Diagnosing adrenal insufficiency

Experts debate glucose control in the ICUTesting, steroid use common questions in the ICU.

Given the interest from the critical care community in the concept of acquired corticosteroid insufficiency during critical illness, should all ICU patients undergo formal testing of adrenal function to determine if their treatment regimens should include corticosteroids?

Not according to Mark S. Cooper, MD, PhD, consultant and senior lecturer in endocrinology at the University of Birmingham, United Kingdom, who led a session on the topic at the Endocrine Society's annual meeting in Toronto in June. “Context matters,” he explained. Dr. Cooper is the lead author of a review article on the topic that appeared in the Feb. 20, 2003, New England Journal of Medicine.

Considering the limitations of testing for both basal cortisol levels and for adrenocorticotrophic hormone (ACTH) stimulation levels or increments in critically ill patients, evidence suggests that only patients with “clinical features suggestive of adrenal insufficiency” or those with severe sepsis or in septic shock will benefit from such tests. For these extremely sick patients, some steroid treatment that may eventually prove to have been unnecessary—a predictable result of a liberal testing strategy—will probably be “outweighed by the benefit,” said Dr. Cooper.

Otherwise, he stressed, do the tests only if symptoms suggest adrenal insufficiency. However, identifying adrenal insufficiency during critical illness is difficult, and when and how to treat it remain controversial.

While critical illness—whether trauma, infection, burns or myocardial infarction—usually increases cortisol production, sometimes that protective mechanism is inadequate. When this occurs during acute severe illness, it is called “functional adrenal insufficiency.” (A related phenomenon is called “relative adrenal insufficiency,” which is when cortisol levels remain high but are insufficient to control the inflammatory response that develops during the course of severe illness.) In these cases, treating the patient with additional corticosteroids can help.

Functional adrenal insufficiency can be difficult to determine, Dr. Cooper explained, because increases in cortisol levels in response to critical illness vary by illness type and severity. Not only is it hard to define normal ranges, he said, but tissue resistance to corticosteroids varies depending on the patient's condition. Both high and low cortisol levels are associated with poor prognosis.

Moreover, critically ill patients in the ICU sometimes have pre-existing, but unrecognized, hypoadrenalism, which can be very difficult to diagnose and is treated differently than a transient insufficiency. According to Dr. Cooper, some diagnostic clues of hypoadrenalism include hemodynamic instability despite adequate fluid resuscitation and ongoing evidence of inflammation without an obvious source that does not respond to empiric treatment.

Exposure to certain drugs can also cause adrenal dysfunction in critically ill patients. An important one is etomidate, which is widely used to induce general anesthesia and is known to suppress steroid synthesis in the adrenal cortex. If an ICU patient develops sepsis, Dr. Cooper said, it is important to ask if etomidate was administered and, if so, to start steroids without even testing.

Although short-term low-dose steroid treatment is generally safe, Dr. Cooper said, steroid use does have some downsides. It can affect the functioning of the adrenal gland for months after cessation of corticosteroid treatment. Possible dangers of steroid use include damage to the immune system, polyneuropathies and delays in weaning patients off ventilators, he said.

Despite major limitations, determinations of corticosteroid insufficiency have usually been based on random cortisol levels or results of the ACTH stimulation test.

The evidence isn't overwhelming, Dr. Cooper said, but in general it appears that critically ill patients may benefit from supplemental corticosteroid treatment if their random cortisol levels (measured at any time of the day) fall below 15 µg/dL.

Cutoff values in the ACTH stimulation tests likewise are “somewhat arbitrary,” he said, and “may not accurately reflect corticosteroid action at the tissue level.” Dr. Cooper noted as well that using results of corticotropin stimulation tests outside of the scenario of septic shock has not been studied and may not apply.

Dr. Cooper also cautioned that knowing actual levels of cortisol may not predict how well the body is fighting its acute illness.

“The threshold which best identifies persons with clinical features of corticosteroid insufficiency or who would benefit from corticosteroid replacement has yet to be determined and is likely to vary depending on the underlying illness,” Dr. Cooper said.

Another complication is that adrenal insufficiency can develop later in an illness, too, Dr. Cooper said. As a result, it is unclear when or how often in the course of an illness corticotropin stimulation tests should be done or cortisol levels measured.