The first report of the use of glucocorticoid hormone drugs in humans dates back to 1948, when Philip S. Hench used "substance E" (cortisone) to treat a 29-year-old woman with rheumatoid arthritis. Since the effect of the therapy was clearly pronounced, the drug was soon used in other patients with rheumatoid arthritis, as well as in patients with other diseases such as ankylosing spondylitis, psoriatic arthritis, and systemic lupus erythematosus [1]. By 1960, 7 more glucocorticoids had been discovered and used in clinical practice. Furthermore, their main side effects were described, among which adrenocortical insufficiency (ACI), revealed after drug withdrawal, was the most discussed. In the 1970s, several schemes were proposed for gradual withdrawal of glucocorticoids, which came down to either infrequent drug intake (less than once a day) and their use in the early morning [2], or to dose reduction (for example, by 5% monthly) [1].

Currently, systemic glucocorticoids are taken by 1 to 3% of people in the general population, including up to 1.8% for a long time [3]. At the same time, about 50% of patients taking these drugs orally have glucocorticoid-induced adrenocortical insufficiency (GIACI) after their withdrawal [4]. Despite the scale of the problem and the length of time it has been studied, there are currently no standard algorithms for reducing the dose or discontinuing glucocorticoids. The evidence base on this issue is formed by relatively few clinical studies, which are extremely heterogeneous in terms of the populations studied, designs, glucocorticoid regimens used, and diagnostic approaches to GIACI. Thus, the strength of evidence on this issue remains low; as a consequence, each healthcare institution has its own approaches to dose tapering and discontinuing glucocorticoids, or none at all. This paper summarizes the current knowledge on GIACI, aimed at improving approaches to tapering or discontinuing glucocorticoids.

A meta-analysis by Broersen et al., which included 3753 patients, showed a high prevalence of GIACI, which is diagnosed due to biochemical investigations in a laboratory, including functional tests. GIACI was most often detected in individuals who received glucocorticoids orally (48.7%), intra-articularly (52.2%), or by a mixed pathway (42.7%); an increased risk of its development was also noted upon inhalation (7.8%), intranasal (4.2%) and topical administration (4.7%) [4]. The highest risk of ACI was revealed in patients with hemato-oncological diseases (60.0%), after kidney transplantation (56.2%), as well as with nasal polyps (46.2%), cystic fibrosis (49.0%), Crohn's disease (52.2%), and rheumatoid arthritis/osteoarthritis (39.4%) [4][5]. A low risk of ACI was associated with low doses of drugs and a short duration of use [4].

It is worth noting that the prevalence of GIACI decreases over time after discontinuation of glucocorticoids but in some patients, it may persist for a long time. Thus, 40% of people may have GIACI 6 months after discontinuation of glucocorticoids, 20% after 2 years, and 5% after 3 years [6]. According to the study by Draoui et al., the median time to recovery of adrenal function is 12 months [7].

The secretion of cortisol depends on the functional state of the hypothalamus-pituitary-adrenal (HPA) system. Corticotropin-releasing hormone (CRH), otherwise stated as corticoliberin, which is produced in the hypothalamus, stimulates the synthesis and secretion of adrenocorticotropic hormone (ACTH) in the corticotrophs of the anterior pituitary, which leads to the release of cortisol from the cells of the zona fasciculata and androgens from the cells of the zona reticularis of the adrenal cortex. Cortisol is a powerful factor suppressing the work of the hypothalamus pituitary system, which is the basis of negative feedback. Excess cortisol has an inhibitory effect on the release of CRH in the hypothalamus and ACTH in the pituitary.

Normally, cortisol secretion is characterized by high variability depending on the time of day and the body's needs, namely the hormone level increases in the early morning hours, during stress, hunger, and a number of other conditions.

Taking glucocorticoids entails the inhibition of the synthesis and secretion of ACTH and corticoliberin. Long-term use of the drugs leads to atrophy of the fascicular and reticular zones of the adrenal cortex, as well as corticotrophs in the pituitary. As a result, after the withdrawal of glucocorticoids, the HPA cortex system may be unable to provide a sufficient level of glucocorticoid hormones for the body promoting ACI development [8]. Of interest are the data received by Gomez et al. on the dynamic examination of patients with endogenous hypercortisolism after surgical treatment. The study participants underwent tests with the introduction of sheep corticoliberin. Researchers found that corticotrophs retained sensitivity to the introduced hormone indicating that ACI is of a tertiary nature, i.e. it develops as a consequence of injury to neurons of the hypothalamus or at a higher level. Concurrently, the reaction in terms of cortisol secretion was worse than in the control group, attesting to atrophy of the adrenal cortex [9]. However, considering that there is no consensus among researchers on the mechanism of ACI development after glucocorticoid withdrawal, as well as an insufficient study of the pathogenesis of this condition, the most correct term at present seems to be "GIACI".

The process of restoring the normal activity of the HPA system after discontinuation of glucocorticoids cannot be considered well studied either. Apparently, the restoration of corticoliberin, ACTH, and cortisol levels occurs sequentially one after another [3][9].

The duration of glucocorticoid use before the development of GIACI is individual. It is recognized that the risk is low in the case of taking the drugs for less than 3 weeks but there are cases when hypocortisolism was detected in the functional biochemical tests after just one week of glucocorticoid administration [10].

It should be noted that after taking glucocorticoids at supraphysiological doses for several months, as well as after surgery for ACTH-independent Cushing's syndrome, the degree of suppression of the HPA system often does not correspond to clinical symptoms since symptoms of ACI may be pronounced against the background of slightly reduced or even normal blood cortisol values. A condition called glucocorticoid withdrawal syndrome (GWS) was described; its pathogenesis is associated not with suppressed adrenal function but with low sensitivity of peripheral tissues to glucocorticoids, which develops as a result of long-term hypercortisolism. GWS is diagnosed in the presence of symptoms of adrenal cortex insufficiency in the absence of biochemical test signs of ACI. Along with that, GWS may have its own clinical features. In particular, mood and cognitive disorders, hypersomnia, muscle weakness, myalgia, and fatigue may more often be manifested upon GWS than ACI, while other symptoms such as arterial hypotension, vomiting, and hypoglycemia are not typical for GWS [11]. It should be noted that this condition is widespread. For instance, according to Hurtado et al., GWS is detected in 67% of patients after adrenalectomy for Cushing's syndrome [12].

Clinical symptoms of chronic GIACI are nonspecific and include malaise, weakness, fatigue, dizziness, gastrointestinal symptoms (nausea, vomiting, diarrhea, constipation, loss of appetite), weight loss, hypotension, headache, arthralgia, myalgia, pale skin, hypoglycemia, frequent respiratory infections, and others. After discontinuation of glucocorticoids, a number of patients develop acute ACI, which is a life-threatening condition. It is accompanied by hypotension or hypovolemic shock, severe weakness, fever, and impaired consciousness (up to coma); in addition, hypoglycemia, hyponatremia, lymphocytosis, and eosinophilia can be detected by clinical biochemical tests [3].

The results of the Broersen meta-analysis indicate that out of 98 patients with ACI, clinical symptoms were detected in only 10 patients [4]. Considering that the clinical signs of hypocortisolism are nonspecific and often difficult to distinguish from the symptoms of concomitant diseases, it cannot be excluded that the negative results of the Broersen analysis are due to the fact that the symptoms of ACI in the studies included in the meta-analysis were not properly assessed and therefore were not recorded by the researchers.

Of interest are the data of the meta-analysis by Mebrahtu, which included data from 70,638 patients with chronic inflammatory diseases such as inflammatory bowel disease, systemic lupus erythematosus, rheumatoid arthritis, etc., who took glucocorticoid hormones orally. The development of GIACI was detected in 183 patients taking glucocorticoids [13]. Among the 74 deceased patients with GIACI, only 13 had acute ACI leading to death, while the most common cause of death was infections (44.6%). It can be assumed that some patients who had infection as the cause of death actually died from acute ACI [13]. These causes of death may be confused in routine clinical practice considering that chronic cortisol deficiency often becomes acute in patients with severe infection and that the clinical picture of acute ACI does not have pathognomonic features.

GWS differs from ACI in its clinical picture. In particular, according to He et al., mood disorders and cognitive impairment, anorexia, and general physical health deterioration are more characteristic of GWS, while arterial hypotension, hypoglycemia, and vomiting are more characteristic of primary ACI [11]. A number of patients with GWS note deterioration in well-being upon reducing the drug doses, despite the fact that they remain supraphysiological. There is no risk of acute ACI in this condition [14].

The laboratory diagnostics of GIACI are performed on the basis of functional tests including insulin, ACTH, and corticoliberin analogs, as well as metyrapone. Tests are conducted after reaching a daily dose of glucocorticoids of 5 mg or less, calculated as the equivalent amount of prednisolone, and switching to short-acting drugs. Besides, hydrocortisone is discontinued 24 hours before the test [15].

The most commonly used test in the world is the ACTH analog test. More specifically, the serum cortisol level is measured before and 30 and 60 minutes after intravenous injection of 250 μg of the drug. An increase in cortisol of more than 18–20 μg/dl (500–550 nmol/l), depending on the laboratory method used for the hormone quantification, enables one to exclude disorders at the level of the adrenal cortex. In the case of pituitary or hypothalamus function failure, for example, in the early stages of GIACI, false negative results may be recorded due to low test sensitivity [16]. The accuracy of the ACTH test in patients with GIACI is unknown. Specific threshold values for diagnosing the condition depend on the method used to determine blood cortisol, as well as on the target indicators of sensitivity and specificity. Upon using modern laboratory diagnostic methods, the threshold for the cortisol level can be set at 420 nmol/l. In the conditions of Russian clinical practice, the test is not performed due to the lack of ACTH preparations in the country. Several studies showed that a simple measurement of the cortisol level in the morning (using a threshold of, for example, 270 nmol/L) could have an equal diagnostic value in identifying GIACI [12][17]. The disadvantage of cortisol determination is the variability of the indicator due to the secretion rhythm, a multitude of concomitant factors affecting cortisolemia, and the need to take into account the method of laboratory testing of cortisol when choosing a threshold value for establishing GIACI [3].

The gold standard for diagnosing central (secondary) ACI is the insulin test. The patient is given intravenous insulin to induce hypoglycemia, which stimulates the secretion of corticoliberin in the hypothalamus and ACTH in the pituitary. This test is rarely used in clinical practice due to its labor-consuming nature, danger to patients, a number of contraindications, and the dependence of the results on the degree of hypoglycemia achieved. In this regard, the ACTH test is more preferable as it has a satisfactory diagnostic value, greater safety, and ease of implementation [18].

When it is not possible to conduct a hypoglycemic test, some authors recommend an overnight test with metyrapone. Metyrapone suppresses the hydroxylation of 11-deoxycortisol, blocking cortisol production. With normal pituitary function, a decrease in the concentration of cortisol compensatorily promotes the production of ACTH, which through stimulating steroidogenesis leads to the accumulation of 11-deoxycortisol. Ultimately, if the low level of 11-deoxycortisol remains, the test diagnoses ACTH deficiency. According to Baker, the metyrapone test is not inferior in accuracy to the insulin test and has certain advantages over the ACTH test [19]. This test is limited by the availability of metyrapone and the specificity of the 11-deoxycortisol blood test used in practice.

The corticoliberin test is used very rarely due to the unavailability of the drug and the low sensitivity of the test.

Plasma ACTH determination has little value for diagnosing GIACI, since its level may be either low or normal under physiological conditions. However, a significantly elevated ACTH level (>100 pg/ml) indicates primary ACI confirmed in the laboratory [19].

The main measures for the prevention of ACI upon using systemic glucocorticoids are the following:

However, following these rules is not always possible in clinical practice. Many patients must receive glucocorticoids for a long time and in fairly high doses. In these cases, the main measure for preventing GIACI is the gradual withdrawal of the drugs. Before reducing the dose, patients or their relatives should be informed about the symptoms of hypocortisolism, and the need to immediately contact a doctor in case of clinical manifestations of ACI, as well as about an increase in the glucocorticoid dose to the previous or higher levels.

Currently, there is no single scheme for reducing the dose of glucocorticoids. When making such a decision, it is necessary to take into account the individual risk of ACI, the risk of exacerbation of the disease, for which the drugs were used, previous experience of their withdrawal, as well as the possibility of transferring the patient to short-acting drugs (hydrocortisone, prednisolone), and reducing the frequency of drug administration.

Upon taking glucocorticoids more than 60 mg/day in terms of the equivalent amount of prednisolone, it is possible to reduce the dose by 10–20 mg once every 1–2 weeks [10]. In patients receiving 20–40 mg/day, the dose can be reduced by 5–10 mg weekly until reaching 20 mg/day. Afterwards, a reduction by 1.0–2.5 mg once every 1–4 weeks is recommended until complete withdrawal [3]. Besides, schemes are proposed for transferring patients to alternating regimens of administration after reaching a low daily dose of drugs, in particular by doubling the dose on one day and not taking it on the next [10].

Functional biochemical investigations can help predict the risk of developing GIACI in patients under laboratory conditions. For example, Prete et al. suggest performing a blood test for cortisol after reaching a dose of 5 mg/day in terms of the equivalent amount of prednisolone. If the hormone level is <100 nmol/l, GIACI is diagnosed, at a level of 100–270 nmol/l, the possibility of developing this condition is recorded, at 270–350 nmol/l, the risk under basic conditions (without stress factors) is considered small, at >350 nmol/l, it is unlikely. Based on the assessment of the GIACI risk, modifications to the glucocorticoid reduction scheme are considered possible [3]. Patients with a significant risk of hypocortisolism are recommended to assess the cortisol level every few months.

To date, there are no reliable markers of the GIACI risk. A promising method for determining the GIACI risk is the assessment of the indicators of adrenal cortex steroidogenesis in the blood or in daily urine [13]. Some studies indicated the advisability of studying CD16 (cluster of differentiation 16, expressed by cells of lymphoid origin) and ADAM17 (a metalloproteinase responsible for shedding CD16 from the surface of natural killer cells) in the blood, since their levels were found to be elevated in patients with ACI [14]. ACI [14]. The study by Draoui et al. revealed a positive correlation between morning serum cortisol levels and body mass index in patients after glucocorticoid withdrawal [7].

In the absence of normalization of the HPA axis function, the tactics of patient management should be aimed at continuous replacement therapy with glucocorticoids, as in the case of chronic ACI. Under basic conditions, a physiological dose of glucocorticoids can be recommended. About 6–11 mg/m2 of cortisol (hydrocortisone) are synthesized in the body per day, which corresponds to approximately 15–20 mg of the drug. The drug is prescribed in 2–3 doses, and 2/3 of the dose should be taken in the morning to simulate the circadian rhythm of cortisol secretion. Meanwhile, mineralocorticoid replacement therapy is not required. Currently, drugs with modified release of hydrocortisone are being developed, which will allow its constant physiological level in the blood to be maintained [15]. In the future, glucocorticoid drugs with tissue-specific action may be created that will not cause, for example, osteopenia and Cushing's syndrome [16–18]. People with ACI should be informed about the rules for augmenting the dose of glucocorticoids in case of increased demand for them, for example, during infections, injuries, severe concomitant diseases, before surgery, and under stress.

In case of development of acute ACI, immediate initiation of therapy is required, which includes correction of hypotension and water-electrolyte disturbances through parenteral administration of hydrocortisone (100 mg as an initial bolus, then 50–100 mg/6 h).

The basis of GWS treatment is the selection of the lowest dose of the drug to relieve the symptoms of the disease with its subsequent reduction over months. An increase in the dose of drugs is not required in case of stressful events and concomitant diseases [19].

GIACI has been well known to clinicians worldwide for many decades but it should be noted that this condition has not been sufficiently studied. Its prevalence and risk factors have not been fully investigated to date. Data on its pathogenesis are obtained from individual small-scale scientific studies. To diagnose ACI, clinicians use standard methods; in such circumstances, there is a demand to clearly differentiate between GIACI and GWS, in particular at an early stage, on the basis of symptoms, time, and conditions of their development. This problem requires further research and is of great clinical importance. Moreover, there is no generally accepted clear algorithm for GIACI prevention by gradually reducing glucocorticoid doses based on the principles of evidence-based medicine. Despite the fact that little attention has traditionally been paid to these issues, their importance is beyond doubt.

Patients treated with glucocorticoids for non-endocrine diseases are at risk of developing a life-threatening condition called acute ACI. Long-term administration of glucocorticoids results in decreased function of the HPA axis, the failure of which is most often detected after surgery, trauma, infectious diseases, and a rapid reduction of the glucocorticoid dose. On the other hand, excessive caution upon reducing the dose of glucocorticoids can lead to the progression of GIACI, as well as to the development of iatrogenic Cushing's syndrome. The presented facts substantiate the relevance of further study of GIACI and GWS.

Currently, there are no reliable predictors of the GIACI state and standard approaches to its diagnosis, despite the widespread use of glucocorticoid drugs in clinical practice and the high GIACI prevalence. Furthermore, evidence-based algorithms for glucocorticoid dose reduction have not yet been developed. Enhancement in the tactics of managing patients who have been receiving glucocorticoids for a long time may reduce the risk of GIACI complications, including life-threatening ones, as well as significantly improve the quality of life of patients.

1. [Rational pharmacotherapy of endocrine system diseases and metabolic disorders. Under the general editorship of Dedov I.I., Melnichenko G.A. Moscow: Litterra, 2006] (Rus).