Content area
Objective: Glucocorticoid (GC) pharmacotherapy is an effective treatment for a range of diseases, but exposure can suppress the hypothalamic-pituitary-adrenal axis, leading to glucocorticoid-induced adrenal insufficiency (GC-AI) in some patients. However, the incidence of diagnosed GC-AI and the associated health burden, including the incidence of adrenal crises (ACs), are unknown. Although GC-AI treatment is based on well-established principles, there are no agreed protocols regarding the peri-operative management of exposed patients. The aims of this study were to assess the incidence of diagnosed GC-AI in hospital patients and review current approaches to peri-operative management of surgical patients with GC exposure.
Methods: An analysis of hospital admission data concerning adult patients diagnosed with GC-AI and a review of published recommendations for peri-operative GC cover.
Results: Between 2001 and 2013, admission with a diagnosis of GC-AI in New South Wales, Australia was rare (annual average of 22.5 admissions/year) and ACs were even more rare (n = 3). Almost two-thirds (64.4%, n = 188) of the patients with diagnosed GC-AI were aged between 50 and 79 years and 45.2% (n = 132) had a comorbid infection. The current approach to peri-operative management of patients with GC exposure appears to be influenced by both the absence of clear guidelines and historic practices. This results in the exposure of some patients to supraphysiologic doses of GCs during the peri-operative period.
Conclusion: Hospital admission with a diagnosis of GC-AI (with or without an AC) is very rare. Clear guidelines on peri-operative GC cover are necessary to avoid overreplacement with supraphysiologic doses in susceptible patients. (Endocr Pract. 2018;24:437-445)
ABSTRACT
Objective: Glucocorticoid (GC) pharmacotherapy is an effective treatment for a range of diseases, but exposure can suppress the hypothalamic-pituitary-adrenal axis, leading to glucocorticoid-induced adrenal insufficiency (GC-AI) in some patients. However, the incidence of diagnosed GC-AI and the associated health burden, including the incidence of adrenal crises (ACs), are unknown. Although GC-AI treatment is based on well-established principles, there are no agreed protocols regarding the peri-operative management of exposed patients. The aims of this study were to assess the incidence of diagnosed GC-AI in hospital patients and review current approaches to peri-operative management of surgical patients with GC exposure.
Methods: An analysis of hospital admission data concerning adult patients diagnosed with GC-AI and a review of published recommendations for peri-operative GC cover.
Results: Between 2001 and 2013, admission with a diagnosis of GC-AI in New South Wales, Australia was rare (annual average of 22.5 admissions/year) and ACs were even more rare (n = 3). Almost two-thirds (64.4%, n = 188) of the patients with diagnosed GC-AI were aged between 50 and 79 years and 45.2% (n = 132) had a comorbid infection. The current approach to peri-operative management of patients with GC exposure appears to be influenced by both the absence of clear guidelines and historic practices. This results in the exposure of some patients to supraphysiologic doses of GCs during the peri-operative period.
Conclusion: Hospital admission with a diagnosis of GC-AI (with or without an AC) is very rare. Clear guidelines on peri-operative GC cover are necessary to avoid overreplacement with supraphysiologic doses in susceptible patients. (Endocr Pract. 2018;24:437-445)
Abbreviations:
AC = adrenal crisis; ACTH = adrenocorticotropic hormone; AI = adrenal insufficiency; CI = confidence interval; GC = glucocorticoid; GC-AI = glucocorticoidinduced adrenal insufficiency; HPA = hypothalamicpituitary-adrenal; OR = odds ratio
INTRODUCTION
GC pharmacotherapy is a widely used and effective treatment for a range of inflammatory and neoplastic diseases. It is estimated that, when all forms of administration (oral, injected [intravenous/intramuscular/intraarticular], topical, nasal, or inhaled routes) are considered together, GC treatment is used by approximately 3% of the population at any point in time (1). However, GC administration at a systemic dose of 5 mg or more of prednisolone (or equivalent) for longer than 2 to 4 weeks can lead to the development of GC-AI due to feedback inhibition of the hypothalamic-pituitary-adrenal (HPA) axis (2). The extent to which suppression occurs varies between individuals and modes of GC administration (3) and many affected patients remain both asymptomatic and undiagnosed, particularly when they are receiving ongoing baseline GC dosing. Perioperative GC therapy for those with potential GC-AI has been prescribed since the 1950s to ensure adequate GC effect in the setting of peri-operative stress. This practice arose after reports of adrenal crises (ACs) in GC-exposed patients; supraphysiologic doses were recommended in an effort to prevent an AC in the belief that high GC doses caused little, if any, harm (4). This approach has persisted, despite increased knowledge about undesirable outcomes from high doses of peri-operative GC cover and demonstration of the safety and efficacy of more physiologically based regimens (4-7). Consequently, patients with an intact HPA axis are frequently given supraphysiologic peri-operative doses of GC, increasing the risks of hyperglycemia, hypokalemia, and neurocognitive activation. Although the likelihood of these adverse outcomes has not been quantified, hyperglycemia, in particular, is frequently encountered, especially in patients with comorbid diabetes mellitus.
There are a number of issues relating to the health consequences of GC exposure that remain unresolved. Importantly, there is no agreement on the role of clinical investigations for detection of GC-AI, nor are there consensus guidelines for the provision of appropriate GC cover to prevent an AC in surgical patients at risk of GC-AI. Hence, most patients are treated as though they have secondary AI. In addition, there are no trials comparing physiologic with supraphysiologic GC dosing in peri-operative patients. Surveys of selected samples of patients have increased the body of evidence on which the clinical management of GC-AI is based (2,3), but the development of practice guidelines is limited by the paucity of good quality data on the incidence of severe illness, especially AC due to GC-AI. The objective of this paper is to describe the current knowledge of the burden of morbidity from GC-AI in populations; to provide novel data on the epidemiology of medically diagnosed GC-AI in hospitalized patients; and to suggest approaches to the peri-operative management of patients receiving GC pharmacotherapy, with or without confirmed GC-AI.
Physiology
Exogenous GCs given in daily doses of 5 mg or more of prednisolone or its equivalent are sufficient to suppress the HPA axis. HPA suppression results from feedback inhibition of hypothalamic paraventricular corticotropinreleasing hormone (CRH), with secondary and/or superadded decreased adrenocorticotropic hormone (ACTH) production and eventual atrophy of the cortisol-producing zona fasciculata cells in the adrenal cortices (8). Although the degree to which HPA axis suppression occurs in patients is unpredictable, and susceptibility appears to vary between individuals (3), the likelihood of GC-AI developing is related to the duration of the GC pharmacotherapy, the dose, and its metabolism, often by the CYP3A4 enzyme (1,6). The development of GC-AI can be potentiated by concurrent use of pharmacotherapy that affects the CYP3A4 enzyme (1), and genetic factors may also play a role, such as polymorphisms in the central nervous system (CNS) drug efflux protein MDR-1, which may lead to different levels of CNS exposure to administered GC (9). Other factors may influence the peripheral inactivation of cortisol or prednisolone by 11 ß hydroxysteroid dehydrogenase enzyme (11 ß HSD-1), such as thyroid status.
GC-induced suppression of cortisol secretion is thought to be a central phenomenon (hypothalamic-pituitary) resulting initially from negative feedback at GC and mineralocorticoid receptors (10). Prolonged HPA axis suppression ensues when supraphysiologic doses of GCs are administered for at least 2 to 4 weeks; recovery follows in 2 to 18 months (11,12). Although the mechanism of prolonged suppression is not known, the rapid restoration of cortisol secretion after intravenous CRH infusion following curative hypophyseal surgery for pituitary Cushing disease suggests that either the hypothalamic CRH neuron or higher regulatory input is the locus of defective HPA axis activity after weaning of GC therapy (13). This finding is in keeping with hypothalamic-lesioned rodent data (8). Conventional testing for adrenal suppression, however, relies on the degree of zona fasciculata atrophy resulting from prolonged deficiency of circulating ACTH (14).
Chronically, most patients with GC-AI are asymptomatic, provided the GC dose is equal to or greater than that required for replacement (approximately 5 mg or more of prednisolone or its equivalent in adults). However, during periods of physiologic stress, patients receiving low doses of synthetic GCs are at risk of developing an AC, with symptoms and signs including hypotension, nausea, vomiting, abdominal pain, and delirium, together with hyponatremia and hyperkalemia (usually in primary AI) (15,16). Patients with AI often require hospital admission and those with an AC mostly require hospital admission, for intravenous fluids and parenteral hydrocortisone. Times of particular risk for development of AI/AC are abrupt withdrawal of GCs, which may be deliberate or inadvertent, or physiologic stress such as bacterial infection or injury that require increased GC levels in the circulation.
Diagnosis
There is no evidence-based definition of GC-AI. The measurement of morning cortisol, while interrupting the exogenous GC for a period longer than its tissue effect (e.g., 24 hours for prednisolone) can provide an index of endogenous HPA activity. Morning plasma cortisol levels of <34 nmol/L or above 340 nmol/L predict subnormal or normal responses to ACTH 1-34 (Synacthen®, Cosyntropin®) testing, respectively (1). Such cut-offs are assay dependent. Critical illness may, however, affect cortisol measurement via several factors including variation in serum corticosteroid-binding globulin concentrations and impaired cortisol metabolism (17,18). Although morning cortisol is used as a convenient screening tool, ACTH stimulation testing with its proven correlation to the less safe insulin hypoglycemia testing, is considered more definitive than a single morning cortisol assessment (3,19).
However, the extent to which subnormal cortisol levels after ACTH stimulation should guide the use of supplemental hydrocortisone during episodes of physiologic stress such as an acute illness or surgical procedure remains unclear. Despite this, recent evidence suggests that patients who are on supraphysiologic GC doses (>5 mg prednisolone or equivalent) with abnormal ACTH stimulation tests may be expected to remain well during stress when maintained on their existing dosage without any supplemental therapy (20,21).
Incidence, Morbidity, and Mortality
Given the relative rarity of structural diseases causing AI and the frequency with which GC therapy is used, GC-AI is likely the most common cause of AI in the population (1,2,22). It has been estimated that approximately half the patients taking oral GCs and one-quarter using inhaled GCs have adrenal suppression, according to ACTH stimulation tests (1). Individuals using GCs intermittently, such as those taking high-dose dexamethasone as part of chemotherapy regimens for nonsurgical emergencies may also be at risk of AI/AC events. However, the extent to which the widespread use of GC pharmacotherapy in its different formulations manifests in serious illness secondary to AI (symptomatic AI/AC), even among those who have suppression on ACTH testing, has not been extensively researched and is largely unknown. Clinical experience suggests that AC events in these patients are both rare and/or mild, probably reflecting incomplete HPA axis suppression in many patients, but the absence of substantial research in this area means that there are few studies to support or refute this assumption. Evidence supporting the likelihood of milder levels of clinically significant suppression is derived from a sample of renal transplant patients taking 5 to 10 mg prednisolone daily, in which one-third were found to have normal ACTH-stimulated cortisol levels, indicating lack of suppression and, therefore, a likely low risk of AC events (21). Studies on primates subjected to an adrenalectomy then cholecystectomy have indicated that physiologic cortisol levels are sufficient and produce equivalent survival to supraphysiologic cortisol levels, whereas cortisol replacement at onetenth of the basal level is associated with higher operative mortality (23).
METHODS
Epidemiologic Data from New South Wales, Australia
There are few available sources of information about the health burden from GC-AI in populations. Hospital admissions are one source of data on the incidence of diagnosed GC-AI in hospital patients. We evaluated the aggregated records of patients admitted to all hospitals in New South Wales, Australia between January 2001 and December 2013 where the principal or any secondary diagnosis was E27.3 (drug-induced AI), providing this code was paired with the rubric "Y42.0 - GCs adverse effects treatment use" (24). We compared the demographic and disease profiles of these patients with those from a random sample of admissions whose records did not include a diagnosis of any type of AI from the same database. The details of this dataset and the selection of the patient samples, which specifically excluded patients with a record of surgery, have been described previously (25).
RESULTS
This analysis demonstrated that hospital admission with either a principal or comorbid diagnosis of GC-AI was very uncommon. In a population of approximately 5 million adults (26), 165,000 of whom may be taking some form of GC, there were 345 admissions with a diagnosis of GC-AI. Of these, 45 (13.0%) admissions were for a malignancy and a further 8 had a principal diagnosis of a fracture. The remaining 292 medical admissions corresponded to an average annual rate of 22.5 admissions/year or, alternatively, an approximate incidence (diagnosis made among hospital patients) of 4.5/million/year. Three (1.0%) ACs were recorded, suggesting that the risk of inducing an AC in GC-AI patients is lower than previously thought (perhaps less than 0.01% annually), particularly in comparison to the AC risk in both primary and secondary AI (estimated to be 5 to 10 ACs/100 patient years) (15,16,27). A further 57 (19.5%) patients had a principal diagnosis of GC-AI, meaning that symptomatic AI due to GC-AI was the main reason for admission in a total of one-fifth (20.5%) of the patients.
Those patients treated in hospital for GC-AI were predominantly in the older age groups, with nearly twothirds (64.4%) aged between 50 and 79 years (Table 1). This patient group had a number of clinically significant comorbid conditions: 45.2% (n = 132) had an infection, gastroenteritis was identified in 10.6% (n = 31), pneumonia (lower respiratory tract infection) was a common comorbidity (n = 57, 19.5%), and over-one third (37.3%, n = 109) had a respiratory illness (Table 2). Patients with GC-AI were hospitalized for a longer period than patients in the non-AI group (Table 1). Among these patients, there were 6 (2.1%) deaths, but none were identified among patients with a record of an AC or a principal diagnosis of GC-AI. The in-hospital mortality rate for patients with GC-AI was comparable to that found in patients without AI (Table 2).
A number of multivariate analyses were conducted to examine whether the observed disparities between the GC-AI and control groups with regard to significant health outcomes (in-hospital mortality and infection) and the occurrence of significant related comorbid illnesses such as type 2 diabetes mellitus, osteoporosis, and cardiac disease may have been affected by the age differences between the 2 samples. Logistic regression models, which included age as a covariate, were used for this purpose. A significant Wald statistic for the addition of group into each model was used to determine the effect of group on each outcome.
With regard to in-hospital mortality, logistic regression showed that age was significantly associated with mortality but that the group effect (GC-AI or control) was not. By comparison, after adjustment for age, group was significantly associated with infection (Wald [1] = 30.1, P<.0001; odds ratio [OR] [95% confidence interval (CI)] for GC-AI: 2.0 [1.6, 2.5]); type 2 diabetes mellitus (Wald [1] =13.3, P<.0001; OR [95% CI] for GC-AI: 1.7 [1.3, 2.2]); and osteoporosis (Wald [1] = 3.9, P<.0001; OR [95% CI] for GC-AI: 34.8 [18.6, 65.1]). Cardiac disease showed a significant group effect (Wald [1] = 7.1, P<.01; OR [95% CI] for GC-AI: 1.5 [1.1, 2.0]), although age was found to have a stronger association (Wald [1] = 186.8, P<.0001) with this particular outcome. Gastroenteritis, which is often associated with admission among patients with AI, showed a significant association with group after adjusting for age (Wald [1] = 10.0, P<.01, OR [95% CI] for GC-AI: 1.9 [1.3, 2.9]). Further, a linear regression model was used to assess the effect of group on length of stay after adjustment for age. This model demonstrated that group was a significant predictor of an increased length of stay (t = 18.7, P<.0001).
Overall, these data suggest that, given the low incidence of diagnosed GC-AI in hospitalized patients, serious illness occurring in association with, or as a direct result of, GC-AI is less common than was previously thought. This is consistent with the impression among clinicians that while GC-AI may occur quite commonly among patients with exposure to GC pharmacotherapy, the clinical effect, with regards to AI-related illness may not be as severe as had been assumed. In contrast, these data do provide some evidence for the increased risk of adverse sequelae such as type 2 diabetes mellitus and osteoporosis among patients following GC exposure.
There are, however, a number of caveats to these results. The most important of these is that only those patients whose record included a coded diagnosis of GC-AI were selected into the study sample. It is likely that there were a greater number of patients with undiagnosed adrenal suppression following GC exposure admitted during this time but this was either not identified as a significant comorbid condition or was not noted in the medical record or coded by health information staff. It is also possible that some patients were miscoded as having other forms of AI. Conversely, it is possible that some AC/ AI events are overdiagnosed in patients with intervening illnesses producing similar clinical features. These limitations notwithstanding, the data provide some support for the widely held view that the health impact of GC-AI may have been overestimated.
Medical Management
Established HPA axis suppression can only be managed by progressive, stepwise withdrawal of exogenous GC. Risks of GC withdrawal include: (1) recrudescence of the initially treated inflammatory disorder; (2) hormone withdrawal syndrome, including in those dose ranges above the physiologic range; and (3) adrenal insufficiency at doses below physiologic replacement levels. Generally, provided features of (1) and (2) are not present, GC doses can be weaned rapidly from supraphysiologic to physiologic doses. Since recovery of endogenous HPA axis function may take 6 to 18 months, dose reductions of only 1 mg prednisolone (or equivalent) per month are often necessary. Recovery can be monitored, usually 2 to 3 times monthly, using morning serum cortisol, with higher or normal values suggesting recovery of endogenous cortisol secretion and the capacity for more rapid weaning of the exogenous GC.
HPA axis suppression is more likely with higher doses of high potency, long-acting GCs. Suppression risk may be reduced if GCs can be given on an alternate day basis, but for many diseases a clear difference in effectiveness between alternate day and daily GC (usually prednisolone/prednisone) is not established. However, HPA axis suppression risk is lower with alternate-day dosing or if patients can be converted to such dosing before weaning (28,29).
GC-AI, Surgery, and the Peri-Operative Period
A review of publications on the peri-operative management of patients with GC-AI or GC exposure in PubMed, MEDLINE, and Cochrane was conducted. There is a paucity of evidence on the best management of such patients undergoing a surgical procedure and an absence of relevant clinical guidelines. Not surprisingly, the treatment of peri-operative patients has been shown to vary considerably between clinicians, even among those from the same institution (30). Such variability has its origins in the interplay between the response to the early reports of perioperative AC events in patients with GC exposure (31-34) and the absence of high-quality studies in the ensuing years on which management decisions could be soundly based (34-36). The relevant studies are listed in Table 3. This problem is further compounded by the belief that too much GC is preferable to too little, and for most patients this is true, but excessive GC dosing may induce hyperglycemia, hypokalemia, and neurocognitive activation, which may hamper the recovery process, particularly among the elderly and those with diabetes mellitus. Finally, minor perturbations such as postoperative nausea or postural dizziness may, in the case of GC-treated patients, be perceived as being symptoms of AI, reinforcing the concept of a need for supraphysiologic high-dose GC cover in the peri-operative period. The nonspecific antinausea effects of exogenous GCs may also create an impression of the existence of frequent episodes of symptomatic AI or ACs in this patient group.
DISCUSSION
The first account of a patient presumed to have had a peri-operative AC appeared in 1952 (4,31). This patient had a lengthy course of cortisone (25 mg twice daily) for rheumatoid arthritis, which was ceased 2 days prior to surgery, after which the patient experienced cardiovascular collapse (31). Another report the following year described a patient in similar circumstances, who had ceased cortisone treatment 1 day before knee surgery (32). However, it was not until 1961 that the first confirmed episode of surgery-related cardiovascular collapse, or AC due to GC-AI, was described (36,40). Indeed, in many of the cases reported between the 1950s and early 1970s the diagnosis of acute AI/AC was not proven, with the outcome found to be potentially attributable to other factors (4,36,40). This weaker than expected association between GC exposure and peri-operative acute AI/AC was reinforced by the results of contemporaneous cohort studies (41,42), which demonstrated that ACs were very uncommon, occurring in less than 1% of the total patients considered to be at risk (4).
Expert guidance on peri-operative GC coverage at the time, however, reflected a belief in a greater likelihood of peri-operative AC than the evidence indicated, recommending large supplementary, supraphysiologic regimens (typically equivalent to a fourfold increase in the usual dose) given for periods of up to 2 weeks postoperatively (33). These were well in excess of the cortisol production rates in normal subjects (4,43) and were at least partially influenced by inaccurately high estimations of these rates at that time (4,43). The recommended doses were also in excess of the estimated increase in plasma cortisol concentrations in response to surgery, which were found to peak 4 to 6 hours post-procedure, returning to baseline after 24 hours, with the exception of major surgery where the elevated levels could be sustained for 48 to 72 hours (44,45). Limitations in the methods used to assess HPA function in susceptible patients at that time compounded the difficulties in identifying patients at risk of a peri-operative AC (30).
This approach to the peri-operative management of patients with GC exposure continued for many years, although there was minimal modification to these depending on the level and duration of GC exposure. The broad definition of "at-risk" levels of GC exposure remained. However, evidence from further studies conducted during this time showed that patients with GC-AI according to ACTH tests could withstand the physiologic stress of a surgical procedure without the need for supplementary cover, providing their usual GC dose was maintained (4,22,46). A 1994 review by Salem et al (4) concluded that continuation of the usual dose of GC at the time of surgery was sufficient to prevent acute hypoadrenalism during the peri-operative period in patients on exogenous GCs and, that where supplementation was desirable, it should only be given in doses equivalent to the cortisol response in a nonglucocorticoid-exposed patient. These authors also proposed that that the administration of supplementary treatment should be commensurate with the extent and severity of the surgical procedure.
Later in the same decade, Nicholson et al (45) revisited the apparent paradox between the widespread use of peri-operative supplementary GCs in patients with GC exposure and the evidence establishing the need for such treatment. They noted, in particular, that the practice had not abated despite an increased weight of evidence supporting a move away from supplementary steroids and observed that this persistence may also be influenced by a widespread misconception about the innocuous nature of GC supplementation for surgical procedures. They further added to the concept of stratification of the level of surgical stress by providing a more nuanced grading system, ranging from "minor" operations that cause little or no stress to major surgery such as cardiovascular bypass grafting (45).
The conundrum of persistently high use of supplementary GCs in the surgical setting was re-examined in 2 systematic reviews approximately 10 years later (34,35). Marik and Varon (2008) (34) reviewed the extant literature consisting of 7 cohort studies and 2 randomized trials on 315 patients who underwent 389 surgical operations (34). They found no difference in the incidence of hypotension or an AC between patients who received supplemental GCs and those who were maintained on usual therapy without supplementation (34). They concluded that stress doses of GCs were not required for surgical procedures, providing the usual therapeutic dose was maintained. They also recommended against the routine use of ACTH testing in this context, due to the inability of this test to identify patients who may develop an AC (34). By comparison, Yong et al (2012) (35) examined the same randomized trials (38,39) and concluded that the studies did not demonstrate a difference in adverse outcomes due to peri-operative hypo-adrenalism between patients given and not given supplemental therapy. However, they added some caveats including the small number of studies and cases, and that in one of the investigations, patients had been operated on under local anesthetic (39). In summary, these authors concluded that there was insufficient evidence on which to base recommendations on whether additional GCs are required at the time of surgery (35).
Current Situation
Although there is evidence supporting the benefits and appropriateness of more physiologically based GC supplementation, the approach to peri-operative GC dosing remains highly variable (30). Typically, in most institutions, peri-operative GC cover is given on an ad hoc basis, although some institutions have local protocols. These dosing schedules can be divided into (1) supraphysiologic dosing, employing doses that will achieve cortisol levels well beyond those seen in patients with intact adrenal function, often for prolonged periods (>48 h); (2) physiologic dosing where the GC dose is likely to produce cortisol levels comparable to normal physiology, and given for <48 h; and (3) the inevitable situation where supplemental GCs are not given, either with continued or discontinued chronic dosing. Recommendations for peri-operative GC administration according to the type of surgical procedure are given in Table 4.
CONCLUSION
Despite the high prevalence of GC use in the community, with quite well documented rates of adrenal suppression, our data and those from other studies suggest a low risk of AC in these patients. These data provide some evidence to support the view that GC-AI entails a low risk of AC events. Given that GC dosing in these patients ranges from high dose to physiologic dose to no stress supplement, it appears that few AC events result overall. There is, however, clear consensus that oral GC therapy must not be stopped abruptly, especially during the peri-operative period. The paucity of such events suggests that physiologic stress-dose replacement is likely to prevent AC and minimize harm from excessive GC doses peri-operatively. Confirmation of the safety of peri-operative physiologic stress dose GC should be demonstrated in a prospective clinical study.
ACKNOWLEDGMENT
We than the NSW Ministry of Health for access to admission data.
DISCLOSURE
The authors have no multiplicity of interest to disclose.
REFERENCES
1. Woods CP, Argese N, Chapman M, et al. Adrenal suppression in patients taking inhaled glucocorticoids is highly prevalent and management can be guided by morning cortisol. Eur J Endocrinol. 2015;173:633-642.
2. Hahner S, Allolio B. Management of adrenal insufficiency in different clinical settings. Expert Opin Pharmacother. 2005;6: 2407-2417.
3. Henzen C, Suter A, Lerch E, Urbinelli R, Schorno XH, Briner VA. Suppression and recovery of adrenal response after short-term, high-dose glucocorticoid treatment. Lancet. 2000;355: 542-545.
4. Salem M, Tainsh RE Jr, Bromberg J, Loriaux DL, Chernow B. Perioperative glucocorticoid coverage. A reassessment 42 years after emergence of a problem. Ann Surg. 1994;219:416-425.
5. Jung C, Inder WJ. Management of adrenal insufficiency during the stress of medical illness and surgery. Med J Aust. 2008;188: 409-413.
6. Broersen LH, Pereira AM, Jørgensen JO, Deckers OM. Adrenal Insufficiency in Corticosteroids Use: Systematic Review and Meta-Analysis. J Clin Endocrinol Metab. 2015;100: 2171-2180.
7. de Lange DW, Kars M. Perioperative glucocorticosteroid supplementation is not supported by evidence. Eur J Intern Med. 2008;19:461-467.
8. Levin N, Shinsako J, Dallman MF. Corticosterone acts on the brain to inhibit adrenalectomy-induced adrenocorticotropin secretion. Endocrinology. 1988;122:694-701. 25.
9. Karssen AM, Meijer OC, van der Sandt CJ, et al. Multidrug resistance P-glycoprotein hampers the access of cortisol but not of corticosterone to mouse and human brain. Endocrinology. 26. 2001;142:2686-2694.
10. Reul JHM, de Kloet ER. Two receptor systems for corticoste- 27. rone in rat brain: microdistribution and differential occupation. Endocrinology. 1985;117:2505-2511.
11. Jamilloux Y, Lioxon E, Pugnet G, et al. Recovery of adrenal function after long-term glucocorticoid therapy for giant cell arte- 28. ritis: a cohort study. PLoS One. 2013;8:e68713.
12. Einaudi S, Bertorello N, Masera N, et al. Adrenal axis function 29. after high-dose steroid therapy for childhood acute lymphoblastic leukemia. Pediatr Blood Cancer. 2008;50:537-541.
13. Gomez MT, Magiakou MA, Mastorakos G, Chrousos GP. The 30. pituitary corticotroph is not the rate limiting step in the postoperative recovery of the hypothalamic-pituitary-adrenal axis in patients with Cushing's syndrome. J Clin Endocrinol Metab. 1993 ;77: 173-177. 31.
14. Bornstein SR, Allolio B, Arlt W, et al. Diagnosis and Treatment of Primary Adrenal Insufficiency: An Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab. 2016;101:364-389. 32.
15. Rushworth RL, Torpy DJ, Falhammar H. Adrenal crises: perspectives and research directions. Endocrine. 2017;55:336-345.
16. Allolio B. Extensive Experience in Endocrinology: Adrenal Crisis. 33. Eur J Endocrinol. 2015;172:R115-R124.
17. Boonen E, Vervenne H, Meersseman P, et al. Reduced corti- 34. sol metabolism during critical illness. N Engl J Med. 2013;368: 1477-1488.
18. Nenke MA, Rankin W, Chapman MJ, et al. Depletion of high- 35. affinity corticosteroid-binding globulin corresponds to illness severity in sepsis and septic shock; clinical implications. Clin Endocrinol (Oxf). 2015;82:801-807. 36.
19. Stewart PM, Corrie J, Seckl JR, Edwards CR, Padfield PL. A rational approach for assessing the hypothalamo-pituitary-adrenal 37. axis. Lancet. 1988;1:1208-1210.
20. Bromberg JS, Alfrey EJ, Barker CF, et al. Adrenal suppression and steroid supplementation in renal transplant recipients. Transplantation. 1991;31:385-390.
21. Bromberg JS, Baliga P, Cofer JB, Rajagopalan PR, Friedman RJ. Stress steroids are not required for patients receiving a renal allograft and undergoing operation. J Am Coll Surg. 1995;180: 532-536.
22. Crowley RK, Argese N, Tomlinson JW, Stewart PM. Central hypoadrenalism. J Clin Endocrinol Metab. 2014; 99:4027-4036.
23. Udelsman R, Ramp J, Gallucci WT, et al. Adaptation during surgical stress. A reevaluation of the role of glucocorticoids. J Clin Invest. 1986;77:1377-1381.
24. National Centre for Classification in Health, 2006. National Centre for Classification in Health. International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Australian Modification (ICD-10-AM), Australian Classification of Health Interventions (ACHI) and Australian Coding Standards (ACS) (Fifth ed.). National Centre for Classification in Health, Sydney; 2006.
25. Rushworth RL, Torpy DJ. A descriptive study of adrenal crises in adults with adrenal insufficiency: increased risk with age and in those with bacterial infections. BMC Endocr Disord. 2014;14:79.
26. http://www.healthstats.nsw.gov. au/Indicator/dem_pop_age/dem_ pop_age
27. Hahner S, Spinnler C, Fassnacht M, et al. High incidence of adrenal crisis in educated patients with chronic adrenal insufficiency: a prospective study. J Clin Endocrinol Metab. 2015;100: 407-416.
28. Krasner AS. Glucocorticoid-induced adrenal insufficiency. JAMA. 1999;18:282:671-676.
29. Ackerman GL, Nolsn CM. Adrenocortical responsiveness after alternate-day corticosteroid therapy. N Engl J Med. 1968;22:278:405-409.
30. Lamore RF 3rd, Hechenbleikner EM, Ha C, et al. Perioperative glucocorticoid prescribing habits in patients with inflammatory bowel disease: a call for standardization. JAMA Surg. 2014;149:459-466.
31. Fraser CG, Preuss FS, Bigford WD. Adrenal atrophy and irreversible shock associated with cortisone therapy. J Am Med Assoc. 1952:23;149:1542-1543.
32. Lewis L, Robinson RF, Yee J, Hacker LA, Eisen G. Fatal adrenal cortical insufficiency precipitated by surgery during prolonged continuous cortisone treatment. Ann Intern Med. 1953;39:116-126.
33. Bayliss RI. Surgical collapse during and after corticosteroid therapy. Br Med J. 1958;2:935-936.
34. Marik PE, Varon J. Requirement of perioperative stress doses of corticosteroids: a systematic review of the literature. Arch Surg. 2008;143:1222-1226.
35. Yong SL, Coulthard P, Wrzosek A. Supplemental perioperative steroids for surgical patients with adrenal insufficiency. Cochrane Database Syst Rev. 2012;12:CD005367.
36. Sampson PA, Brooke BN, Winstone NE. Biochemical confirmation of collapse due to adrenal failure. Lancet. 1961;1:1377.
37. Friedman RJ, Schiff CF, Bromberg JS. Use of supplemental steroids in patients having orthopaedic operations. J Bone Joint Surg Am. 1995;77:1801-1806.
38. Glowniak JV, Loriaux DL. A double-blind study of perioperative 42. steroid requirements in secondary adrenal insufficiency. Surgery. 1997;121:123-129.
39. Thomason JM, Girdler NM, Kendall-Taylor P, Wastell H, 43. Weddel A, Seymour RA. An investigation into the need for supplementary steroids in organ transplant patients undergoing gingival surgery. A double-blind, split-mouth, cross-over study. J 44. Clin Periodontol. 1999;26:577-582.
40. Cope CL. The adrenal cortex in internal medicine. I. Br Med J. 1966;8;2:847-853 . 45.
41. Plumpton FS, Besser GM, Cole PV. Corticosteroid treatment and surgery. 1. An investigation of the indications for steroid cover. 46. Anaesthesia. 1969;24:3-11.
42. Kehlet H. Clinical Course and Hypothalamic-Pituitary - Adrenocortical function in glucocorticoid-treated surgical patients. Copenhagen: FADL's Forlag, 1976.
43. Mohler JL, Michael KA, Freedman AM, McRoberts JW, Griffen WO Jr. The evaluation of postoperative function of the adrenal gland. Surg Gynecol Obstet. 1985;161:551-556.
44. Alford WC Jr, Meador CK, Mihalevich J, et al. Acute adrenal insufficiency following cardiac surgical procedures. J Thorac Cardiovasc Surg. 1979;78:489-493.
45. Nicholson GI, Burrin JM, Hall GM. Peri-operative steroid supplementation. Anaesthesia. 1998;53:1091-1104.
46. Thorén L. General metabolic response to trauma including pain influence. Acta Anaesthesiol Scand Suppl. 1974;55:9-14.
Copyright Allen Press Publishing Services May 2018