| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Division of Endocrinology and Metabolism, Department of Internal Medicine, Emory University School of Medicine (L.M.T.), Atlanta, Georgia 30322; and the Division of Endocrinology and Diabetes, Department of Internal Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee 37232
Address all correspondence and requests for reprints to: Leonard M. Thaler, M.D., Division of Endocrinology, 1639 Pierce Drive, Room 1301 WMRB, Atlanta, Georgia 30322. E-mail: lthaler{at}emory.edu
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
Many other tests, of course, are also used in the evaluation of AI. Excellent reviews are available for detailed descriptions (1, 2). The insulin tolerance test (ITT), considered by many to be the gold standard, and metyrapone test are considered extremely accurate in the diagnosis of AI. They are physician intensive, however, and are contraindicated in certain patients. To avoid these tests, the rapid ACTH stimulation test is commonly used. Lack of continuous trophic stimulation of the adrenal glands by endogenous ACTH in patients with pituitary disease leads to adrenal atrophy and hyporesponsiveness to exogenous ACTH. Thus, the ACTH stimulation test is an indirect measurement of pituitary function. The conventional rapid ACTH stimulation test is performed by administering an iv or im injection of 250 µg ACTH; cortisol levels are determined at 0, 30, and 60 min. An ACTH-stimulated cortisol level above 500 nmol/L constitutes a normal response. This test has been described as "... the ideal method for evaluating adrenal function in all cases except those involving recent hypothalamic and/or pituitary dysfunction" (2). Although this test is highly specific, its sensitivity has been questioned (3, 4, 5, 6, 7, 8).
In recent years, several investigators have published substantial evidence for a more sensitive ACTH stimulation test using a lower dose of ACTH (1 µg) (4, 5, 6, 7). It is performed in a similar manner as the higher dose ACTH stimulation test, thereby retaining its ease, safety, and low cost. A serum sample for cortisol determination is obtained 30 min after the iv injection of 1 µg ACTH; a value over 500 nmol/L is considered a normal response. A baseline value is not necessary, and the test may be performed at any time of day. The basis for its superior accuracy in the diagnosis of AI may be that very low levels of circulating ACTH keep the adrenal cortex in a state capable of transiently responding to a supraphysiological, but not a physiological, dose of ACTH (8). The entire stored pool of endogenous ACTH in the anterior pituitary is approximately 600 µg (9). Thus, the conventional 250-µg dose is expected to be quite an adrenal stimulus. Far lower doses of ACTH, such as 1 µg, may provide a more realistic level of stimulation to which only normal adrenals can respond.
| A Review of the Literature |
|---|
|
|
|---|
For reasons that are unclear, the practice of using 250 µg for an ACTH stimulation test has persisted to this day and has become the standard of evaluation. The realization that much lower doses of ACTH will maximally stimulate cortisol secretion has been slow to translate into an updated clinical test for AI. Major endocrinology textbooks recommend the 250-µg test as the primary screening tool for AI, describing it as the most convenient and reliable test, except in cases of recent pituitary injury when the adrenal glands would still be expected to respond to an exogenous dose of ACTH (14).
Several reports document normal responses to 250 µg ACTH in patients in whom a diagnosis of AI was established by the ITT or metyrapone stimulation test (3, 5, 7, 8, 15). Cunningham et al. compared the effectiveness of the 250-µg ACTH stimulation test with either the metyrapone test or ITT (3). Of 20 cases of AI diagnosed with either metyrapone or the ITT, only 8 were detected by the conventional ACTH stimulation test. In 1987, Lindholm published a study reevaluating the clinical utility of the 250-µg ACTH stimulation test (15). Of 162 patients who achieved adequate hypoglycemia with the ITT, 26 were found to have AI. Importantly, testing with 250 µg ACTH missed the diagnosis in 7 of these 26 patients. However, 2 of the 7 patients had evidence of recent pituitary injury and should be eliminated from the analysis. Regardless, relying on the standard ACTH stimulation test would have failed to identify approximately one fifth of cases of true AI.
In 1991, Dickstein et al. reported results of low dose ACTH stimulation (4). They noted that in normal subjects, 30 min cortisol values were similar after the injection of 250, 5, or 1 µg ACTH. They also analyzed six patients receiving chronic steroids who had normal cortisol responses to 250 µg ACTH. Five of them had subnormal cortisol responses to 1 µg ACTH. Unfortunately, no gold standard test was performed to confirm the diagnosis of AI in these patients. Tordjman et al. compared various doses of ACTH with either metyrapone or ITT (5). They studied a control group, a group documented to have an impaired hypothyroid-pituitary-adrenal (HPA) axis due to pituitary disease, and a group with similar pituitary pathology but with a normal HPA axis. Using 1 µg ACTH, all subjects with AI were identified. However, 7 of 10 cases of AI were missed using the 30 min cortisol values after the administration of both the 5- and 250-µg doses. Interestingly, 2 patients in the group with known pituitary disease but normal HPA axis had abnormal responses during the 1-µg ACTH test. These findings may suggest that the low dose test is an even more sensitive indicator of adrenal dysfunction than the metyrapone test or ITT. As Streeten et al. have shown, reliance on the conventional ACTH stimulation test and resultant failure to make a diagnosis of AI can have serious consequences for affected patients (8).
Recent data also confirm the specificity of the 1-µg ACTH stimulation test (16). One microgram of ACTH has produced normal cortisol responses after 30 min in over 130 healthy subjects. This report demonstrates a low false positive rate, and hence high specificity, in addition to an extremely high sensitivity.
| Further Analysis |
|---|
|
|
|---|
The LR for a positive test result [(+)LR] is calculated most easily by the formula, sensitivity/(1 - specificity). The LR for a negative test result [(-)LR] is calculated using the formula, (1 - sensitivity)/specificity. Once the LRs are calculated, and the pretest probability of disease is estimated by the physician, calculating the posttest probability of disease is straightforward. First, the pretest probability is converted to odds using the formula: [probability/(1 - probability)] = odds. This result multiplied by the LR yields the posttest odds. The odds are then converted into the posttest probability using the formula: odds/(odds + 1) = probability. Employing these calculations, one can show that a LR for a positive or negative test result of 1 does not change the pretest probability. LRs of greater than 10 and less than 0.1 generate large changes from pretest to posttest probability. LRs of 510 and 0.10.2 generate moderate changes from pretest to posttest probability. LRs of 25 and 0.20.5 generate small (but often important) changes in probability (18). Increases in LRs do not yield linear increases in probability; the relationship is more logarithmic (19). For example, if a clinicians pretest probability of disease is 50%, then a change in (+)LR from 2 to 20 improves posttest probability from 67% to 95%, whereas a change in (+)LR from 20 to 200 improves posttest probability from 95% to 99.5%.
The several studies cited above provide enough information to calculate
LRs for both the conventional and low dose ACTH stimulation tests. When
a comparison of LRs is performed, inclusion of confidence intervals
(CIs) for the LRs provides a more meaningful interpretation (20).
Because the conventional ACTH stimulation test is extremely reliable
when a positive result occurs, its (+)LR will be consistently high.
However, given the reports of false negative results, its (-)LR may be
less powerful. When investigating a condition such as AI, false
negative results have serious repercussions. Therefore, the (-)LR will
be the focus of this analysis and the means to compare the performances
of the 250- and 1-µg ACTH tests (Table 1
).
|
(95% CI, 0.85,
), and the (-)LR is 0.58
(95% CI, 0.39,0.85). As reported above, a (-)LR in this range only
confers a small decrease from the pretest probability of disease. For
example, if the estimated pretest probability of having AI were 50%
and the stimulated cortisol level was more than 500 nmol/L, the
calculations described above compute a posttest probability of 37%;
thus, even with a normal ACTH stimulation test, the chances of real AI
remain high. Lindholms study of the 250-µg ACTH stimulation test is perhaps one of the strongest in favor of its accuracy (15). After eliminating the two patients with evidence of recent pituitary injury from the analysis, the (+)LR is 108 (95% CI, 15.1,767), and the (-)LR is 0.21 (95% CI, 0.10,0.46). These results indicate assurance of the diagnosis when the test is positive due to the markedly high (+)LR. For instance, if the pretest probability of AI is 50%, a positive test result magnifies this to 99%. Conversely, given the moderately low (-)LR, a negative test result provides a less striking change from pretest probabilities. For instance, given the same pretest probability of 50%, a negative test result reduces this to 17%, clearly not ruling out the presence of AI.
At least two published reports permit calculations of LRs for both the
conventional and low dose tests (Table 1
). Using the data from Tordjman
et al. (5) for the 250-µg test the (+)LR is
(95% CI,
0.62,
), and the (-)LR is 0.70 (95% CI, 0.47,1.06). Corresponding
values for the 1-µg ACTH stimulation test are 8.0 (95% CI,
2.18,29.25) and 0 (95% CI, 0.0,0.81). For Rasmuson et al.
(7), the (+)LR for the 250-µg test is 9.63 (95% CI, 1.47,62.95), and
the (-)LR is 0.14 (95% CI, 0.04,0.51). For the 1-µg test, the (+)LR
is
(95% CI, 1.45,
), and the (-)LR is 0.06 (95% CI,0.01,0.42).
The data from these two studies may be combined to increase the power
of a comparison. The integrated data for the 250-µg test reveal a
(+)LR of 17.65 (95% CI, 2.53,123.29) and a (-)LR of 0.36 (95% CI,
0.21,0.61). Corresponding data for the 1-µg test reveal a (+)LR of
12.98 (95% CI, 3.41,49.37) and a (-)LR of 0.04 (95% CI, 0.01,0.28).
As expected, CIs for the (+)LRs overlap because both tests are accurate
when a positive result occurs. CI values for the (-)LRs slightly
overlap, probably because the sample size is limited. A 90% CI is
0.008,0.21 for the (-)LR of the 1-µg test and 0.23,0.56 for the
250-µg test. These CIs do not overlap and indicate a significantly
better (-)LR for the 1-µg ACTH test with 90% confidence. This
translates to greater assurance of normal adrenal function with a
normal test result. Applying the same example used in the analysis of
Lindholms study illustrates this point well. If the pretest
probability of AI is 50%, a negative result with the 250-µg test
reduces this to 26%; however, a negative result with the 1-µg test
would reduce this to 4%.
| Conclusions and Recommendations |
|---|
|
|
|---|
We propose that the 1-µg ACTH stimulation test replace the standard 250-µg test when evaluating for central AI. A cortisol level below 500 nmol/L after 30 min signifies adrenal dysfunction. A baseline level is unnecessary, and the test may be performed at any time of day. If this low dose test results in a borderline value, for instance a 30 min cortisol value of 450500 nmol/L, and the clinical picture is not consistent with AI, then an ITT should be performed. This strategy should result in an extremely high diagnostic accuracy. The 1-µg test should not be used if recent pituitary injury is suspected. Pharmaceutical companies should be encouraged to provide synthetic ACTH in 1-µg vials to facilitate testing, and until that time, extreme care is warranted when diluting a 250-µg vial. If too little of the drug is administered, the 30 min cortisol value may appear low, and a patient may be falsely suspected of having AI and thus subjected to unnecessary treatment with corticosteroids. If too much of the drug is inadvertently administered, the 30 min cortisol value may be normal (as seen with 250 µg ACTH), and a patient with true AI may have potentially life-saving treatment withheld. Our current policy is to add 250 µg ACTH to 250 mL 0.9% saline. One milliliter is then removed from the solution for iv injection. Preferably, a 1-mL syringe connected directly to a needle is used to inject the solution into a vein; this should be done because additional plastic tubing may reduce the amount of ACTH delivered (21). Although the stability of diluted ACTH has been established (4), our practice has been to discard the solution at the end of each day.
| Acknowledgments |
|---|
Received December 9, 1997.
Revised March 13, 1998.
Accepted May 11, 1998.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
I. Wallace, S. Cunningham, and J. Lindsay The diagnosis and investigation of adrenal insufficiency in adults Ann Clin Biochem, September 1, 2009; 46(5): 351 - 367. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Giordano, M. Balbo, A. Picu, L. Bonelli, R. Berardelli, A. Falorni, E. Ghigo, and E. Arvat Corticotrope hypersecretion coupled with cortisol hypo-responsiveness to stimuli is present in patients with autoimmune endocrine diseases: evidence for subclinical primary hypoadrenalism? Eur. J. Endocrinol., September 1, 2006; 155(3): 421 - 428. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Agha, J. W. Tomlinson, P. M. Clark, G. Holder, and P. M. Stewart The Long-Term Predictive Accuracy of the Short Synacthen (Corticotropin) Stimulation Test for Assessment of the Hypothalamic-Pituitary-Adrenal Axis J. Clin. Endocrinol. Metab., January 1, 2006; 91(1): 43 - 47. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Agha, B. Rogers, M. Sherlock, P. O'Kelly, W. Tormey, J. Phillips, and C. J. Thompson Anterior Pituitary Dysfunction in Survivors of Traumatic Brain Injury J. Clin. Endocrinol. Metab., October 1, 2004; 89(10): 4929 - 4936. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Miozzari and P. M. Ambuhl Steroid withdrawal after long-term medication for immunosuppressive therapy in renal transplant patients: adrenal response and clinical implications Nephrol. Dial. Transplant., October 1, 2004; 19(10): 2615 - 2621. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. E. George, V. Sharma, J. Jacobson, S. Simon, and A. J. Nopper Adverse Effects of Systemic Glucocorticosteroid Therapy in Infants With Hemangiomas Arch Dermatol, August 1, 2004; 140(8): 963 - 969. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Giordano, M. Pellegrino, S. Oleandri, M. Baldi, M. Balbo, S. Laureti, A. Falorni, E. Ghigo, and E. Arvat Adrenal Sensitivity to Adrenocorticotropin 1-24 Is Reduced in Patients with Autoimmune Polyglandular Syndrome J. Clin. Endocrinol. Metab., February 1, 2004; 89(2): 675 - 680. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. P. Nasrallah and B. M. Arafah The Value of Dehydroepiandrosterone Sulfate Measurements in the Assessment of Adrenal Function J. Clin. Endocrinol. Metab., November 1, 2003; 88(11): 5293 - 5298. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. I. Dorin, C. R. Qualls, and L. M. Crapo Diagnosis of Adrenal Insufficiency Ann Intern Med, August 5, 2003; 139(3): 194 - 204. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Schmiegelow, U. Feldt-Rasmussen, A. K. Rasmussen, M. Lange, H. S. Poulsen, and J. Muller Assessment of the Hypothalamo-Pituitary-Adrenal Axis in Patients Treated with Radiotherapy and Chemotherapy for Childhood Brain Tumor J. Clin. Endocrinol. Metab., July 1, 2003; 88(7): 3149 - 3154. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Bone, M. Diver, A. Selby, A. Sharples, M. Addison, and P. Clayton Assessment of Adrenal Function in the Initial Phase of Meningococcal Disease Pediatrics, September 1, 2002; 110(3): 563 - 569. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. J. Inder and P. J. Hunt Glucocorticoid Replacement in Pituitary Surgery: Guidelines for Perioperative Assessment and Management J. Clin. Endocrinol. Metab., June 1, 2002; 87(6): 2745 - 2750. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Karlsson, J. Kallio, K. Irjala, S. Ekblad, J. Toppari, and P. Kero Adrenocorticotropin and Corticotropin-Releasing Hormone Tests in Preterm Infants J. Clin. Endocrinol. Metab., December 1, 2000; 85(12): 4592 - 4595. [Abstract] [Full Text] |
||||
![]() |
E. Arvat, L. Di Vito, F. Lanfranco, M. Maccario, C. Baffoni, R. Rossetto, G. Aimaretti, F. Camanni, and E. Ghigo Stimulatory Effect of Adrenocorticotropin on Cortisol, Aldosterone, and Dehydroepiandrosterone Secretion in Normal Humans: Dose-Response Study J. Clin. Endocrinol. Metab., September 1, 2000; 85(9): 3141 - 3146. [Abstract] [Full Text] |
||||
![]() |
R. Rossi, L. Tauchmanova, A. Luciano, M. Di Martino, C. Battista, L. Del Viscovo, V. Nuzzo, and G. Lombardi Subclinical Cushing's Syndrome in Patients with Adrenal Incidentaloma: Clinical and Biochemical Features J. Clin. Endocrinol. Metab., April 1, 2000; 85(4): 1440 - 1448. [Abstract] [Full Text] |
||||
![]() |
E. J. Nye, J. E. Grice, G. I. Hockings, C. R. Strakosch, G. V. Crosbie, M. M. Walters, and R. V. Jackson Comparison of Adrenocorticotropin (ACTH) Stimulation Tests and Insulin Hypoglycemia in Normal Humans: Low Dose, Standard High Dose, and 8-Hour ACTH-(1-24) Infusion Tests J. Clin. Endocrinol. Metab., October 1, 1999; 84(10): 3648 - 3655. [Abstract] [Full Text] [PDF] |
||||
![]() |
M. Zarkovic, J. Ciric, M. Stojanovic, Z. Penezic, B. Trbojevic, M. Drezgic, and M. Nesovic Optimizing the Diagnostic Criteria for Standard (250-{micro}g) and low Dose (1-{micro}g) Adrenocorticotropin Tests in the Assessment of Adrenal Function J. Clin. Endocrinol. Metab., September 1, 1999; 84(9): 3170 - 3173. [Abstract] [Full Text] |
||||
![]() |
T. A. M. Abdu, T. A. Elhadd, R. Neary, and R. N. Clayton Response to Dr. Oelkers: There Is Enough Evidence in Favor of Low Dose ACTH Test--Authors' Response J. Clin. Endocrinol. Metab., August 1, 1999; 84(8): 2973a - 2974. [Full Text] |
||||
![]() |
D. H. P. Streeten Editorial: Shortcomings in the Low-Dose (1 {micro}g) ACTH Test for the Diagnosis of ACTH Deficiency States J. Clin. Endocrinol. Metab., March 1, 1999; 84(3): 835 - 837. [Full Text] |
||||
![]() |
T. A. M. Abdu, T. A. Elhadd, R. Neary, and R. N. Clayton Comparison of the Low Dose Short Synacthen Test (1 {micro}g), the Conventional Dose Short Synacthen Test (250 {micro}g), and the Insulin Tolerance Test for Assessment of the Hypothalamo-Pituitary-Adrenal Axis in Patients with Pituitary Disease J. Clin. Endocrinol. Metab., March 1, 1999; 84(3): 838 - 843. [Abstract] [Full Text] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |