| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Endocrine Care |
Department of Medicine (W.J.I.), St. Vincent’s Hospital and University of Melbourne, Melbourne, Australia 3065; and Department of Endocrinology (P.J.H.), Christchurch Hospital, Christchurch, New Zealand 8001
Address all correspondence and requests for reprints to: Warrick Inder, M.D., Department of Medicine St. Vincent’s Hospital, 41 Victoria Parade, Fitzroy VIC 3065, Australia. E-mail: . winder{at}medstv.unimelb.edu.au
Abstract
Patients undergoing surgical resection of pituitary adenomas are frequently given perioperative glucocorticoid therapy. There are no randomized controlled studies assessing the need for such steroids; however, several studies have documented changes in the hypothalamic-pituitary-adrenal (HPA) axis associated with pituitary surgery. Based on the evidence available, this article details recommendations for the perioperative management of glucocorticoid therapy in patients with pituitary tumors. For patients with proven ACTH deficiency preoperatively [usually based on response to a short ACTH 1–24 (Synacthen) test], 48 h of supraphysiological glucocorticoid therapy should be administered perioperatively (e.g. hydrocortisone, 50 mg every 8 h on d 0, 25 mg every 8 h on d 1, and 25 mg at 0800 h on d 2). For patients with intact HPA function preoperatively, and in whom selective adenomectomy is possible, perioperative glucocorticoids are not necessary. Early postoperative assessment depends on daily clinical assessment of the patient and 0800 h plasma cortisol levels. Cortisol levels over 450 nM (16 µg/dl) reflect normal HPA function, and levels less than 100 nM (3.6 µg/dl) are consistent with ACTH deficiency. Cortisol levels between 100 and 250 nM (3.6–9 µg/dl) may be ACTH deficient and should receive morning hydrocortisone replacement until definitive HPA axis testing. Cortisol levels between 250 and 450 nM (9–16 µg/dl) are unlikely to be ACTH deficient but should receive additional steroids for stress until a definitive test is performed. For those requiring definitive testing, the insulin tolerance test, the overnight metyrapone test, or the glucagon stimulation test are appropriate and may be performed as early as d 7–10 or, if more convenient, wk 4–6. Following the guidelines suggested here should reduce the use of unnecessary glucocorticoids, while ensuring the safety of patients is not compromised.
CURRENTLY, THERE EXIST a variety of local practices for the management of glucocorticoid replacement around the time of pituitary surgery. Though randomized controlled trials have not been performed to assess the necessity of steroid cover, there are a number of studies that have investigated changes in the hypothalamic-pituitary-adrenal (HPA) axis associated with pituitary surgery. The main purpose of this article is to review the evidence for the use of perioperative glucocorticoids and to produce some recommendations for perioperative assessment and management in non-ACTH-secreting pituitary adenomas. The special situation of perioperative management of ACTH-secreting adenomas is also briefly discussed.
Preoperative assessment of HPA axis function
Before pituitary surgery, all patients with pituitary disease should be subjected to a diagnostic test of HPA axis function. It is beyond the scope of this review to make a detailed assessment of the strengths and weaknesses of each test, a topic which has been extensively examined by others. Traditionally, the insulin tolerance test (ITT) has been the gold standard test for integrity of HPA function (1) and, in some units, is still the preferred preoperative test (2). However, because these patients require reassessment of adrenal function postoperatively, subjecting them to an ITT at this stage may be considered excessive. Most endocrinologists prefer a version of the short ACTH 1–24 (Synacthen) test as their initial preoperative test of HPA axis function in a patient with a known pituitary tumor (3, 4). Opinion is divided concerning the optimal dose of ACTH 1–24, with both the standard 250-µg and the low-dose 1-µg tests having their supporters (5, 6, 7, 8). In the setting of long-standing ACTH deficiency, the 250-µg ACTH 1–24 test generally correlates well with the ITT (9); however, some cases of ACTH deficiency may be missed (2, 10, 11). Both the definition of the so-called normal cutoff (12) and the cortisol assay used (13) affect the interpretation of the test. We have used a cutoff of more than 550 nM (20 µg/dl), at 60 min, to define a normal cortisol response to 250 µg ACTH 1–24 in our population. This was changed from 500 nM (18 µg/dl) in 1998 after reviewing our local data, which showed a number of patients with proven ACTH deficiency who had peak cortisol levels between 500 and 550 nM (unpublished observation). Different centers have used cutoff values of between 500–600 nM in published reports (2, 4, 6, 12, 14). Higher cutoff values improve the sensitivity of the test in the diagnosis of ACTH deficiency but may misclassify some normal subjects as being deficient (reduced specificity) (12).
When performed early in the morning, an ACTH 1–24 test also allows an assessment of 0800 h (or in some centers, 0900 h) basal cortisol levels. For the purposes of deciding preoperatively whether an individual has a high or low probability of ACTH deficiency, a basal early morning cortisol followed by either the 1-µg or 250-µg test, depending on local protocol and clinician preference, provides sufficient information.
Perioperative management
The first decision regarding the perioperative use of glucocorticoid cover depends on the result of the preoperative screening test (15).
Abnormal ACTH 1–24 (Synacthen) test.
If the ACTH 1–24 test is abnormal, the patient should be commenced on standard maintenance doses of glucocorticoid (15–30 mg hydrocortisone daily, depending on factors such as age, sex, and body habitus) in the lead-up to surgery. These patients should be treated perioperatively with 48 h of supraphysiological glucocorticoid therapy (15, 16), with rapid dose reduction. A suggested dose regimen, using hydrocortisone, is 50 mg every 8 h on d 0, 25 mg every 8 h on d 1, and 25 mg at 0800 h on d 2. An alternative regimen, using dexamethasone, is 4 mg at induction of anesthesia, 2 mg at 0800 h on d 1, and 0.5 mg at 0800 h on d 2. Dexamethasone has the advantage of a long half-life, which allows for single daily administration. Though there are no data from randomized controlled trials, it is clear from detailed study of cortisol production rates, around the time of major surgery, that most traditional perioperative glucocorticoid replacement schedules markedly overtreat patients (17). The above dosage recommendations provide adequate cover during the perioperative period, based on normal subjects’ cortisol response to surgical stress. Individuals with normal HPA axis function show a return to baseline cortisol secretion within 48 h of major surgery, providing that no complications have occurred (17). In addition, many patients show a rapid increase in pituitary hormone secretion after pituitary adenomectomy, presumably as a result of increased flow of hypothalamic releasing hormones through the hypothalamic-hypophysial portal system (18). Thus, there is the potential for a patient with hypopituitarism preoperatively to gain full functional recovery, providing that viable normal pituitary tissue remains in situ. Prolonged treatment with high doses of glucocorticoids postoperatively may result in adrenal suppression and mask those who spontaneously recover function.
Provided there are no postoperative complications, glucocorticoid supplementation should be withdrawn, after 48 h, and measurement of 0800 h plasma cortisol levels performed daily between d 3 and 5 postoperatively. Daily clinical assessment of the patient, together with these cortisol results, will determine the subsequent use of glucocorticoid replacement therapy (see below).
Normal ACTH 1–24 (Synacthen) test.
If the preoperative ACTH 1–24 test is normal and the neurosurgeon is able to perform a selective adenomectomy, we recommend that no perioperative glucocorticoid cover be given (15). There is good evidence that such an approach is safe in this patient group (15, 19). If selective adenomectomy is not possible and surgery is more extensive, then it is prudent to provide glucocorticoid cover for 48 h and proceed as above. Otherwise, we recommend measuring 0800 h plasma cortisol on d 1–3 postoperatively. Again, the subsequent use of glucocorticoid replacement depends on these results (see below) and the clinical state of the patient.
ACTH-secreting adenomas
Cushing’s disease presents a special situation, with all patients requiring perioperative glucocorticoid cover. The presence of low plasma cortisol levels postoperatively implies a surgical cure (20), although the level of plasma cortisol defining a so-called cure remains debated (21). Patients with low plasma cortisol levels after pituitary surgery for Cushing’s disease require physiologic hydrocortisone replacement therapy with ongoing review of their HPA axis. Here, the aim is to gradually withdraw the hydrocortisone as the ACTH suppression recovers. Normal plasma cortisol levels postoperatively usually indicate ongoing autonomous ACTH secretion. The clinician must decide whether early repeat surgery is indicated (22) or use an alternative modality of treatment, such as radiotherapy (23) or adrenolytic therapy. Estimation of 24-h urinary free cortisol may help in this decision and may be performed while receiving physiological replacement with dexamethasone (0.5–0.75 mg daily). The CRH test and ACTH 1–24 test have also been used in this setting but, overall, seem to add little to the basal measurements of plasma and urinary free cortisol in predicting relapse. For a detailed review of this topic, see Ref. 21 .
Early postoperative assessment
The decision regarding ongoing glucocorticoid replacement should initially be based on the levels of 0800 h plasma cortisol in the early postoperative period (15, 16). These measurements should be made on d 1–3 in patients not treated with glucocorticoids and d 3–5 in patients covered with glucocorticoids for the initial 48 h. If there are clinical symptoms and signs of cortisol deficiency or there are significant delays in obtaining a cortisol level, replacement hydrocortisone should be given until the result is available. Ideally, this should be as a single morning dose of 15–20 mg hydrocortisone; but if the patient is medically unstable, an initial 50-mg iv dose of hydrocortisone may be necessary.
Review of the literature demonstrates several conflicting opinions regarding what constitutes a safe level of 0800 h (or 0900 h) plasma cortisol (14, 15, 16). Table 1
outlines suggested guidelines that allow selective use of provocative tests of HPA axis function and does not result in patients with normal HPA axis function being overtreated. Patients with an 0800 h cortisol of more than 450 nM postoperatively have an extremely low risk of adrenal insufficiency and should be regarded as having normal HPA axis function (14). No further dynamic testing is necessary.
|
Patients with cortisol levels between 100 and 250 nM are possibly ACTH deficient and should be treated with a single (nonsuppressive) morning dose of hydrocortisone (10–20 mg), with instructions to increase levels during times of illness or other stress, until definitive testing of the HPA axis is performed.
Patients with cortisol levels between 250 and 450 nM are unlikely to be ACTH deficient but may require definitive testing (14). These patients should not receive replacement glucocorticoid therapy but should be given a supply of hydrocortisone tablets and instructions for their use during stress, until a definitive test has been performed. It has been suggested that, in this group, an 0800 h plasma cortisol level more than 350 nM at d 7 is highly predictive of a normal response to insulin-hypoglycemia (16). Therefore, repeat 0800 h cortisol at d 7 may reduce the number of patients requiring a more invasive test of their pituitary function.
It is in this group that the most controversy exists regarding the need for a definitive postoperative test of the HPA axis. A number of investigators have suggested that postoperative cortisol levels of more than 250–270 nM are highly predictive of a normal response to ITT (14, 16, 26). However, in their 1994 review, Jones et al. (27) determined that a basal cortisol level of 500 nM was required before a normal response to ITT could be predicted with any degree of certainty. They used a peak cortisol of 580 nM to define a normal cortisol response to ITT, which is higher than many centers would use. Different centers use a peak cortisol of between 500 and 600 nM to define a normal response (28). Reviewing their data, a morning (0830–0930 h) cortisol of more than 350 nM was predictive of a peak cortisol response to ITT of more than 550 nM, with just 2 of 35 subjects with basal cortisol above 350 nM, demonstrating borderline fails the ITT (peak cortisol of close to 500 nM) (27). Examination of earlier work from Pavord et al. (29) and Hagg et al. (30) also suggests that basal cortisol more than 350 nM strongly predicts a normal response to ITT. Of over 100 patients in the Pavord et al. series, only four with cortisol more than 350 nM seem to have failed the ITT, with peak cortisol between 500 and 550 nM in 3 subjects, and the other failure recorded a peak cortisol of 400 nM despite a basal cortisol of 500 nM (29). Hagg et al. demonstrated 1 ITT failure in 33 patients with basal cortisol more than 350 nM (30). Overall, the data show that less than 4% of patients with morning cortisol of more than 350 nM fail the ITT, and these are usually associated with peak cortisol levels close to or above 500 nM.
Definitive postoperative assessment
There are three main questions to answer:
Who needs a definitive test of their HPA axis?
Reviewing the available literature, one could make a case for performing a provocative test of HPA function only in those patients with postoperative 0800 h cortisol levels between 100 and 350 nM (14, 16, 18, 26). Based on these studies, this approach misclassified less than 4% of patients, where basal cortisol levels more than 350 nM were associated with a cortisol response to ITT of less than 550 nM (see above). Several of these had peak cortisol levels over 500 nM, which may represent an adequate response or, at least, require repeat testing before a definitive diagnosis of secondary adrenal insufficiency could be made. A repeat 0800 h plasma cortisol at 4–6 wk, 24 h after the last dose of hydrocortisone, may provide additional evidence for normal HPA function. Plasma cortisol levels more than 350 nM are highly suggestive of normal pituitary-adrenal function. Individual endocrinologists seeking a degree of certainty greater than 96% should use 450 nM as their cutoff.
Those patients with levels less than 100 nM are likely to remain ACTH deficient; however, there are cases where late recovery of the HPA axis is documented (18, 26). A repeat 0800 h cortisol, 24 h after the last dose of hydrocortisone, performed between 4 and 6 wk postoperatively, should identify such patients. Those still less than 100 nM require lifelong replacement therapy, whereas those whose cortisol levels have risen to more than 100 nM should undergo a definitive test.
Which test should be performed?
In all tests of HPA axis function, the first problem is defining a normal response, as already discussed for the ACTH 1–24 test. We would encourage laboratories to establish their own local reference range for the test(s) they routinely perform, taking into account the characteristics of the local population and assay used (14, 31, 32).
It is clear that ACTH 1–24 testing in the early postoperative period is not the investigation of choice (2, 14), presumably, in part, because of the inability of this test to detect recent onset ACTH deficiency. The standard 250-µg dose test will underdiagnose ACTH deficiency, whereas the 1-µg test may misclassify individuals as being deficient (14). Commonly, in North America, where access to metyrapone may be difficult, an ACTH 1–24 test is performed at the 6-wk time point. Though, in most cases, this may be adequate, there is potential for cases of secondary adrenal insufficiency to be misdiagnosed (11). Unfortunately, there are no studies specifically examining the cortisol response to ACTH 1–24 at 6 wk after pituitary surgery, compared with a test of the entire HPA axis, such as the ITT or metyrapone test in nonfunctioning adenomas. Further research on this specific topic is required, investigating both the 250- and 1-µg tests.
The gold standard test for normal HPA function has been the ITT (1, 2, 15). Despite a low incidence of serious adverse events (1), many clinicians are reluctant to subject their patients to ITT in situations where another test of the HPA axis might suffice. There is considerable literature concerning the place of the ITT vs. other tests in this assessment (2, 32), including a recent commentary addressing the issue of a normal response to ITT (28). The original work by Plumpton and Besser (33) defined a normal cortisol response, using a fluorometric assay, to be more than 580 nM. However, with improvements in assay specificity, more recent work has shown a mean 96-nM positive bias using the fluorometric assay (34). Therefore, a normal cutoff of more than 500 nM has been proposed and is supported by some recent studies in normal subjects (35). Ultimately, as in defining a normal response to ACTH 1–24, results are assay dependant, and there are centers where 550 nM is used (14).
There are several advantages to using the ITT. First, it examines the entire HPA axis, providing a stimulus at the level of the hypothalamus. Second, recent studies have demonstrated satisfactory reproducibility of the cortisol response to ITT in normal subjects and in determining cortisol deficiency vs. normality (36, 37). Though there was significant variability in the peak cortisol value in hypopituitary patients, repeat testing did not misclassify any of the subjects (37). However, ITT is contraindicated in patients with a history of ischemic heart disease, arrhythmia, or epilepsy and should be used with caution in the over-65 age group. It is also more time consuming for nursing and medical staff and has greater consumable and assay costs.
An alternative is the overnight metyrapone test (38, 39). This can be performed safely on an outpatient basis, using 30 mg/kg metyrapone taken at 2400 h, with food, to minimize gastrointestinal side effects. Published data (38) and our own unpublished observations using this test, reveal a very low rate of adverse effects. Plasma cortisol and 11-deoxycortisol are measured at 0800 h on the morning after metyrapone administration (39). The metyrapone test also assesses the entire HPA axis but relies on the fall in glucocorticoid negative feedback, rather than direct stimulation of hypothalamic ACTH secretagogues, to stimulate ACTH secretion. The usual criteria for a normal response is an 11-deoxycortisol level of more than 200 nM (7 µg/dl) (32, 38, 39). Some investigators believe that the measurement of ACTH adds sensitivity (subjects with a subnormal ACTH response, but adequate 11-deoxycortisol response, being labeled as having a partial ACTH deficiency) (39). In normal subjects, a reference range for the ACTH response more than 33 pM (150 pg/ml) has been proposed, with levels between 13 and 33 pM (60–150 pg/ml) suggesting a partial response, usually associated with an adequate 11-deoxycortisol level (39). Until more data are available on the normal range in ACTH response to metyrapone and on the relevance of a reduced ACTH (but adequate 11-deoxycortisol) level, we would recommend clinical decision making be undertaken on the basis of the 11-deoxycortisol level. It has been shown that some patients with a normal ITT may have an abnormal metyrapone test; however, the clinical implication of this observation is unknown (32, 40). The finding of a normal metyrapone test can reassure the clinician and patient alike that glucocorticoid replacement is unnecessary.
An alternative test used in some units is the glucagon stimulation test (41, 42, 43). The recommended dose is 1 mg, by the im route (42); sc glucagon may be used, but responses are more variable (41, 44). There is a moderately high incidence of adverse side effects, with approximately 20–50% experiencing nausea and 10% vomiting (41, 42, 43, 44). The criteria for an adequate peak cortisol response, based on a large study in normal subjects, is more than 500 nM (18 µg/dl). Again, centers making use of this test on a regular basis are encouraged to establish their own reference range.
The CRH test (1 µg/kg or 100 µg CRH as an iv bolus) is a useful investigation in the differential diagnosis of Cushing’s syndrome (45) but has also been studied as a test of secondary adrenal insufficiency (46). Two patterns of response may be seen, depending on whether the defect is at the pituitary or hypothalamic level. Though the CRH test correlates well with the ITT in the setting of exogenous glucocorticoid therapy (47), its performance as a test of HPA axis function in the setting of pituitary disease is inferior to that of basal 0800 h cortisol (48). In the difficult group of patients with intermediate 0800 h cortisol levels (200–440 nM), it provides no further information (48), which is precisely the group where a definitive test is required. Therefore, we do not recommend its routine use in this setting.
When is the optimal time to perform the test?
Here the literature is unable to give any fixed time-point. There is some evidence that ACTH deficiency in the early postoperative period may recover by 3 months (15, 16). However, most studies that have examined the HPA axis by ITT or metyrapone testing between 7 and 14 d post surgery show a strong predictive value between these results and the need for long-term glucocorticoid requirement (2, 16). We recommend the measurement of an 0800 h plasma cortisol on d 7 with plans to perform the definitive test, if required, between d 10 and 14. This would ensure that all necessary assay results would be available for a routine outpatient review at 3–4 wk postoperatively. Nonsuppressive doses of hydrocortisone (where used) can be continued until the results are available.
Another alternative is to routinely perform a definitive dynamic test of pituitary function between 4 and 6 wk postoperatively (32). This may have the slight advantage of identifying those patients with later recovery of cortisol secretion. In units with an early discharge policy (2–3 d after pituitary surgery), patients treated perioperatively with glucocorticoids (and those with low 0800 h plasma cortisol levels on d 1 and 2) should receive physiological glucocorticoid replacement until later assessment of functional recovery. Physiological hydrocortisone replacement for this period of time after pituitary surgery does not suppress the HPA axis (49). Ultimately, any decision regarding the optimal time for provocative testing will depend on local practice. Because patients are often discharged within the first week after trans-sphenoidal surgery, and there may be geographical concerns, some centers may prefer assessment at 4–6 wk
Conclusions
The majority of patients undergoing selective trans-sphenoidal adenomectomy do not require routine glucocorticoid replacement in the perioperative period. Using the available literature, we have outlined a suggested protocol for investigation and management after pituitary surgery (Fig. 1). The adoption of such a protocol may reduce the unnecessary use of steroids without compromising the safety of our patients, while allowing for early clinical decision-making regarding the necessity of long-term glucocorticoid replacement. Many patients will be able to have their need for replacement glucocorticoids based on postoperative 0800 h plasma cortisol levels. Those requiring a definitive test should have an ITT, metyrapone or glucagon test, because the use of ACTH 1–24 is unreliable in the postoperative setting and the CRH test adds little to the basal cortisol level. Finally, we would encourage further research in this area, to better define optimal postoperative glucocorticoid management after pituitary surgery.
|
Acknowledgments
Footnotes
Abbreviations: HPA, Hypothalamic-pituitary-adrenal; ITT, insulin tolerance test.
Received June 29, 2002.
Accepted February 15, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  A. S. Dumont, E. C. Nemergut II, J. A. Jane Jr, and E. R. Laws Jr Postoperative Care Following Pituitary Surgery J Intensive Care Med, May 1, 2005; 20(3): 127 - 140. [Abstract] [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]
|
|
|
|
|
|
C. H. Courtney, A. S. McAllister, P. M. Bell, D. R. McCance, H. Leslie, B. Sheridan, and A. B. Atkinson Low- and Standard-Dose Corticotropin and Insulin Hypoglycemia Testing in the Assessment of Hypothalamic-Pituitary-Adrenal Function after Pituitary Surgery J. Clin. Endocrinol. Metab., April 1, 2004; 89(4): 1712 - 1717. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. A. F. S. Rollin, N. P. Ferreira, M. Junges, J. L. Gross, and M. A. Czepielewski Dynamics of Serum Cortisol Levels after Transsphenoidal Surgery in a Cohort of Patients with Cushing's Disease J. Clin. Endocrinol. Metab., March 1, 2004; 89(3): 1131 - 1139. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
, if selective adenomectomy is not possible, then treat as per the abnormal ACTH 1–24 test.