Adrenal Insufficiency

Home , ACTH stimulation test, Adrenal crisis, Adrenal insufficiency, Methylprednisolone

Downloaded from www.jama.com at Johns Hopkins University, on November 23, 2005

GRAND ROUNDS CLINICIAN’S CORNER AT THE JOHNS HOPKINS BAYVIEW MEDICAL CENTER

Roberto Salvatori, MD A 44-year-old woman reported several weeks of fatigue, somnolence, pain CASE PRESENTATION in the large joints, nausea, and decreased appetite. She had also noted an A 44-year-old woman whose medical unintentional 11-kg weight loss over a period of 6 months. She had a re- history was remarkable only for chon- mote history of amenorrhea, but she was presently menstruating regularly. drocalcinosis of the knees was referred She was taking no medications, with the exception of acetaminophen as needed to the endocrinology department to for knee pain. The diagnosis of adrenal insufficiency (AI) was considered. exclude a metabolic or hormonal prob- Serum cortisol level after adrenocorticotropin hormone (ACTH) stimulation lem. The patient complained of several was abnormal. Because her plasma ACTH level was not increased, a diag- weeks of fatigue, somnolence, pain in the large joints, nausea, and decreased nosis of secondary AI (due to deficiency in ACTH) was made. Magnetic reso- appetite. Her symptoms had appeared nance imaging of the brain performed to exclude the presence of a sellar or gradually, without a clear precipitating suprasellar mass showed reduction in size of the pituitary gland and an in- event. She had also noted an uninten- creased cerebrospinal fluid content within the sella, consistent with a par- tional 11-kg weight loss over a period tially empty sella. The patient’s symptoms improved rapidly with hydrocor- of 6 months, associated with reduced tisone therapy but during follow-up, the dose of hydrocortisone was found appetite and occasional nausea. Her to be excessive. Important differences exist between primary and secondary past medical history was remarkable AI, and the diagnosis of secondary AI may be challenging. The therapy of AI for the fact that in her early 30s she had had a period of amenorrhea lasting should be carefully tailored to the requirements of the individual patient. approximately 1 year. After blood work, JAMA. 2005;294:2481-2488 www.jama.com a physician had told her that “my pituitary was not working and that I would not be able to become pregnant.” Sub- with a body mass index (BMI, calcu- Upon examination of her labora- sequently, however, her periods returned lated as weight in kilograms divided by tory studies, the electrolytes were nor- spontaneously, and she had 4 natural the square of height in meters) of 25.2. mal (sodium 141 mEq/L, potassium 3.6 pregnancies. Her recumbent blood pressure was mEq/L), as were her creatinine and On the review of systems, the pa- 115/80 mm Hg, her heart rate 74/min, blood urea nitrogen levels. Blood he- tient denied any change in her skin her respiratory rate 12/min, and her moglobin was normal (13 g/dL). At color or texture, in bowel habits, diz- temperature 36.5°C. The patient had no 10 AM, serum cortisol and plasma ad- ziness, chest pain, shortness of breath, significant blood pressure or heart rate renocorticotropin hormone (ACTH) polyuria, polydypsia, headaches, or vi- change after 1 minute in the upright po- levels were 4.0 µg/dL (normal range, sual changes. Her periods were com- sition. The skin and buccal mucosa had 5-20 µg/dL) and 9 pg/mL (normal ing monthly, with normal flow. She was no hyperpigmentation or hypopigmen- range, 6-60 pg/mL), respectively. The taking no medications, with the excep- tation. The thyroid gland was pal- serum level of free thyroxine (T4) was tion of acetaminophen as needed for pable but without abnormalities. There 1.2 ng/dL (15.4 pmol/L) (normal range, knee pain. Physical examination was no cervical or axillary adenopa- thy. Examination of the heart showed Author Affiliation: Department of Medicine, Divi- showed a well-appearing woman who sion of Endocrinology, Johns Hopkins University School weighed 69.5 kg and stood 166 cm tall, a regular rate and rhythm with nor- of Medicine, Baltimore, Md. mal cardiac sounds and no murmurs. Corresponding Author: Roberto Salvatori, MD, De- partment of Medicine, Division of Endocrinology, Johns See also Patient Page. The lungs were clear to auscultation and Hopkins University School of Medicine, 1830 E Monu- the abdomen was benign. Visual fields ment St, Suite 333, Baltimore, MD 21287 (salvator CME available online at were full by confrontation, and the neu- @jhmi.edu). www.jama.com Grand Rounds Section Editor: David S. Cooper, MD, rological examination was nonfocal. Contributing Editor, JAMA.

0.18-1.6 ng/dL [2.32-20.59 pmol/L]), adenoma that infarcted, leaving ids in the adrenal glands and in the cen- with a thyroid-stimulating hormone of an “empty sella.”1 Although ACTH tral nervous system. Therefore, the 3.3 mIU/mL (normal range, 0.5-4.5 deficiency with normal thyroid- clinician should perform a careful neu- mIU/mL). An ACTH stimulation test stimulating hormone and gonadotro- rological examination in young males was performed. Sixty minutes after the pin secretion (judged by the resump- with primary AI, keeping in mind that intravenous injection of 1 to 24 syn- tion and continuation of normal AI can appear long before neurologi- thetic ACTH (250-µg intravenous), the menses) is rare, it has been described cal symptoms.6 Several infectious pro- serum cortisol was 7 µg/dL (normal, before in a similar setting.2 The low cesses known to be associated with Ͼ18 µg/dL). serum insulin-like growth factor 1 AIDS, particularly Cytomegalovirus, My- Magnetic resonance imaging of the in a patient with hypopituitarism cobacterium tuberculosis, Cryptococcus brain, performed to exclude the pres- strongly suggests growth hormone neoformans, Toxoplasma gondii, Myco- ence of a sellar or suprasellar mass, deficiency.3 bacterium avium intracellulare, Pneu- showed reduction in size of the pitu- The etiology, diagnosis, and therapy mocystis jiroveci, and Histoplasma cap- itary gland, mild leftward deviation of of primary and secondary AI, as well as sulatum, may lead to adrenal gland the pituitary stalk, and an increased ce- the differences between these 2 dis- destruction. Among medications, ke- rebrospinal fluid content within the ease states, are reviewed herein. The toconazole (an antifungal) and etomi- sella, consistent with a partially empty major hormones of the hypothalamic- date (a general anesthetic) can cause sella, but no tumor mass or cyst. pituitary-adrenal (HPA) axis are shown AI.7 The patient started taking hydrocor- in the FIGURE. Secondary AI. The most common tisone (20 mg in the morning and cause of secondary AI is the abrupt dis- 10 mg in the afternoon). Two months Classification of AI continuation of long-term administra- later, she reported significant improve- Adrenal insufficiency develops when a tion of glucocorticoids. There is a great ment of all her symptoms and a large part of the function of the adre- individual variability in susceptibility 6-kg weight gain. Six months later she nal glands is lost. Primary AI is caused to suppression of the HPA axis by ex- was feeling well, but her weight had by processes that damage the adrenal ogenous glucocorticoids; therefore, the increased to 94 kg (24.5 kg in 9 glands or by drugs that block cortisol presence of AI cannot be predicted re- months of therapy; BMI, 34.1), with synthesis. In contrast, secondary AI re- liably by the dose and duration of glu- preferential accumulation of adipose sults from processes that reduce the se- cocorticoid use.8 Several weeks of ex- tissue in her abdomen. Her hydrocor- cretion of ACTH by the pituitary gland ogenous glucocorticoid administration tisone dose was reduced to 10 mg in due to a pituitary or hypothalamic pa- are required for the development of AI.9 morning and 5 mg in afternoon, with- thology; although the latter, due to defi- Intramuscular, intra-articular, and even out any relapse of her symptoms. cit of corticotrophin-releasing hor- inhaled or topical glucocorticoids can Three and half years later, she was still mone (CRH), is sometimes called cause significant suppression of the HPA doing well with the same hydrocorti- tertiary AI. In this review, both forms axis.10,11 Similarly, megestrol acetate, a sone dose and her weight was 87 kg are included in the secondary AI group. synthetic progestational agent used to (BMI, 31.5). She was still having regu- Both primary and secondary AI can de- increase appetite in a variety of ca- lar periods. The insulin-like growth velop either slowly, over several weeks chexia-inducing illnesses, can sup- factor 1 was low at 71 ng/mL (normal or months, or acutely, with cata- press ACTH secretion at doses of more range, 90-360 ng/mL). strophic consequences that may lead to than 160 mg/d, leading to secondary cardiovascular collapse and death. The AI.12 DISCUSSION classification and common etiologies of Secondary AI can be caused by pi- This patient has symptoms and bio- primary and secondary AI are shown tuitary adenomas, craniopharyngio- chemical evidence of adrenal insuffi- in the Figure. mas, pituitary surgery,lymphocytic hy- ciency (AI), as shown by her abnormal Primary AI. In developed coun- pophysitis, and a wide variety of other cortisol response to ACTH (the inter- tries, the most frequent cause of pri- neoplastic, inflammatory, and infec- pretation criteria of the ACTH stimu- mary AI is autoimmune adrenalitis. In tious processes involving the sellar or lation test will be discussed below). the developing world, on the other suprasellar area. Pituitary or whole Her AI is secondary to reduced ACTH hand, tuberculosis most likely re- brain irradiation can cause AI up to sev- secretion rather than due to a primary mains the more common cause of ad- eral years after its completion.13 Fi- disease of the adrenal glands, as indi- renal failure.4,5 In young males, adre- nally, traumatic brain injury and sub- cated by the inappropriately normal noleukodystrophy (or the less severe arachnoid hemorrhage can cause AI that plasma ACTH level. The past history adrenomyeloneuropathy) must be con- may not manifest itself until several of amenorrhea is consistent with the sidered. This disease, transmitted as a months after the acute event.14 possibility that the patient has a pitu- recessive X-linked trait, is due to the ac- In the failing pituitary gland, ACTH itary pathology, possibly a pituitary cumulation of very long chain fatty ac- secretion is usually the last function to

ADRENAL INSUFFICIENCY be lost. As the present case illustrates, Figure. Hypothalamic-Pituitary-Adrenal Axis and Causes of Primary and Secondary Adrenal however, exceptions to this rule oc- Insufficiency cur. The patient had ACTH deficiency (and possibly growth hormone deficiency) despite the presence of nor- Hypothalamic-Pituitary- Primary Adrenal Insufficiency mal thyroid and gonadotropin func- Adrenal Axis tion. Such exceptions are more frequent Stress Stress in postirradiation hypopituitarism.13 Activation Causes Acute Onset Clinical Presentation Hypothalamus Hypothalamus Adrenal Hemorrhage The development of acute AI is poten- or Infarction tially lethal. The physician must con- Corticotropin-Releasing Slow Onset Increased CRH∗ sider AI as a possible cause of unex- Hormone (CRH) Autoimmune Disease plained deterioration of cardiovascular Adrenalitis Activation Infectious Diseases status. Adrenal hemorrhage or infarc- Tuberculosis tion occurs in patients who are se- AIDS-Related Infections verely ill from underlying conditions, Anterior Pituitary Anterior Pituitary Cancer Lymphoma including sepsis, pulmonary embo- Metastases Adrenocorticotropin lism, acute renal failure, acute myocar- Increased ACTH Drugs dial infarction, and heart failure.15,16 The Hormone (ACTH) Ketoconazole Etomidate Activation presence of antiphospholipid antibod- Inhibition Other ies or pharmacological anticoagula- Congenital Adrenal Hyperplasia tion can lead to adrenal hemorrhage or Adrenal Gland Adrenal Gland Adrenoleukodystrophy 17 infarction. (Males) Patients with slow onset AI usually complain of being chronically fa- Cortisol Decreased Cortisol tigued. They often report joint pain, lack of appetite, unintentional weight loss, Secondary Adrenal Insufficiency abdominal pain, nausea, and diarrhea. Primary AI, in contrast with second- Hypothalamic Etiology Pituitary Etiology ary AI, is often associated with lack Stress Stress Or Causes of aldosterone as well as cortisol. Con- Acute Onset sequently, symptoms and signs of Pituitary Apoplexy mineralocorticoid deficiency (salt- Hypothalamus Hypothalamus Pituitary or Hypothalamic Surgery craving, postural hypotension, electro- Traumatic Brain Injury lyte abnormalities) usually indicate pri- Decreased CRH Increased CRH∗ Slow Onset mary AI. For any given level of morning Autoimmune Disease serum cortisol, patients with primary Lymphocytic Hypophysitis Infectious Disease AI have more severe symptoms than Tuberculosis 18 patients with secondary AI. Patients Pituitary Gland Pituitary Gland Cancer with secondary AI are often able to Pituitary or Hypothalamic Tumors Lymphoma function relatively well during un- Decreased or Decreased or Trauma or Other Injury stressed periods, and they may mani- Inappropriately Inappropriately Traumatic Brain Injury fest cardiovascular instability or hypo- Normal ACTH Normal ACTH Subarachnoid Hemorrhage Radiation glycemia only when they undergo Drugs physical stress. Adrenal Gland Adrenal Gland Megestrol Acetate Hyperpigmentation of the skin and Other Discontinuation of of the mucosae resulting from the mel- Exogenous Glucocorticoids anocyte-stimulating activity of ␤ li- Decreased Cortisol Decreased Cortisol Sarcoidosis potropin, which derives from the same Empty Sella Syndrome precursor as ACTH, is observed only in primary AI. Although hyperkalemia is Affected Structure Causing Adrenal Insufficiency observed only in primary AI, hypona- tremia can also occur in secondary AI, as the result of reduced glomerular fil- *Although CRH in the hypophyseal portal system cannot be measured, it is likely increased.

ADRENAL INSUFFICIENCY tration rate, increased antidiuretic hor- ing this test, as shift workers or patients tisol as an index of AI.20 Unfortunately, mone secretion, and possible concomi- with unusual sleep-wake patterns may direct measurement of free serum cor- tant central hypothyroidism. Patients have altered cortisol circadian rhythm.19 tisol is not yet widely available, and there with primary AI caused by autoim- This test is helpful if the serum corti- is no formula to correct serum cortisol mune adrenalitis are at risk for other sol level is more than 18 µg/dL (nor- according to albumin or total protein manifestations of autoimmune dis- mal adrenal function) or less than 3 level. For this reason, patients diag- ease, such as vitiligo, Hashimoto thy- µg/dL (indicates AI).4 In the latter case, nosed with AI in the intensive care unit roiditis, pernicious anemia, and type 1 plasma ACTH level will distinguish be- based on serum cortisol should be re- diabetes mellitus. TABLE 1 illustrates tween primary and secondary AI. In pri- tested in the outpatient setting before some of the contrasts between primary AI, the ACTH level is almost in- being committed to a life-long gluco- mary and secondary AI. variably more than 100 pg/mL, while corticoid treatment. in secondary AI, plasma ACTH If an ambulatory patient has a morn- Diagnosis of AI can be either low or inappropriately nor- ing serum cortisol level of between 3.1 Once AI is suspected, a variety of tests mal (when serum cortisol is reduced). and 17.9 µg/dL, some sort of dynamic may be used to evaluate adrenal func- In patients with severe physical stress, test of adrenal function is essential. The tion. A stepwise approach helps main- serum cortisol level should be more than test options for the assessment of ad- tain cost-effectiveness in screening all 18 µg/dL at any time of the day.4 One ex- renal function are summarized in the patients in whom this potentially ception is the patient with severe hypo- TABLE 2. None of the available stimu- dangerous (and treatable) disease is proteinemia, because albumin and cor- lation tests is ideal in terms of sensi- suspected. tisol–binding globulin normally bind tivity and specificity.Consequently,the Nonstimulated Serum Cortisol. The approximately 90% of circulating se- test results must be interpreted in the initial test in an ambulatory patient rum cortisol. In patients who are se- context of the clinical scenario. The should be a 6- to 8-AM measurement of verely ill with serum albumin values of choice of the test depends on several serum cortisol and plasma ACTH. This less than 2.5 g/L, the total cortisol is of- factors, including the experience of the approach exploits the fact that serum ten markedly lower than 18 µg/dL; how- physician and practical consider- cortisol level peaks in the early morn- ever, the free fraction of cortisol may be ations related to test performance. ing hours. A careful social and work his- normal. In such patients, the total se- Insulin-Induced Hypoglycemia (In- tory should be obtained before order- rum cortisol is less reliable than free cor- sulin Tolerance Test). The insulin tolerance test (ITT) measures the patient’s cortisol response to hypoglycemia Table 1. Signs and Symptoms of Primary and Secondary Adrenal Insufficiency (AI) induced by the intravenous adminis- Primary AI Secondary AI tration of insulin. This test is often con- Skin and mucosae Dark Pale sidered the criterion standard because Potassium High Normal it assesses the ability of the entire HPA Sodium Low Normal or low axis to respond to the stressful situa- Associated diseases Primary hypothyroidism, type 1 Central hypogonadism or tion of hypoglycemia.21 Following in- diabetes mellitus, vitiligo, hypothyroidism, growth neurological deficit hormone deficiency, diabetes sulin (0.1 IU/kg) administration, blood (adrenoleukodystrophy, insipidus, headaches, visual is drawn during symptomatic hypogly- males only) abnormalities cemia (glucose should decrease to less

Table 2. Synopsis of the Dynamic Tests Available to Assess Adrenal Function Test

Conventional-Dose ITT Metyrapone CRH ACTH Low-Dose ACTH Stimulus Insulin, 0.1-0.15 IU/kg Metyrapone, 30 mg/kg by Ovine CRH, 1 µg/kg 1-24 ACTH, 250 µg 1-24 ACTH, 1 µg intravenously mouth at midnight intravenously intravenously (or intravenously (maximum, 3000 mg) intramuscularly) Blood drawing time 30, 45, 60, and 90 min 8 AM following morning 15, 30, and 60 min 30 or 60 min 30 min points Measurements Serum cortisol, blood Serum cortisol and Serum cortisol Serum cortisol Serum cortisol glucose 11-deoxycortisol Cutoff Serum cortisol Ͼ18 µg/dL, Serum 11-deoxycortisol Serum cortisol Ͼ18.5 Serum cortisol Ͼ18 Serum cortisol Ͼ18 if glucose decreased Ͼ7 µg/dL, with serum µg/dL µg/dL µg/dL Ͻ40 mg/dL cortisol Ͻ5 µg/dL Abbreviations: ACTH, adrenocorticotropin hormone; CRH, corticotrophin-releasing hormone; ITT, insulin tolerance test. SI conversion: To convert glucose to mmol/L, multiply by 0.0555.

ADRENAL INSUFFICIENCY than 40 mg/dL [Ͻ2.22 mmol/L]). In CRH Stimulation Test. The CRH be diluted to reach lower concentra- obese patients with insulin resistance, stimulation test measures the ability of tions (ie, 1 µg/mL), and it can be kept the usual dose of insulin should be in- the pituitary gland to secrete ACTH in at 4°C for up to 4 months without any creased to 0.15 IU/kg.22 A serum cor- response to CRH and the ability of the decline in biological activity.31 tisol peak of more than 18 µg/dL is con- adrenal gland to respond with an in- In several studies, the low-dose sidered normal. The ITT has the crease in cortisol to the increase in cir- ACTH stimulation test has been shown advantage that it can also test the culating ACTH. Ovine CRH (1 µg/kg) to be more sensitive for diagnosing mild growth hormone reserve in patients is injected intravenously, and cortisol secondary AI than the traditional with pituitary or hypothalamic dis- is measured after 15, 30, and 60 min- 250-µg test, but not as sensitive as the ease. It is contraindicated in patients utes. The test has been proposed as a ITT or metyrapone tests.24,32-35 How- older than 60 years, and in those with way to differentiate secondary (pitu- ever, a recent meta-analysis has found history of seizures, or with docu- itary disease) from tertiary (hypotha- the sensitivity and specificity of the 2 mented or suspected coronary artery lamic disease) AI. However, a serum tests to be not significantly different in disease. It requires close medical su- cortisol cutoff that allows high sensi- the diagnosis of secondary AI.28 Both pervision and trained personnel. Even tivity (18.5 µg/dL) is faulted by very low doses perform equally well for the di- the ITT has some limitations in terms specificity (33%).27 The lack of mul- agnosis of primary AI, even in mild of reproducibility and clear cutoff lev- tiple studies involving large numbers of cases.28,36 Using the low-dose test, most els.21,23 It has been proposed that to im- patients and the high cost of CRH have authors recommend using a 30- prove its specificity glucose should de- greatly limited its use. minute serum cortisol cutoff of 18 µg/ crease to less than 30 mg/dL (Ͻ1.67 ACTH Stimulation Test. The ACTH dL. When the results of this test are bor- mmol/L), with accompanying poten- stimulation test is based on the inabil- derline (peak serum cortisol between tial higher risks.23 ity of a diseased adrenal gland to re- 15 and 18 µg/dL), confirmation by Metyrapone Test. The metyrapone spond acutely to the injection of ACTH either the metyrapone test or ITT test measures the ability of the HPA axis by secreting cortisol. In the conven- should be performed.24,28 to respond to an acute reduction in se- tional ACTH stimulation test, 250 µg There are sporadic reports of pos- rum cortisol levels. Metyrapone inhib- of synthetic (1-24) ACTH is injected in- sible dangers of missing the diagnosis its 11-hydroxylase, the enzyme intravenously (or intramuscularly), and of secondary AI using the 250-µg test.37 volved in the last step of cortisol serum cortisol levels are measured at 30 However, it has not yet been proven in synthesis. This inhibition causes a de- or 60 minutes. A serum cortisol level a large cohort whether patients who crease in cortisol that results (in a of more than 18 µg/dL at either time pass the high-dose test but fail the low- healthy patient) in a compensatory in- point constitutes a normal response, in- dose test actually benefit from gluco- crease in ACTH and in the cortisol dependent of baseline level and time of corticoid therapy. precursor, 11-deoxycortisol. This test day. Although some healthy individu- As long-term reduction in ACTH se- is simple, relatively inexpensive, als can peak below this cutoff level, a cretion is needed to develop adrenal at- very sensitive, and requires a single peak of less than 15 µg/dL is invari- rophy, any form of ACTH stimulation blood drawing.24 The administration of ably abnormal.28 test generates false-negative results in metyrapone (30 mg/kg; maximal dose, The choice of the 250-µg dose is patients with a recent onset of ACTH 3000 mg) occurs at midnight, and blood based solely on the fact that ACTH deficiency, such as those who have re- is drawn the following morning at 8 AM comes in 250-µg vials. However, with cently undergone pituitary surgery or for cortisol and 11-deoxycortisol.25 In this dose, ACTH reaches plasma lev- had a pituitary infarction. The reduc- response to metyrapone, serum corti- els that are approximately 1000 times tion in ACTH secretion must be chronic sol level should decrease to less than 5 the values observed in maximally (Ͼ1 month) and probably severe to µg/dL, and 11-deoxycortisol should in- stressed healthy individuals, thereby po- cause atrophy of the adrenal glands. crease to more than 7 µg/dL. Recently, tentially causing a falsely normal cor- Because the low-dose test is simple it has been proposed that the sum of tisol response by an adrenal gland that to perform, potentially more sensitive cortisol and 11-deoxycortisol should be is in fact partially impaired.29 Based on than the high-dose test, and less ex- more than 16.5 µg/dL.25 Although some this observation, Dickstein et al30 pro- pensive (many patients can be tested authors recommend that this test posed the use of a more physiological with a single 250-µg ACTH vial), it should be performed only during a hos- low ACTH dose (1 µg) to identify should be the routine procedure used pital admission, a large retrospective se- milder forms of secondary AI. They in the evaluation of potential cases of ries found it to be safe and suitable for showed that in healthy individuals, the both primary and secondary AI. the outpatient setting.26 Regrettably,the 1-µg dose causes a similar increase in intermittent availability of metyra- serum cortisol at 30 minutes, with a de- Identifying the Cause of AI pone limits its applicability in the cline at 60 minutes.30 Because ACTH is Once the diagnosis of AI is established, United States. packaged only in 250-µg vials, it must the physician must determine its cause.

ADRENAL INSUFFICIENCY

corticoids, respectively.45,46 In patients Table 3. Equivalency Table of Commonly Used Glucocorticoids* who are taking protease inhibitors, sig- Glucocorticoids nificant adrenal suppression may oc- Cortisone cur within few months of therapy with Hydrocortisone Acetate Prednisone Prednisolone Dexamethasone inhaled steroids.47 Relative potency 1 0.8 4 5 25-50 Women usually do not need adjust- Equivalent dose, mg 20 25 5 4 0.5 ment in their dose of glucocorticoids Biological half-life, h 8-12 8-12 18-36 18-36 36-54 during pregnancy, but in case of *Adapted from Krasner.9 protracted first trimester vomiting, par- enteral administration may be needed. In primary AI, the age of the patient, as- when, together with replacement Every patient with AI should wear a sociated morbidities, clinical picture, and therapy,one wants to suppress ACTH se- medical alert bracelet or necklace stat- medical history must guide the deci- cretion, such as in congenital adrenal ing the need for glucocorticoids in sions regarding work-up. Measure- hyperplasia. case of an emergency. An emergency ment of anti-adrenal antibodies may help Because there is no good biochemi- glucocorticoids injection kit should be the diagnosis of autoimmune adrenal- cal index to help determine the right prescribed to patients who live or itis. This test is highly specific but not glucocorticoid dose to reduce long- travel to remote areas, where they may 100% sensitive.38 Measurement of very term adverse effects on bone and other not have readily available medical long chain fatty acids must be obtained tissues, the smallest amount that im- assistance. if adrenoleukodystrophy is suspected. proves the patient’s symptoms is rec- Mineralocorticoids. Patients with Computed tomography imaging of the ommended.42 In patients with gluco- primary AI also need mineralocorti- adrenals may unveil hemorrhagic, meta- corticoid-induced AI, using a small dose coid therapy. The only drug available static, or infectious diseases. Adrenal bi- may accelerate the recovery of the HPA is fludrocortisone. Its dose ranges in opsy may occasionally be needed. In sec- axis.9 In some patients who have mild most cases between 0.05 and 0.2 mg/d, ondary AI, in the absence of a history of secondary AI and no symptoms dur- usually in a single dose. The dose is ad- exogenous glucocorticoid exposure, ing everyday life, hydrocortisone may justed according to the patient’s symp- magnetic resonance imaging of the sel- be required only during periods of toms, including orthostatic dizziness lar region is mandatory, as secondary AI physical stress. In contrast with the ap- and salt craving, and to serum potas- may be the presenting feature of a neo- proach to treating primary hypothy- sium and plasma renin activity. Sup- plastic process in the pituitary or the roidism, when restoration of a normal pressed plasma renin activity is a use- hypothalamus. thyroid-stimulating hormone is an im- ful marker of fludrocortisone overdose. portant target of therapy, in the treat- In the absence of orthostatic symp- Therapy ment of primary AI, attempting to nor- toms or of significant changes in the Glucocorticoids. Glucocorticoids are malize plasma ACTH levels is not blood pressure or heart rate during or- the main therapy for all forms of AI. Al- recommended because this invariably thostatic maneuvers, however, the though several kinds of glucocorti- results in overdosing and the induc- fludrocortisone dose should not be in- coids can be used, hydrocortisone (10- tion of an iatrogenic Cushing syn- creased even if the plasma renin activ- 12.5 mg/m2 per day) is preferred because drome state.43 In patients who are also ity is increased to mild or moderate lev- its short half-life mimics most closely the hypothyroid, thyroid hormones should els. Overdosing of fludrocortisone can normal cortisol circadian rhythm. This never be replaced before administer- cause fluid retention, edema, and hy- dose of hydrocortisone is not associ- ing glucocorticoids; euthyroidism may pertension. ated with reduced bone mineral den- trigger an adrenal crisis by accelerat- Androgen. In women with AI, dehy- sity.39 The downside is that hydrocorti- ing the metabolism of cortisol. Simi- droepiandrosterone replacement im- sone must be given twice or 3 times a larly, growth hormone therapy can ac- proves well-being and sexuality.47 As day.40 The classic dose of 30 mg/d (20 celerate cortisol metabolism, thereby dehydroepiandrosterone is considered a mg in the morning and 10 mg in the af- requiring adjustment of hydrocorti- dietary supplement in the United States, ternoon) is probably excessive in most sone dosing.44 it is not regulated by the US Food and patients, particularly patients with sec- The cytochrome P450 CYP3A4 iso- Drug Administration. Arlt et al48 showed ondary AI, as shown by today’s patient, zyme is involved in the hepatic me- that patients treated with 50 mg/d of de- who gained a significant amount of tabolism of glucocorticoids. There- hydroepiandrosterone increase their se- weight with this dose. Other glucocor- fore, patients who are taking drugs that rum dehydroepiandrosterone-sulfate and ticoids may be administered, keeping in may increase (phenytoin, rifampin, bar- testosterone levels; therefore, these mea- mind their potencies relative to that of biturates) or decrease (protease inhibi- surements can be used to determine ad- hydrocortisone (TABLE 3).9,41 Long- tors) the function of this enzyme may equateness of the supplement and the acting glucocorticoids are preferred need higher or lower doses of gluco- patient’s adherence.

ADRENAL INSUFFICIENCY

Therapy During Stress. During ill- CONCLUSION pituitary-adrenal axis and adrenocortical function. J Endocrinol Invest. 2004;27:380-386. nesses, the glucocorticoid dose should The symptoms of AI are nonspecific and 6. Moser H, Dubey P, Fatemi A. Progress in X-linked be increased. One reasonable ap- this disease often goes undiagnosed for adrenoleukodystrophy. Curr Opin Neurol. 2004; 17:263-269. proach is to double the maintenance extended periods of time, with the risk 7. Magill SS, Puthanakit T, Swoboda SM, et al. Im- dose in the setting of fevers (tempera- of potential serious consequences. There- pact of fluconazole prophylaxis on cortisol levels in criti- tures Ͼ38°C), major dental proce- fore, physicians must suspect it in the cally ill surgical patients. Antimicrob Agents Chemother. 2004;48:2471-2476. dures (tooth extractions or root ca- presence of the symptoms described ear- 8. Schlaghecke R, Kornely E, Santen RT, Ridder- nals), or invasive diagnostic procedures lier, particularly in patients with a his- skamp P. The effect of long-term glucocorticoid therapy on pituitary-adrenal responses to exogenous cortico- (gastroscopy, colonoscopy, cystos- tory of autoimmune diseases or sign or trophin-releasing hormone. NEnglJMed. 1992;326: copy, or bronchoscopy). Patients with symptoms consistent with hypopituita- 226-230. AI are more likely to develop cardio- rism. The work-up must be step-wise and 9. Krasner AS. Glucocorticoid-induced adrenal insufficiency. JAMA. 1999;282:671-676. vascular instability when they have a should start from the least expensive test 10. Mahachoklertwattana P, Sudkronrayudh K, Direk- disease that causes vomiting or diar- (morning serum cortisol). Adrenal insuf- wattanachai C, Choubtum L, Okascharoen C. De- creased cortisol response to insulin induced hypogly- rhea. In these situations, in addition to ficiency needs to be suspected in severely caemia in asthmatics treated with inhaled fluticasone the increased glucocorticoid require- ill patients with cardiovascular instabil- propionate. Arch Dis Child. 2004;89:1055-1058. 11. Mader R, Lavi I, Luboshitzky R. Evaluation of the ment, they may not properly absorb the ity. In this setting, glucocorticoids should pituitary-adrenal axis function following single intra- oral therapy. For this reason, patients be administered empirically while wait- articular injection of methylprednisolone. Arthritis with AI should go the emergency de- ing for the results of a pretreatment Rheum. 2005;52:924-928. 12. Ron IG, Soyfer V, Goldray D, Inbar MJ, Weis- partment if they experience multiple serum cortisol. man Y. A low-dose adrenocorticotropin test reveals episodes of vomiting and/or diarrhea. Confirming the diagnosis of partial impaired adrenal function in cancer patients receiving megestrol acetate therapy. Eur J Cancer. 2002;38: The glucocorticoid doses commonly secondary AI may be challenging and 1490-1494. used during major stress (major sur- repeated testing may be needed before 13. Toogood AA. Endocrine consequences of brain irradiation. Growth Horm IGF Res. 2004;14 gery,severe infection, myocardial infarc- committing a patient to a lifelong glu- (suppl A):S118-S124. tion)—80 to 100 mg of hydrocortisone cocorticoid therapy. As both lack of 14. Aimaretti G, Ambrosio MR, Di Somma C, et al. every 8 hours—are probably excessive needed treatment and not warranted Traumatic brain injury and subarachnoid haemor- rhage are conditions at high risk for hypopituitarism: and not based on any clear evidence. Pa- glucocorticoids are dangerous, in du- screening study at 3 months after the brain injury. Clin tients with intact adrenal function se- bious cases the decision of whether to Endocrinol (Oxf ). 2004;61:320-326. 15. Xarli VP, Steele AA, Davis PJ, Buescher ES, Rios crete between 75 and 150 mg/d in re- treat a patient with suspected AI should CN, Garcia-Bunuel R. Adrenal hemorrhage in the adult. sponse to major surgery.49 Therefore, the be left to practitioners who have expe- Medicine. 1978;57:211-220. maximal dose recommended is 50 mg of rience in the field. Glucocorticoids need 16. Siu SC, Kitzman DW, Sheedy PF 2nd, Northcutt RC. Adrenal insufficiency from bilateral adrenal hydrocortisone every 8 hours. When to be prescribed judiciously, using the hemorrhage. Mayo Clin Proc. 1990;65:664-670. such a dose is administered, there is no lowest possible dose that improves the 17. Espinosa G, Santos E, Cervera R, et al. Adrenal involvement in the antiphospholipid syndrome: clini- need to prescribe fludrocortisone even patient’s symptoms to avoid long- cal and immunologic characteristics of 86 patients. in patients with primary AI, because the term adverse effects on bone, blood Medicine (Baltimore). 2003;82:106-118. 18. Vallotton MB. Endocrine emergencies: disorders high doses of hydrocortisone will acti- pressure, body composition, and glu- of the adrenal cortex. Baillieres Clin Endocrinol Metab. vate the mineralocorticoid receptor. It is cose metabolism. 1992;6:41-56. possible that even this dosage is exces- 19. Touitou Y, Motohashi Y, Reinberg A, et al. Effect Financial Disclosures: None reported. of shift work on the night-time secretory patterns of sive in patients with secondary AI. A Acknowledgment: I thank John Stone, MD, and melatonin, prolactin, cortisol and testosterone. small study50 showed that among pa- Simeon Margolis, MD, PhD, from the Department of Eur J Appl Physiol Occup Physiol. 1990;60:288-292. Medicine, Johns Hopkins University, for critically re- 20. Hamrahian AH, Oseni TS, Arafah BM. Measure- tients with glucocorticoid-induced AI viewing the manuscript. No compensation was given ments of serum free cortisol in critically ill patients. (taking a mean daily prednisone dose of to either of them. N Engl J Med. 2004;350:1629-1638. 18.8 mg) undergoing major surgery, a 21. Pfeifer M, Kanc K, Verhovec R, Koijianic A. Re- REFERENCES producibility of the ITT for assessment of GH and cor- group treated with the usual predni- tisol responses in normal and hypopituitaric adult men. sone dose and a group that received 1. Robinson DB, Michaels RD. Empty sella resulting from Clin Endocrinol (Oxf ). 2001;54:17-22. the spontaneous resolution of a pituitary macroadenoma. 22. Cordido F, Alvarez-Castro P, Isidro ML, Casanueva stress-dose glucocorticoids had no dif- Arch Intern Med. 1992;152: 1920-1923. FF, Dieguez C. Comparison between insulin toler- ference in perioperative complications. 2. Hatazoe T, Murakami Y, Umaki I, Sohmiya M, Hu ance test, growth hormone (GH)-releasing hormone HY, Kato Y. Isolated adrenocorticotropic hormone de- (GHRH), GHRH plus acipimox and GHRH plus GH- Patients with septic shock whose total ficiency associated with growth hormone deficiency releasing peptide-6 for the diagnosis of adult GH de- serum cortisol increases by less than 9 and empty sella. Intern Med. 1995;34:688-691. ficiency in normal subjects, obese and hypopituitary µg/dL in response to 250 µg of ACTH 3. Hartman ML, Crowe BJ, Biller BM, Ho KK, Clem- patients. Eur J Endocrinol. 2003;149:117-122. mons DR, Chipman JJ; HyposCCS Advisory Board and 23. Nye EJ, Grice JE, Hokgings GI, et al. The insulin have lower mortality when treated with US HypoCCS Study Group. Which patients do not re- hypoglycemia test: hypoglycemic criteria and hydrocortisone (50 mg every 6 hours) quire a GH stimulation test for the diagnosis of adult reproducibility. J Neuroendocrinol. 2001;13: 524- 51 GH deficiency? J Clin Endocrinol Metab. 2002;87: 530. plus fludrocortisone (0.05 mg/d). This 477-485. 24. Soule S, van Zyl Smit K, Parolis G, et al. The low benefit is likely a pharmacological effect 4. Oelkers W. Adrenal insufficiency. N Engl J Med. dose ACTH stimulation test is less sensitive than the 1996;335:1206-1212. overnight metyrapone test for the diagnosis of sec- and does not necessarily classify these 5. Kelestimur F. The endocrinology of adrenal tuber- ondary hypoadrenalism. Clin Endocrinol (Oxf ). 2000; patients as having AI. culosis: the effects of tuberculosis on the hypothalamo- 53:221-227.

ADRENAL INSUFFICIENCY

25. Berneis K, Staub JJ, Gessler A, Meier C, Girard J, in the assessment of hypothalamo-pituitary-adrenal and does it matter? Clin Endocrinol (Oxf ). 1997;46: Muller B. Combined stimulation of adrenocorticotro- axis immediately after pituitary surgery. J Clin Endo- 255-261. pin and compound-S by single dose metyrapone test crinol Metab. 2000;85:3713-3719. 43. Oelkers W, Diederich S, Bahr V. Diagnosis and as an outpatient procedure to assess hypothalamic- 34. Courtney CH, McAllister AS, McCance DR, et al. therapy surveillance in Addison’s disease: rapid adre- pituitary-adrenal function. J Clin Endocrinol Metab. Comparison of one week 0900 serum cortisol, low and nocorticotropin (ACTH) test and measurement of 2002;87:5470-5475. standard dose synachten test witha4to6weeks in- plasma ACTH, renin activity and aldosterone. J Clin 26. Fiad TM, Kirby JM, Cunningham SK, Mckenna TJ. sulin hypoglycemia test after pituitary surgery in as- Endocrinol Metab. 1992;75:259-264. The overnight single-dose metyrapone test is a simple sessing HPA axis. Clin Endocrinol (Oxf ). 2000;53: 44. Swords FM, Carrol PV, Kisalu J, Wood PJ, Taylor and reliable index of HPA axis. Clin Endocrinol (Oxf ). 431-436. NF, Monson JP. The effects of growth hormone de- 1994;40:603-609. 35. Suliman AM, Smith TP, Labib M, Fiad TM, ficiency and replacement on glucocorticoids expo- 27. Schmidt IL, Lahner H, Mann K, Petersenn S. Di- McKenna TJ. The low-dose ACTH test does not pro- sure in hypopituitary patients on cortisone acetate and agnosis of adrenal insufficiency: evaluation of the cor- vide a useful assessment of the hypothalamic-pituitary- hydrocortisone therapy. Clin Endocrinol (Oxf ). 2003; ticotropin-releasing hormone test and basal serum cor- adrenal axis in secondary adrenal insufficiency. Clin 59:613-620. tisol in comparison to the insulin tolerance test in Endocrinol (Oxf ). 2002;56:533-539. 45. Lamberts SWJ, Bruining HA, De Jong FH. Corti- patients with hypothalamic-pituitary-adrenal disease. 36. Laureti S, Arvat E, Candeloro P, et al. Low dose costeroid therapy in severe illness. N Engl J Med. 1997; J Clin Endocrinol Metab. 2003;88:4193-4198. (1 µg) ACTH test in the evaluation of adrenal dys- 337:1285-1292. 28. Dorin RI, Qualis CR, Crapo LM. Diagnosis of function in pre-clinical Addison’s disease. Clin Endo- 46. Eagling VA, Back DJ, Barry MG. Differential in- adrenal insufficiency. Ann Intern Med. 2003;139: crinol (Oxf ). 2000;53:107-115. hibition of cytochrome P450 isoform by the protease 194-204. 37. Streeten DH, Anderson GH Jr, Bonaventura MM. inhibitors ritonavir, sequinavir and indinavir. BrJClin 29. Mayenknecht J, Diederich S, Bahr V, Plockinger The potential for serious consequences from misin- Pharmacol. 1997;44:190-194. U, Oelkers W. Comparison of low and high dose terpreting normal responses to the rapid adrenocor- 47. Samaras K, Pett S, Gowers A, McMurchie M, Coo- corticotrophin stimulation test in patients with pitu- ticotropin test. J Clin Endocrinol Metab. 1996;81:285- per DA. Iatrogenic Cushing’s syndrome with osteo- itary disease. J Clin Endocrinol Metab. 1998;83:1558- 290. porosis and secondary adrenal failure in human im- 1562. 38. Falorni A, Laureti S, Nikoshkov A, et al. 21- munodeficiency virus-infected patients receiving 30. Dickstein G, Shechner C, Nicholson WE, et al. Ad- hydroxylase autoantibodies in adult patients with en- inhaled corticosteroids and ritonavir-boosted prote- renocorticotropin stimulation test: effects of basal cor- docrine autoimmune diseases are highly specific for ase inhibitors: six cases. J Clin Endocrinol Metab. 2005; tisol level, time of the day and suggested new sensi- Addison’s disease: Belgian diabetes registry. Clin Exp 90:4394-4398. tive low dose test. J Clin Endocrinol Metab. 1991;72: Immunol. 1997;107:341-346. 48. Arlt W, Callies F, van Vlijmen JC, et al. Dehydro- 773-778. 39. Chikada N, Imaki T, Hotta M, Sato K, Takano K. epiandrosterone replacement in women with 31. Dickstein G, Arad E, Shechner C. Low-dose ACTH An assessment of bone mineral density in patients with adrenal insufficiency. N Engl J Med. 1999;341:1013- stimulation test. Endocrinologist. 1997;7:285-293. Addison’s disease and isolated ACTH deficiency treated 1020. 32. Abdu TAM, Elhadd TA, Neary R, Clayton RN. with glucococrticoid. Endocr J. 2004;51:355-360. 49. Salem M, Tainsh RE Jr, Bromberg J, Loriaux DL, Comparison of the low dose short synachten test 40. Mah PM, Jenkins RC, Rostami-Hodjegan A, et al. Chernow B. Perioperative glucocorticoid coverage: a (1 µg), the conventional dose short synachten test (250 Weight-related dosing, timing and monitoring hydro- reassessment 42 years after emergence of a problem. µg), and the insulin tolerance test for assessment of cortisone replacement therapy in patients with adre- Ann Surg. 1994;219:416-425. the hypothalamo-pituitary-adrenal axis in patients with nal insufficiency. Clin Endocrinol (Oxf ). 2004;61: 50. Glowniak JV, Loriaux DL. A doube-blind study of pituitary disease. J Clin Endocrinol Metab. 1999;84: 367-375. perioperative steroid requirements in secondary ad- 838-843. 41. Malchoff CD, Carey RM. Adrenal insufficiency. renal insufficiency. Surgery. 1997;121:123-129. 33. Dokmetas HS, Colak R, Kelestimur F, Selcuklu A, In: Current Therapy in Endocrinology and Metabo- 51. Annane D, Sebille V, Charpentier C, et al. Effect Unluhizarci K, Bayram F. A comparison between the lism. St Louis, Mo: Mosby; 1997:142-147. of treatment with low doses of hydrocortisone and 1-microg adrenocorticotropin (ACTH) test, the short 42. Peacey SR, Guo CY, Robinson AM, et al. Gluco- fludrocortisone on mortality in patients with septic ACTH (250 microg) test, and the insulin tolerance test corticoid replacement therapy: are patients over treated, shock. JAMA. 2002;288:862-871.