related to the adrenal response to an adrenocorticotropic hormone (ACTH) challenge is increasing. What does it all mean? Are we in a new era of adrenal disease? Or is there something about the medical intensive care unit (MICU) clinical setting that should prompt a reevaluation of our concept of adrenal insufficiency?
The diagnosis of adrenal insufficiency depends on the Cortrosyn stimulation test. In this test, the ability of the adrenal glands to secrete cortisol in response to an ACTH challenge (Cortrosyn is a synthetic form of ACTH) is measured. Cortrosyn, 250 µg, is administered as an intravenous bolus, and plasma cortisol is measured 45 and 60 minutes later. A normal response is a plasma cortisol level greater than 20 µg/dl.1 This is a simple and very reliable test. What could go wrong in such a way that this straightforward criterion for diagnosing adrenal insufficiency may no longer apply?
The most likely answer lies in the way cortisol circulates in the blood. Seventy percent of circulating cortisol is tightly bound to cortisol-binding globulin (CBG); 20% is loosely bound to albumin; and 10% is free. It is the free cortisol that is regulated by the adrenal-hypothalamic-pituitary feedback axis. At a plasma cortisol concentration of 20 µg/dl, 14 µg of cortisol is bound to CBG, 4 µg is bound to albumin, and 2 µg is free. This analysis assumes a normal plasma concentration of albuminabout 4 g/dl. What if the level falls to 2 g/dl? The CBG-bound cortisol and the free cortisol levels remain the same, but the level of albumin-bound cortisol drops from 4 µg/dl to 2 µg/dl. Thus, the total plasma cortisol level drops from 20 µg/dl to 18 µg/dljust enough to trigger the diagnosis of adrenal insufficiency.
Many, if not most, patients in the ICU are hypoalbuminemic. Relevant published evidence supports the theoretical discussion above.2 Thus, the problem may not be an increase in adrenal insufficiency but the lack of new criteria to interpret a widely used test in the setting of the ICU, where many patients have a low albumin level.
Perhaps of more interest is the idea that it is the difference between the baseline level of plasma cortisol and the level of plasma cortisol following Cortrosyn stimulation that has important clinical value, not simply the magnitude of the stimulated value. For example, it has been shown that if this difference is less than 9 µg/dl in patients with sepsis, these patients have a better outcome if treated with modest doses of intravenous glucocorticoid.3 The idea is that this situation may represent a state of partial adrenal insufficiency. A patient in such a state, however, might have a response that goes from 50 µg/dl to 52 µg/dl. This patient has plenty of circulating cortisolso much, in fact, that he or she cannot make more. Does the idea of partial adrenal insufficiency really make sense here?
There is an alternative interpretation: the smaller the difference between baseline and stimulated levels of plasma cortisol following Cortrosyn stimulation, the greater the homeostatic stress. Expressed in another way, the adrenal glands are maximally stimulated only in the worst cases. Do these patients do better with exogenous doses of glucocorticoid? The literature suggests they might. But why? Why would more cortisol be helpful when it would seem that every receptor for cortisol is occupied? The question raises the old specter of glucocorticoid resistance. Does it exist? Clinicians commonly see cases of asthma that required 40 mg of dexamethasone to break. In such cases, the asthma abates while the rest of the body suffers all of the untoward effects of too much glucocorticoid. The same phenomenon is seen in lupus and rheumatoid arthritis: local resistance to the effects of glucocorticoid. Perhaps the cardiovascular system can suffer the same fate, or perhaps in some cases, acquired glucocorticoid resistance can be general in nature. Maybe the way to identify this new disorder is with our old friend the Cortrosyn stimulation test.
Here is the larger point. Many of the tests we use on a day-to-day basis in clinical medicine were developed and standardized before there was such a thing as an MICU or a chronic critically ill person. All of these tests will have to be rethought in light of this new medical reality. There are certain to be some surprises.
http://www.acpmedicine.com/wnim/acp_0805.htm#L1
