Central adrenal insufficiency (CAI) can be difficult to diagnose, especially when the deficiency in production of adrenocorticotropic hormone (ACTH) is partial. In addition to non-specific symptoms, diagnostic testing for this disorder—including the gold standard insulin-induced hypoglycemia (IIH) test—poses challenges. This led researchers to evaluate the role of serum dehydroepiandrosterone (DHEA) and DHEA-sulfate (DHEA-S) in the work-up of suspected CAI. Their findings are the subject of this issue of Strategies.

Suspected CAI presents a diagnostic dilemma to physicians, because of both its symptomatology and the limitations of tests for this condition, according to Laure Sayyed Kassem, MD, an assistant professor of medicine at Case Western Reserve University and practicing endocrinologist at University Hospitals Case Medical Center in Cleveland. "The main challenge with CAI is that it frequently is a partial deficiency of ACTH, so the degree of suppression of glucocorticoid production can be very variable. This results in a wide range of symptom severity," she explained.

Sayyed Kassem also pointed out several testing-related issues for CAI. The gold standard IIH test can be prohibitively unsafe in several populations, including children, the elderly, and patients with cardiac disease, among others. Random cortisol tests are subject to diurnal variation and need to be timed and interpreted appropriately. In addition, low-dose (1 µg) and standard (250 µg) cosyntropin stimulation tests both have low sensitivity. Although the low-dose test has higher sensitivity, it lacks specificity and can give false-negative results, depending on the cutoffs used. For instance, with a cutoff of 18.1 µg/dL as the definition of a normal response to low-dose cosyntropin, as many as one-third of patients with IIH-confirmed CAI will have levels above the cutoff. However, with the cutoff at 22 µg/dL, the percentage of false-negative results drops considerably.

In search of better methods to identify patients with CAI, Sayyed Kassem and her co-authors studied the utility of measuring DHEA and DHEA-S (J Clin Endocrinol Metab 2012;97:3655–62). "They're produced by the adrenal cortex, and ACTH is the most potent, if not the only, stimulator for their production. Since they're adrenal cortical hormones under the control of ACTH then it would be reasonable to assume DHEA and DHEA-S could be complementary markers for adrenal cortical function. This logic is what led us to look at both as complementary markers," explained Sayyed Kassem.

The study involved 155 consecutive patients newly diagnosed with a pituitary mass and 63 healthy subjects. None of the patients had received glucocorticoids, which could impair adrenal androgen secretion. All subjects had random serum cortisol, DHEA, DHEA-S, and plasma ACTH testing at baseline. The authors defined hypothalamic pituitary adrenal (HPA) function based on random serum cortisol results, response to low-dose cosyntropin stimulation, or hypoglycemia related to IIH testing.

Participants with normal ACTH levels, serum cortisol concentrations ≥12 µg/dL, and no symptoms of adrenal insufficiency were considered to have normal HPA function. Patients with initial random serum cortisol levels <5 µg/dL had repeat morning testing, and those still with levels < 5 µg/dL were considered to have impaired HPA function. Subjects who had low-dose cosyntropin test results >23 µg/dL were considered to have normal HPA function, while those with levels <18 µg/dL were considered to have impaired HPA function. Patients with low-dose cosyntropin stimulation test results between 18–22 µg/dL underwent IIH testing, in which case the investigators defined a normal response as a peak serum cortisol ≥18.5 µg/dL, along with glucose levels < 40 mg/dL. In all, 58 subjects were found to have impaired HPA function. Of those patients, 18 had low-dose cosyntropin stimulation test results >18 µg/dL, so in most clinical practice settings they would have been considered to have normal responses. However, in comparison to both patients with normal HPA function and healthy subjects, all 58 patients with impaired HPA function had markedly lower levels of both DHEA and DHEA-S, with P values < 0.001 for all comparisons.

"Individuals with normal adrenal function had a normal response to stimulation in terms of their cortisol as well as age- and gender-appropriate levels of DHEA and DHEA-S," said Sayyed Kassem. "The patients who had abnormal adrenal function, invariably they all had low DHEA-S levels at baseline and a less-than-appropriate response with DHEA to cortisol stimulation. However, a substantial proportion of these patients, about 30 percent, had a normal response to ACTH stimulation with cortisol. So relying on cortisol alone was not going to give us the same result as the gold standard insulin tolerance test."

Sayyed Kassem elaborated that the study redemonstrated problems with the low-dose cosyntropin stimulation test in diagnosing CAI. "Repeatedly it's been shown that some patients determined to have adrenal insufficiency by the gold standard insulin tolerance test have been able to respond very normally to cosyntropin. So relying on cosyntropin alone may lead us to misdiagnose them as having normal adrenal function. We have to have other markers to support our diagnosis."

Novel findings of the study included the additive value of using either DHEA or DHEA-S in teasing out patients with CAI. In a prior study, the authors documented that patients with CAI defined by IIH results have "drastically" low DHEA-S levels, even when they have normal low-dose cosyntropin test results (Endocr Pract 2011;17:261–70).

Sayyed Kassem explained that most endocrinologists today rely primarily on low-dose cosyntropin stimulation test results in evaluating suspected CAI. However, in certain circumstances, either DHEA or DHEA-S tests can help pinpoint the diagnosis. "There is a subset of patients whose baseline cortisol levels fall in a gray zone between 6–12 µg/dL and are therefore neither low enough nor high enough to definitively describe adrenal function. In such cases, the addition of baseline DHEA-S level to the low-dose cosyntropin test can provide valuable information. Supplementary baseline and stimulated DHEA levels may also be helpful," she said. "We're not recommending that everyone have their adrenal androgens assayed, because that would be an unnecessary expense, especially if the patient's adrenal function is obvious otherwise. We suggest reserving adrenal androgen testing for those patients whose cortisol levels fall in the gray zone."

She emphasized that both DHEA and DHEA-S reference ranges should be age- and gender-specific, which is not how all labs report these results. "We see reference ranges that are not helpful in clinical practice because they do not represent the changes in adrenal androgen level as we age. It is well-documented that the values decline decade-by-decade," she said.

David Koch, PhD, who was not involved in the study, encouraged laboratorians to review their reference ranges for these tests. At Grady Memorial Hospital in Atlanta, where he is director of clinical chemistry, toxicology, and point-of-care testing, DHEA and DHEA-S testing volume is low, and both are reference tests. The reference lab reports age- and gender-specific results for both analytes.

Koch also suggested that laboratorians collaborate with physicians around CAI testing. "I always encourage laboratorians to interact with their clinicians. In this case, you would want to have a conversation with an endocrinologist or two who are caring for patients with central adrenal insufficiency and ask them if making this diagnosis has been a challenge. If they're not already measuring DHEA or DHEA-S perhaps they might consider doing so." He also is an associate professor of pathology and laboratory medicine at Emory University.

Sayyed Kassem, emphasized that laboratorians are in a good position to review with physicians the pros, cons, and caveats of DHEA or DHEA-S testing which some clinicians may not be aware of.