0021.972x/9179@&0923$03.00/0
`Journal
`of Clinical
`Endocrinology
`Copylight
`0 1994 by The Endocrine
`
`and Metabolism
`Society
`
`Vol. 79, No. 4
`in U.S.A.
`Pm&d
`
`CLINICAL
`Laboratory
`
`REVIEW
`Assessment
`
`62
`of Adrenal
`
`Insufficiency*
`
`STEVEN
`
`K. GRINSPOON,
`
`AND
`
`BEVERLY
`
`M. K. BILLER
`
`Neuroendocrine
`
`Unit, Massachusetts General Hospital, Boston, Massachusetts
`
`A DRENAL insufficiency results from dysfunction at one
`
`or more sites in the hypothalamic-pituitary-adrenal
`(HPA) axis. Primary hypoadrenalism is caused by bilateral
`adrenal gland destruction, whereas central hypoadrenalism
`is caused by CRH deficiency from hypothalamic dysfunction
`or ACTH deficiency from pituitary destruction. Central
`hypoadrenalism is sometimes termed secondary (pituitary)
`or tertiary (hypothalamic), depending on the level of dys-
`function. The clinical spectrum of adrenal insufficiency is
`broad, ranging from clear hemodynamic compromise to sub-
`tle dysfunction manifest only with stress. Laboratory evalu-
`ation of the HPA axis is performed in two clinical settings.
`The first is the diagnostic evaluation of the patient presenting
`with symptoms suggestive of adrenal insufficiency. In this
`setting, the initial goal is to determine whether cortisol pro-
`duction is adequate. The Cortrosyn stimulation test (Organon
`Diagnostics, West Orange, NJ), the insulin tolerance test
`(ITT), the metyrapone test, and serum cortisol levels are
`routinely used for this purpose. Once the diagnosis of adrenal
`insufficiency has been made, the next step is to localize the
`defect, usually accomplished by drawing a plasma ACTH
`level. In some cases, a prolonged ACTH
`infusion or CRH
`test may be used to establish whether the defect is located
`centrally or at the adrenal glands.
`The second clinical setting in which evaluation of the HPA
`axis is needed is the patient known to be at risk for the
`development of secondary hypoadrenalism. The most com-
`mon risk factor is prior glucocorticoid therapy. Other predis-
`posing factors include hypothalamic-pituitary disease and
`related surgery or radiation. Localization of the defect in such
`patients is usually obvious based on the specific risk factor,
`and the purpose of laboratory testing is to screen for the
`development of inadequate hypothalamic-pituitary
`reserve.
`Biochemical evaluation is based on three important prin-
`ciples of HPA physiology: 1) cortisol exerts feedback inhibi-
`tion at the pituitary and hypothalamic levels; 2) the adrenal
`glands depend on ACTH as a tropic hormone in such a way
`that ACTH deficiency results in a reversible inability
`to
`produce cortisol; and 3) the HPA axis can be activated by
`pharmacological and physiological stimuli that override the
`normal diurnal pattern of cortisol production. Numerous
`
`tests, both static and dynamic, are available to assess HPA
`function in patients who demonstrate symptoms of, or are at
`risk for, hypoadrenalism. The purpose of this article is to
`discuss the rationale, methods, criteria, limitations, and op-
`timal use of these tests in the evaluation of adrenal insuffi-
`ciency (see Table 1 for summary).
`
`Diagnostic Tests
`
`Serum cortisol
`
`level
`
`Cortisol production is a function of the coordinated activity
`of the entire HPA axis. Hormone release is pulsatile through-
`out the day, but exhibits a diurnal pattern with levels being
`the highest in the early morning. Production of cortisol
`increases in response to declining serum cortisol levels and
`to physiological stressors, such as hypoglycemia. Serum cor-
`tisol can be measured at a random point in time, as a morning
`value between 0600 and 0800 h, or as part of a dynamic test
`of HPA function. Interpretation of a serum cortisol level is
`complicated by a number of factors. Hydrocortisone, meth-
`ylprednisolone, and prednisone, but not dexamethasone,
`cross-react in the cortisol assay and should be avoided within
`24 h of testing. Because little of the hormone exists in the
`free or unbound state, the measured level of cortisol is a
`function of the predominant binding protein, cortisol-bind-
`ing globulin (CBG). Estrogen stimulates hepatic production
`of CBG, thereby resulting in higher total serum cortisol levels.
`CBG is decreased in cirrhosis, in the nephrotic syndrome,
`and in hyperthyroidism, but not usually to an extent that
`affects serum cortisol levels. Urine-free cortisol (UFC) is
`unaffected by such variables but is a poor diagnostic test for
`adrenal insufficiency because it is normal in 20% of patients
`with adrenal insufficiency (1). The UFC is also a poor test to
`judge the adequacy of glucocorticoid replacement because it
`reflects only the fraction of circulating cortisol that exceeds
`the binding capacity of CBG.
`
`Morning
`
`cortisol
`
`level
`
`The morning serum cortisol measurement has long been
`used as an index of adrenal function in the unstressed patient
`because it reflects peak endogenous activation of the HPA
`axis. However, the normal range, 9-25 pg/dL, is derived
`from morning cortisol levels in patients without HPA disease
`and does not necessarily differentiate normal subjects from
`those with adrenal dysfunction. For instance, in one series,
`
`923
`
`22, 1994.
`February
`20, 1993. Accepted
`December
`Received
`reprint
`requests
`to Beverly
`M. K. Biller, M.D.,
`Neuroendo-
`Address
`Jackson
`10, Massachusetts
`General
`Hospital,
`Boston, Mas-
`crine Unit,
`02114.
`sachusetts
`* This work was partially
`
`by NIH Grant
`
`RR-01066.
`
`supported
`
`The Endocrine Society. Downloaded from press.endocrine.org by [Deanna Carr] on 13 August 2016. at 08:13 For personal use only. No other uses without permission. . All rights reserved.
`
`
`
`JANSSEN EXHIBIT 2052
`Mylan v. Janssen IPR2016-01332
`
`

`

`ACTH production pulsa-
`
`coid
`exogenous glucocorti-
`within a few hours by
`tile and inhibited
`
`to lab
`and immediately sent
`
`Must be drawn on ice
`
`Not useful as a diagnos-
`
`tic test for AI
`
`May precipitate AI; per-
`
`increase me-
`
`See above
`blockade
`pone and decrease F
`tabolism of metyra-
`barbital
`
`Phenytoin and pheno-
`
`blood draw
`dose after morning
`Possible glucocorticoid
`formed as inpatient
`
`insulin
`needed to overcome
`Increased dose may be
`in attendance
`Physician
`patients
`psychiatric, or seizure
`derly, cardiovascular,
`in el-
`
`Contraindicated
`
`resistance
`
`with recent 2” or 3” AI
`False positive F response
`F increment not reliable
`
`sent
`back inhibition
`levels rise because feed-
`
`is ab-
`
`In primary AI, ACTH
`
`Best test to separate 1”
`
`from central AI
`
`(RIA) 2
`
`See above
`
`See above
`
`in-
`
`blockade
`proximal
`and steroidogenesis
`tion stimulates ACTH
`inhibi-
`ll&hydroxylase
`from
`Hypocortisolemia
`
`to enzymatic
`
`Sensitive test for central
`
`AI
`
`test: S 2 7.0 pg/
`
`stimulus of HPA axis
`Hypoglycemia a powerful
`
`Adrenal glands dependent
`
`ACTH
`on tropic effect of
`
`Directly assesses entire
`
`HPA axis
`
`test: F 2 18 fig/
`
`Indirectly assesses entire
`
`HPA axis
`
`Often indeterminate
`
`ing
`highest levels in morn-
`
`F production pulsatile;
`
`function
`
`Measure of HPA
`
`in stable patient
`
`test: 3 pg/
`
`test: 13
`
`Often indeterminate
`
`Stress activates HPA axis
`
`Measure of HPA function
`
`in acutely ill patient
`
`Liiitations
`
`Rationale
`
`Utility
`
`test: F 2 18 ).tg/
`Criteria
`
`100 pg/mL
`1” AI: ACTH
`
`AI: 17-OHS increase <2-
`
`3~ over baseline
`
`baseline
`crease ~2-3~ over
`test: 17-OHS
`
`Normal
`
`AI: S < 7.0 PgfdL and F
`
`< 5 rtd~
`
`PLg/~
`-.
`7.0 rg/dL and F 2 5.0
`test: S <
`
`Indeterminate
`
`Normal
`
`cemia 540 mg/dL
`symptomatic hypogly-
`
`AI: F 5 18 pg/dL and
`
`dL at any point
`
`Normal
`
`Normal
`test: F 2 18 rg/
`Definite AI: F 5 3 rg/dL
`
`dL at any point
`
`dL c F < 19 /.ig/dL
`
`Indeterminate
`
`Presumptive AI: 5 pg/dL
`
`dL4
`Normal
`
`N;yl
`test: F > 19 pg/
`Definite AI: F < 5 pg/dL
`< F < 18 /.vz/~
`
`4~
`
`Indeterminate
`
`5 F 5 13 pg/dL
`
`dL
`
`simultaneous F
`ill patient, when drawn with
`is made except in the acutely
`Drawn after the diagnosis of AI
`
`metyrapone
`line, during, and day after
`OHS and cr collected at base-
`for 17-
`sequential 24-h urine
`
`500-750 mg po q 4 h x 6 with
`
`level
`
`ACTH
`
`24 h metyrapone test
`
`F and S
`cumbent patient with 0800 h
`
`30 mg/kg po at 2400 h in re-
`
`test
`
`Overnight metyrapone
`
`at 0,30, and 60 min
`lin iv; glucose and F sampled
`0.1-0.15 U/kg short acting insu-
`
`iv; F sampled at
`
`250 pg ACTH
`
`0,30, and 60 min
`
`ITT
`
`test
`
`Cortrosyn stimulation
`
`Drawn between 0600 and 0800 h
`
`Morning cortisop*
`
`Drawn emergently with simulta-
`
`level
`
`neous ACTH
`
`Method
`
`Random cortisol”**
`Name
`
`insufficiency
`
`tests used in the evaluation of adrenal
`
`1. Laboratory
`
`TABLE
`
`The Endocrine Society. Downloaded from press.endocrine.org by [Deanna Carr] on 13 August 2016. at 08:13 For personal use only. No other uses without permission. . All rights reserved.
`
`

`

`$
`
`ij
`2
`F
`
`insufficiency.
`
`status.
`
`As above
`
`in
`re-
`
`to Cortrosyn
`
`1” AI
`spond
`
`will not
`
`Aldosterone
`
`Limited
`
`cen-
`
`dis-
`
`1” from
`reliably
`
`and potassium
`
`pos-
`
`by volume,
`
`tral AI
`tinguish
`
`status
`ture,
`Affected
`
`Test cannot
`
`deficiency
`to mineralocor-
`in re-
`
`increases
`
`PRA
`
`Limited
`
`> 3.0
`
`ticoid
`sponse
`
`dis-
`
`in pituitary
`and a flat
`re-
`in hypothalamic
`ACTH
`
`re-
`
`ease
`sponse
`disease
`sponse
`gerated
`
`tachypnea
`
`rare
`
`Flushing,
`
`an exag-
`
`produces
`
`CRH
`
`for AI
`
`test
`
`diagnostic
`
`future
`
`Possible
`
`necessary
`3” AI, but not often
`
`l”, 2”, and
`
`Distinguishes
`
`As above
`
`As above
`
`test
`
`5-day
`
`that
`
`consuming
`
`time
`
`Less
`
`2 10
`
`5 4 mg/
`
`levels
`and
`17-OHS
`hance F metabolism
`en-
`and phenytoin
`phenobarbital,
`totane,
`
`reduce
`
`mi-
`
`Aminoglutethimide,
`
`to
`
`exposure
`
`with
`
`in cen-
`
`reversed
`atrophy
`
`ACTH
`prolonged
`tral AI
`
`Adrenal
`
`central
`
`from
`
`1”
`
`Separates
`
`AI
`
`over base-
`in-
`
`I7-OHS
`
`AI, adrenal
`renal
`!’ assay.
`
`and
`the
`
`m
`
`17.hydroxysteroids,
`thyroid,
`
`liver,
`cross-react
`
`17-OHS,
`
`by estrogen,
`
`S; cr, creatinine,
`thus affected
`
`and
`
`S, compound
`dependent,
`
`and hydrocortisone
`
`methylprednisolone,
`
`F, cortisol;
`* F is CBG
`” l’reclmsone,
`
`rise after Cortro-
`
`< 5 ng/dL
`with-
`AI: Basal aldos-
`
`wn
`out
`terone
`
`Primary
`
`levels at baseline
`test with
`
`stimulation
`
`and 60 min
`aldosterone
`
`Cortrosyn
`
`test
`
`stimulation
`
`Aldosterone
`
`AI: PRA
`
`ng/dL
`Primary
`
`sample
`
`Random
`
`activity
`
`renin
`
`Plasma
`
`0, 15 30, 60,
`
`I20 min
`at -15,
`
`90, and
`sampled
`
`See Fig. 1
`
`iv; F and ACTH
`
`CRH
`
`1 pg/kg
`
`test
`
`CRH
`
`24 h
`
`I7-OHS
`
`1” AI:
`
`h
`
`mg/24
`
`AI: 17-OHS
`
`Central
`
`r3x
`AI:
`
`line
`crease
`
`Central
`
`infusion
`and cr on 1st and 2nd day of
`with
`3 rg/h
`
`for 17-OHS
`iv over 2 days
`
`24-h urine
`ACTH
`
`test
`each sub-
`
`day of the
`and during
`
`and cr at
`
`for 17-OHS
`
`sequent
`baseline
`urine
`each of 3-5 days with
`
`24-h
`
`IV over 8 h on
`
`rg of ACTH
`
`250
`
`(48 h)
`
`test
`
`infusion
`
`ACTH
`
`Continuous
`
`infusion
`
`days)
`ACTH
`
`(3-5
`
`test
`
`Prolonged
`
`The Endocrine Society. Downloaded from press.endocrine.org by [Deanna Carr] on 13 August 2016. at 08:13 For personal use only. No other uses without permission. . All rights reserved.
`
`

`

`926
`
`GRINSPOON
`
`AND BILLER
`
`Primary
`
`Adrenal
`
`Insufficiency
`
`Secondary
`
`Adrenal
`
`Insufficiency
`
`’ Piiuitary
`I
`
`Lesions
`
`Hypothalamic
`
`Lesions
`
`:
`
`:
`07
`
`A*/
`09
`.
`
`lO\
`
`130
`
`I20
`
`110
`
`loo
`
`J
`
`I
`
`:[
`2400
`
`2200-
`
`Moo-
`1800-
`
`1500-
`1400-
`1200-
`
`looo-
`
`9w-
`
`;
`,p
`?
`
`:
`
`i
`
`s600
`
`!i
`a.
`
`4o02
`40-
`
`30-
`
`M-
`
`10
`
`-
`
`(top) and cortisol
`FIG. 1. Plasma ACTH
`(bottom)
`responses to CRH
`in subjects
`with primary adrenal insufficiency (left)
`or
`secondary
`adrenal
`insufficiency
`(right). Patients with hypothalamic
`le-
`sions had clearly distinct ACTH
`re-
`sponses to CRH, different
`from those in
`three patients with pituitary adrenal
`in-
`sufficiency. Shaded
`area: Absolute range
`from 15 normal subjects. Reproduced
`with permission
`from H. M. Schulte et
`al, J Clin Endocriml
`(35) and The
`Metab
`Endocrine Society.
`
`1
`.
`2
`l
`9 3
`4 ion
`l
`
`prsdniaonel
`
`,
`
`.
`
`0,
`
`“”
`
`’
`
`’
`
`I
`
`I
`
`I
`
`i/
`/y/j /
`
`20
`
`10
`15
`5
`
`n
`
`I
`
`/
`
`,A
`
`/
`
`60
`
`90
`
`120
`
`150
`
`180
`
`TIME
`
`(mini
`
`90
`
`90
`
`70
`
`90
`
`50
`
`40
`
`30
`
`20
`
`10
`
`0
`
`20
`
`15
`
`10
`
`5
`
`0
`
`i
`
`-1501530
`
`50
`
`90
`
`120
`
`150
`
`180
`
`TIME
`
`Imin)
`
`15% of patients with documented primary adrenal insuffi-
`ciency had morning cortisol levels between 9 and 19 &dL
`(1). Furthermore, it is not clear from the literature what level
`of morning serum cortisol accurately predicts an adequate
`response to stress. In one study, a level of 11 pg/dL predicted
`an adequate response to the ITT in 51 of 52 patients evalu-
`ated for adrenal insufficiency (2). In another study, no patient
`with a morning cortisol level above 14 pg/dL failed an ITT
`(3). Thus, it is difficult to know what cutoff to choose, but it
`is well accepted that patients with morning cortisol levels
`above 19 do not need further testing.
`The morning cortisol level has also been used to predict
`the adequacy of adrenal function in the setting of recent
`pituitary surgery and chronic glucocorticoid therapy. It was
`shown in one study that a morning cortisol level over 9 pg/
`dL several days after pituitary surgery predicts a normal
`response to the ITT 2-3 days later (4). These data have not,
`however, been reproduced. Debate also exists about the use
`of the morning cortisol level in patients on chronic glucocor-
`ticoid therapy. Only half of the patients with a morning
`cortisol level greater than 5 pg/dL demonstrated a normal
`
`response to CRH in a recent study (5). It was concluded that
`the morning cortisol level does not reliably predict the re-
`sponse to dynamic testing in this patient population. How-
`ever, the morning cortisol level that was chosen to define
`normal adrenal function, 5 pg/dL, was arbitrary and quite
`low. A higher level might better have predicted a normal
`response to dynamic testing.
`Although a low morning cortisol level may simply reflect
`a nadir between glucocorticoid pulses, there are substantial
`data which confirm that a level of less than 3 rg/dL is always
`diagnostic of adrenal insufficiency (2, 4,5). The major draw-
`back of the morning cortisol level is that many patients have
`indeterminate levels and require additional testing. Never-
`theless, the test is easy to perform and is a good first assess-
`ment of adrenal function in the stable patient because a very
`low or high level obviates the need for further tests.
`
`Random
`
`cortisol
`
`level
`
`A random cortisol level is sometimes useful to evaluate
`adrenal function in the clinical setting of severe stress (when
`endogenous HPA activity should be at a maximum) and
`
`The Endocrine Society. Downloaded from press.endocrine.org by [Deanna Carr] on 13 August 2016. at 08:13 For personal use only. No other uses without permission. . All rights reserved.
`
`

`

`CLINICAL
`
`REVIEW
`
`927
`
`level or dynamic
`for a morning
`there is no time to wait
`when
`testing,
`In
`the acutely
`ill patient, glucocorticoid
`therapy
`should be given
`immediately
`after blood
`is drawn, with
`further
`testing performed at a later time. If dexamethasone
`is chosen
`for glucocorticoid
`replacement, a Cortrosyn
`test
`can subsequently
`be performed.
`is adequate
`It is debated what
`level of cortisol production
`in the acutely
`ill patient. A cortisol
`level below 18 pg/dL
`during severe stress is often taken as evidence of inadequate
`adrenal
`function, but it is extrapolated
`from
`the literature
`regarding
`the rapid ACTH
`test. The range of serum cortisol
`levels in the acutely ill patient is broad. Mean random cortisol
`levels of 22,40, and 45 rg/dL have been reported
`in patients
`with gastrointestinal
`bleeding, respiratory
`failure, and sepsis
`respectively.
`Intensive care patients with cortisol levels below
`13 rg/dL exhibit inadequate Cortrosyn
`stimulation,
`increased
`mortality, and improvement
`in clinical condition after glu-
`cocorticoid administration
`(6). However,
`critically
`ill patients
`with
`random cortisol
`levels as low as 5 pg/dL have been
`known
`to survive and exhibit adequate stimulation
`to ACTH
`(7). A random cortisol
`level below 5 pg/dL during severe
`stress
`is definitive evidence of adrenal
`insufficiency. A level
`below 13 pg/dL
`is presumptive
`evidence of this disorder
`mandating glucocorticoid
`therapy
`to reduce potential mor-
`bidity. Levels above 13 rg/dL
`but below
`18 rg/dL
`are
`indeterminate
`and necessitate
`further
`testing and interim
`glucocorticoid
`therapy.
`
`The Cortrosyn stimulation
`
`test (l-h ACTH
`
`test)
`
`The Cortrosyn stimulation test, in which cortisol is meas-
`ured 0, 30, and 60 min after the iv administration of 250 PLg
`synthetic ACTHlvz4, is an excellent diagnostic test for patients
`suspected of having chronic adrenal insufficiency. In most
`centers, it has supplanted the 8-h infusion test because it is
`easier to perform and achieves similar cortisol responses (8).
`Although
`this test directly measures only the functional
`integrity of the adrenal glands, it also provides an indirect
`assessment of hypothalamic and pituitary
`function because
`the adrenal glands depend on endogenous ACTH
`for its
`tropic effect. When ACTH production is impaired by pitui-
`tary or hypothalamic disease, the adrenal gland loses the
`capacity to respond to exogenous stimulation.
`Two aspects of the Cortrosyn test have generated much
`debate: 1) selection of the best test criterion, i.e. peak cortisol
`response us. increment; and 2) definition of an adequate level
`of cortisol response. Historically, both the increment in cor-
`tisol and the absolute level of cortisol were viewed as impor-
`tant. However, the increase in cortisol following Cortrosyn
`administration is an unreliable index of adrenal function
`because it fails to distinguish normal patients from those
`with adrenal insufficiency. This was demonstrated by a study
`in which one third of normal controls exhibited a rise in
`cortisol of less than or equal to 7 pg/dL (9). Because the
`cortisol increment is inversely proportional to the basal cor-
`tisol level, smaller increases are obtained in the morning,
`when endogenous ACTH and cortisol levels are already high
`(10, 11). The peak cortisol response to Cortrosyn, which is
`unaffected by the time of day, is thus a more useful measure
`
`of adrenal function than the increment (12).
`Peak cortisol levels ranging from 15-25 pg/dL have been
`proposed as criteria for establishing adequate adrenal func-
`tion based on the Cortrosyn test. Values of 18-20 pg/dL are
`well supported in the literature. In one retrospective series,
`cortisol responses to Cortrosyn were examined in a hetero-
`geneous population of 399 patients. In 95% of the tests, the
`peak cortisol level was greater than 19 pg/dL, although no
`data were provided regarding the actual incidence of adrenal
`insufficiency nor correlation made with other tests of adrenal
`function (10). In another series, a value of 18 pg/dL com-
`pletely separated patients with known adrenal insufficiency
`from normal controls (9). The usefulness of the Cortrosyn
`test in detecting adrenal insufficiency is demonstrated by
`comparing it to other dynamic tests. Only 8 discrepancies
`were found when the Cortrosyn stimulation test and the ITT
`were compared in 200 consecutive patients evaluated pro-
`spectively for adrenal insufficiency. In 6 cases the discrep-
`ancies were either minor or were attributed to inadequate
`hypoglycemic response. In 2 cases of acute pituitary dys-
`function, however, the Cortrosyn test was normal whereas
`the ITT was not (13).
`Such cases illustrate the recognized phenomenon of a short
`window period of a few weeks after the onset of pituitary
`dysfunction, during which the adrenal glands are still capable
`of responding to exogenous ACTH (14). In the de nova
`evaluation of patients with hypothalamic-pituitary disease,
`the duration of adrenal insufficiency may not be known, and
`the Cortrosyn test may be misleading, as shown in a study
`demonstrating normal Cortrosyn test results but abnormal
`ITT responses in patients with pituitary disease (15). The
`most common clinical setting in which the test must not be
`used is immediately after pituitary surgery, when corticotrope
`damage may have occurred but the adrenal glands have not
`yet atrophied from the loss of ACTH effect. In this setting,
`it is essential to assess pituitary
`reserve with a morning
`cortisol, metyrapone test, or ITT. After a month, the test of
`choice is probably the Cortrosyn test, although normal Cor-
`trosyn but abnormal ITT responses have been seen up to 3
`months after pituitary surgery (16).
`The most common cause of adrenal insufficiency is HPA
`axis suppression due to the use of exogenous glucocorticoids.
`A key question regarding patients on such therapy is whether
`they can safely tolerate the stress of a surgical procedure
`without steroid administration. Various tests of adrenal func-
`tion, including the Cortrosyn stimulation test, metyrapone
`test, and ITT have been used to predict the adrenal response
`to surgery in this patient population. Cortisol responses to
`preoperative Cortrosyn stimulation correlate remarkably well
`with intraoperative cortisol levels (17). Furthermore, patients
`with a subnormal response to the Cortrosyn test demonstrate
`a smaller perioperative increase in cortisol than that seen in
`the patients with a normal Cortrosyn but abnormal insulin
`tolerance or metyrapone test (18). In contrast, abnormal
`metyrapone responses have been demonstrated in patients
`previously receiving high dose glucocorticoid therapy who
`have a normal Cortrosyn test. However, the data do not
`permit a comparison to be made between the adrenal re-
`
`The Endocrine Society. Downloaded from press.endocrine.org by [Deanna Carr] on 13 August 2016. at 08:13 For personal use only. No other uses without permission. . All rights reserved.
`
`

`

`928
`
`GRINSPOON
`
`AND BILLER
`
`JCE & M - 1994
`Vol79.No4
`
`stress,
`
`and to a severe physiological
`
`to metyrapone
`sponse
`such as surgery
`(16).
`test is the ideal method
`stimulation
`Thus,
`the Cortrosyn
`in all cases except
`those
`function
`for evaluating adrenal
`involving
`recent hypothalamic
`and/or pituitary dysfunction.
`A peak cortisol response of 18 &dL
`at any time during
`the
`test is an adequate response. Minimally abnormal
`tests (13-
`17 pg/dL) should be repeated at least once, because they are
`often normal on repeat
`testing
`(10). The decision about
`whether
`to give replacement glucocorticoid
`therapy
`in such
`a situation will depend on the index of suspicion
`for adrenal
`insufficiency.
`
`ITT
`
`The ITT has long been long considered the standard for
`the diagnosis of adrenal insufficiency because it assesses the
`integrity of the entire HPA axis. Hypoglycemic stress is a
`powerful, but indirect, stimulus to cortisol production via
`activation of the hypothalamus and pituitary. The major
`drawbacks of the ITT are: 1) it is contraindicated in the
`elderly and in those with cardiovascular disease or seizure
`history, and 2) it requires that a physician be present through-
`out the test to monitor for adrenergic and neuroglycopenic
`symptoms. Short-acting regular insulin is injected iv at doses
`ranging between 0.1 U/kg and 0.15 U/kg. The dose of insulin
`may be adjusted upwards in insulin resistant states, such as
`acromegaly or obesity. Downward adjustments may prevent
`prolonged glucose suppression in those with hypopituitar-
`ism, but doses of <O. 1 U/kg do not reliably achieve adequate
`hypoglycemia (19).
`As in the Cortrosyn test, the peak serum cortisol response
`to insulin-induced hypoglycemia is a more reliable criterion
`than the increment in cortisol for confirming adequate HPA
`axis function (20). Adequate adrenal function is demon-
`strated by a peak cortisol level over 18 pg/dL at any time
`during the test, a cutoff that reliably separates normal con-
`trols from patients with adrenal insufficiency (21). Hypoa-
`drenalism is demonstrated by a peak cortisol level of less
`than 18 pg/dL with symptomatic hypoglycemia (glucose <
`40 mg/dL). A rare patient will pass the ITT while exhibiting
`clinical adrenal insufficiency. This was demonstrated in a
`small study of six patients, all of whom had normal ITTs,
`but exhibited hypocortisolemia by frequent sampling and
`amelioration of symptoms with glucocorticoid replacement
`(22). A partial CRH deficiency is thought to explain these
`results. Nevertheless, the ITT is an excellent method for
`assessing dysfunction at any level of the HPA axis and is a
`sensitive test for evaluating pituitary reserve in patients at
`risk for the development of secondary adrenal insufficiency.
`The metyrapone test may be used when the ITT is contrain-
`dicated.
`
`Metyrapone
`
`test
`
`The metyrapone test was developed specifically as a test
`of pituitary reserve (23). Metyrapone
`inhibits the adrenal
`enzyme 1 l/&hydroxylase, which converts 11 -desoxycortisol
`(compound S, 11-S) to cortisol in the final step of adrenal
`
`steroidogenesis. 11 -S does not have glucocorticoid activity
`and therefore does not inhibit ACTH production. When
`metyrapone is given to a normal subject, the decline in serum
`cortisol stimulates ACTH production, driving adrenal ste-
`roidogenesis proximal to the enzyme blockade and causing
`11-S to accumulate. When metyrapone is given to a patient
`with adrenal insufficiency of any type, 11 -S fails to increase.
`As originally described, the metyrapone test was per-
`formed by administering 500-650 mg of the drug orally
`every 4 h for six doses, with measurement of plasma and
`urine 17-hydroxysteroids over 2 days (23). Interpretation of
`urinary 17-hydroxysteroids in this test is shown in Table 1.
`The test has been simplified in recent years, and is now most
`often performed as an overnight test with a single dose of
`30 mg/kg administered orally at 2400 h, with a snack to
`prevent gastric irritation (24). Blood is drawn at 0800 h the
`next morning for measurement of serum cortisol and 11-S.
`Because of the risk of acute adrenal insufficiency which may
`be precipitated by metyrapone, the test should be conducted
`on an inpatient basis. After the morning blood sample has
`been obtained, it is appropriate to administer a prophylactic
`dose of prednisone or hydrocortisone to patients in whom
`hypopituitarism is highly suspected (25).
`Adequate adrenal function is confirmed by an 11-S value
`greater than 7.0 pg/dL, regardless of the cortisol level (24).
`Adrenal insufficiency is diagnosed if the 11-S response is
`less than 7.0 pg/dL and the simultaneously drawn serum
`cortisol is low. A low cortisol level confiis adequate enzyme
`blockade and insures sufficient pituitary stimulation. Cortisol
`values ranging from 2-7.5 pg/dL have been used to establish
`adequate 1 l/3-hydroxylase inhibition, with 5 pg/dL, a com-
`monly accepted but arbitrary value. If the 11-S level is below
`7.0 pg/dL but the cortisol level above 5 pg/dL, the metyra-
`pone test is indeterminate because the hypocortisolemia was
`not adequate to stimulate pituitary production of ACTH.
`The metyrapone test may be the most sensitive method
`for detecting adrenal insufficiency. In one study, patients
`exposed to prior glucocorticoid treatment or recent pituitary
`surgery were prospectively evaluated with the metyrapone,
`insulin tolerance, and Cortrosyn tests. Discordant results
`from the 3 tests were seen in 6 of 25 patients, with abnormal
`response to Cortrosyn in 1, insulin in 3, and metyrapone in
`4 (26). In a recent study of HPA axis function after cranial
`radiation, morning cortisol, ACTH, CRH, and metyrapone
`tests were performed on 31 patients. No patient had symp-
`toms of adrenal insufficiency or an abnormal ACTH test, yet
`35% had an abnormal metyrapone test by traditional criteria
`(27). It is not clear, however, whether the metyrapone test
`detects clinically relevant adrenal hypofunction
`in this set-
`ting. Further studies will be needed to clarify this point.
`Many commonly used drugs such as glucocorticoids, phen-
`ytoin, and phenobarbital can affect the metyrapone test.
`Phenytoin and phenobarbital increase the metabolism of
`metyrapone, thereby reducing plasma metyrapone levels and
`decreasing the ll@hydroxylase blockade (28). The standard
`recommendation is to discontinue phenytoin 2 weeks before
`testing. If this is not possible, the conventional metyrapone
`dose should be given initially because it is usually effective
`
`The Endocrine Society. Downloaded from press.endocrine.org by [Deanna Carr] on 13 August 2016. at 08:13 For personal use only. No other uses without permission. . All rights reserved.
`
`

`

`CLINICAL
`
`REVIEW
`
`929
`
`even in patients on chronic phenytoin therapy (24). If the
`test is indeterminate, the metyrapone dose should be doubled
`WV
`
`Localizing
`
`Tests
`
`Once the diagnosis of adrenal insufficiency has been con-
`firmed by the appropriate test, the second step is to determine
`the site of the HPA axis defect. This is critical because it
`directs the subsequent evaluation, i.e. magnetic resonance
`imaging of the head to visualize the hypothalamus and
`pituitary vs. computed tomography of the abdomen to visu-
`alize the adrenal glands. It is also important to localize the
`site of hormone deficiency because therapy differs with
`respect to the need for mineralocorticoid replacement. In
`secondary adrenal insufficiency, aldosterone production re-
`mains adequate and only glucocorticoid replacement is nec-
`essary. In contrast, patients with primary adrenal insuffi-
`ciency exhibit hypoaldosteronism and require mineralocor-
`ticoid replacement.
`
`ACTH
`
`level
`
`level is the best test to distinguish
`The plasma ACTH
`primary from secondary adrenal insufficiency. ACTH secre-
`tion is pulsatile and increases in response to hypocortisolemia
`if the hypothalamus and pituitary gland are intact. The level
`is invariably above 100 pg/mL by RIA in primary adrenal
`insufficiency. Numerous factors effect interpretation of the
`test, including half-life, stability, and glucocorticoid
`ACTH
`administration. The short plasma half-life of ACTH and its
`vulnerability
`to cellular enzymes mandate that the test be
`drawn on ice and immediately sent to the laboratory to avoid
`falsely low values. To make the diagnosis of primary adrenal
`insufficiency, the ACTH level must be drawn prior to glu-
`cocorticoid administration because the level can decrease to
`normal within a few hours after steroids are begun. This
`phenomenon also accounts for the fluctuating ACTH levels
`in patients with Addison’s disease on replacement therapy
`(29). Despite these drawbacks, the ACTH level is a simple,
`effective means for distinguishing primary from secondary
`insufficiency and should always be drawn simultaneously
`with the initial cortisol level in the workup of adrenal insuf-
`ficiency.
`level is a
`It is important to emphasize that the ACTH
`localizing test and is not useful by itself in the initial evalu-
`ation of suspected adrenal insufficiency because it fails to
`separate normal patients from those with central hypoadren-
`alism. The range of ACTH values among control subjects
`and patients with proven pituitary disease was nearly iden-
`tical in a recent study, 4-81 pg/mL and 8-75 pg/mL respec-
`tively (30).
`
`Prolonged ACTH
`
`infusion
`
`Before the development of reliable ACTH assays, a pro-
`longed ACTH infusion test was required to distinguish pri-
`mary from secondary adrenal insufficiency. To perform the
`test, 250 pg ACTH are infused over 8 h on each of 3-5 days
`
`with measurement of sequential 24-h urines for 17-hydroxy-
`steroids. In primary adrenal insufficiency, the adrenal gland
`is not capable of responding to ACTH stimulation. In con-
`trast, a gradual rise in 17-hydroxysteroids to 3-5 times base-
`line is seen in secondary adrenal insufficiency as the adrenal
`gland gradually recovers function (31). The utility of the test
`is demonstrated by the near zero (1 W) overlap in 17-hydroxy-
`steroid responses between patients with primary and second-
`ary disease by the 5th day (32). This test can also be per-
`formed as a 48-h continuous infusion in which patients
`receive 3 U/h of ACTH over 2 days. A 24-h urine 17-
`hydroxysteroid response on the second day of at least 10
`mg/24 h is seen in patients with secondary adrenal insuffi-
`ciency, whereas a response of less than 4 mg/24 h is seen in
`patients with primary adrenal insufficiency (32). This 48-h
`test compares favorably to the longer ACTH infusion tests
`in sensitivity and ease of performance. Of