Critical Illness-induced Corticosteroid Insufficiency: What It Is Not and What It Could Be.

Critical illnesses are hallmarked by increased systemic cortisol availability, a vital part of the stress response. Acute stress may trigger a life-threatening adrenal crisis when a disease of the hypothalamic-pituitary-adrenal (HPA) axis is present and not adequately treated with stress doses of hydrocortisone. Stress doses of hydrocortisone are also used to reduce high vasopressor need in patients suffering from septic shock, in the absence of adrenal insufficiency. Research performed over the last 10 years focusing on the HPA axis during critical illness has led to the insight that neither of these conditions can be labeled "critical illness-induced corticosteroid insufficiency" or CIRCI. Instead, these data suggested using the term CIRCI for a condition that may develop in prolonged critically ill patients. Indeed, when patients remain dependent on vital organ support for weeks, they are at risk of acquiring central adrenal insufficiency. The sustained increase in systemic glucocorticoid availability, mainly brought about by suppressed circulating cortisol-binding proteins and suppressed hepatic/renal cortisol metabolism, exerts negative feedback inhibition at the hypothalamus/pituitary, while high levels of other glucocorticoid receptor ligands, such as bile acids, and drugs, such as opioids, may further suppress adrenocorticotropic hormone (ACTH) secretion. The adrenal cortex, depleted from ACTH-mediated trophic signaling for weeks, may become structurally and functionally impaired, resulting in insufficient cortisol production. Such a central HPA axis suppression may be maladaptive by contributing to lingering vasopressor need and encephalopathy, hence preventing recovery. Here, we review this concept of CIRCI and we advise on how to recognize and treat this poorly understood condition.

Three Brief Case Presentations

Case 1

A 50-year-old woman was found unresponsive, in acute shock (severe hypotension–tachycardia) hours after elective surgery (nephrectomy for an extensive upper pole renal cell carcinoma). Arterial blood gas analysis revealed severe hyponatremia (119 mmol/L) and hyperkalemia (5.4 mmol/L), and low to normal glycemia (81 mg/dL-4.50 mmol/L); the patient was transferred to the intensive care unit (ICU) for hemodynamic management (volume resuscitation and vasopressor treatment) and respiratory support (invasive ventilation). The medical file documented a 3-month history of lethargy, fatigue, a decrease in appetite and unintentional weight loss with an insidious start, as well as subtle complaints of nausea and pain in the abdomen, joints, and muscles. On ICU admission, plasma total cortisol was low (3 µg/dL-82.8 nmol/L) and plasma adrenocorticotropic hormone (ACTH) substantially elevated (>200 pg/mL->44 pmol/L). Upon clinical examination, several areas of the skin appeared darker, especially on the extensor surfaces (elbow, knees) and in and around the mouth (hyperpigmentation). An intravenous bolus of 100 mg of hydrocortisone (≈cortisol) was given followed by stress doses of 50 mg of hydrocortisone every 6 hours for 24 hours. The shock was swiftly reversed within hours and the electrolyte imbalance improved; the patient woke up and could be extubated on the next day after which she became hypertensive and agitated, which disappeared after tapering the intravenous (IV) hydrocortisone dose to a maintenance dose of 15 mg in the morning and 10 mg in the evening.

Case 2

A 55-year-old man was brought in to the emergency room in respiratory distress, hemodynamically unstable, altered mental status, and with a high-grade fever (40°C). The patient was admitted to the ICU because of septic shock due to community-acquired pneumonia. He was intubated and mechanically ventilated because of respiratory failure, and had persistently low blood pressure despite adequate fluid resuscitation, requiring high-dose vasopressor therapy. A 250-µg synthetic ACTH stimulation test was performed, upon which plasma total cortisol rose from 19 µg/dL to 24 µg/dL (524-662 nmol/L) at 60 minutes. Treatment with 200 mg of IV hydrocortisone per day was started, upon which blood pressure gradually increased. Over the next 72 hours, the dose of vasopressors could be tapered to a low maintenance dose. In the following days, he could be further weaned off from hemodynamic and respiratory support, and hydrocortisone treatment was discontinued.

Case 3

A prolonged critically ill 67-year-old male patient, who initially had been admitted to the ICU for abdominal sepsis due to bowel perforation, required several “re-do laparotomies” over the next weeks with an abdomen that was eventually left open. On ICU day 30, he remained encephalopathic and dependent on vasopressors, mechanical ventilation, and renal replacement therapy. He continued to receive opioids for wound care. He had moderate hypercalcemia, pronounced hyperbilirubinemia, and high plasma bile acid concentrations. Plasma total cortisol measured on the morning blood sample of day 7 and day 14 had been high (28 and 22 µg/dL, respectively or 772 and 607 nmol/L), but now, on day 30 in the ICU, appeared rather low considering the persistent severe illness (8 µg/dL), while plasma carrier proteins albumin and cortisol binding globulin (CBG) remained low with estimated free cortisol declining from several-fold the normal range to values now similar to those of healthy controls. Concurrently, plasma ACTH was low to normal (25 pg/mL). The corticotropin-releasing hormone (CRH) stimulation test (100 µg) revealed low incremental ACTH and cortisol responses. An ACTH stimulation test (250 µg) also showed a low incremental total and free cortisol response. Treatment with 40 mg of hydrocortisone IV in the morning and 20 mg of hydrocortisone IV in the evening was started, upon which vasopressors could be swiftly tapered and stopped. Clinically, the patient improved and regained consciousness. Also, the urine output increased and serum calcium concentrations normalized. Over the next days, the hydrocortisone dose could be further tapered to a maintenance dose, followed by several attempts to stop the hydrocortisone treatment.

Background: 2 Decades of Evolving Insights and Changing Terminology

An immediate increase in systemic glucocorticoid availability in response to severe stress, evoked by trauma, extensive surgery, or a severe medical condition, such as sepsis, is essential for survival. This is long known and evidenced by the uniform lethality following adrenalectomy in laboratory animals and by the historically high mortality rates caused by Addison’s disease, prior to the discovery and widespread availability of glucocorticoid treatment (1, 2). When un- or undertreated, people with Addison’s disease are unable to mount an appropriate HPA axis activation in response to stress and therefore are at high risk of death following such an insult. The normal humoral and neural activation of the HPA axis in response to stressors comprises the release of CRH from the hypothalamic paraventricular nucleus which, together with vasopressin, activates the pituitary corticotropes to release ACTH into the systemic circulation, which in turn stimulates the adrenal cortex to synthetize and secrete cortisol (3). Cortisol exerts feedback inhibition at both the hypothalamic and pituitary level, which is essential to restore homeostasis. At the target tissue, appropriate activation of both glucocorticoid and mineralocorticoid receptors in response to the stress-induced increased cortisol availability is necessary to bring about the essential cardiovascular, metabolic, and immune regulatory responses to allow survival (4).

Thus, an intrinsic failure or insufficiency of either of the 3 organs within the HPA axis followed by a stressful insult may lead to a quick clinical deterioration evolving to a life-threatening condition requiring intensive medical care and appropriately dosed treatment with hydrocortisone. However, the question as to the exact dose of hydrocortisone that is needed to successfully cope with the stressful insult and to protect against death is in fact still open and a matter of debate (5, 6). Yet, for the high levels of stress evoked by critical illness requiring treatment in the ICU, traditionally, doses of 200 to 400 mg of hydrocortisone per day have been recommended, which is the approximate equivalent of 10 to 20 times the substitution/replacement dose (7-9). This recommended dose of 200 to 400 mg hydrocortisone per day was largely based on the assumption that this equals the daily amount of cortisol that human patients with a normal hypothalamus–pituitary–adrenal (HPA) axis would need to produce in response to severe stress, in order to bring about the high plasma concentrations that have been documented in early studies (10, 11).

In 2000, Annane et al performed an observational study in patients with septic shock (12), which suggested that a subset of these patients may develop what was referred to as “relative adrenal insufficiency,” later renamed “critical illness-related corticosteroid insufficiency” (CIRCI) (13, 14). This presumed relative adrenal insufficiency was thought to be due to a maximally activated adrenal cortex which does not result in sufficiently high systemic cortisol availability to allow survival of patients with septic shock. The study suggested that this condition can be diagnosed by the response to a short 250-µg ACTH stimulation test, more specifically by an incremental total cortisol response that is not higher than 9 µg/dL, irrespective of baseline plasma cortisol, which can be high, normal, or low (12). This conclusion was based on the finding that such a suppressed response to ACTH infusion was most predictive for mortality (12). A low incremental total cortisol response to ACTH stimulation was further interpreted as identifying those patients who should be treated with at least 200 mg of hydrocortisone, typically labeled as “stress doses,” in order to increase the odds of survival (13, 15). This hypothesis was initially supported by a randomized controlled trial (RCT) of Annane et al, which found no survival benefit by hydrocortisone in the total population of patients with septic shock, but suggested mortality reduction in the subgroup of ACTH nonresponders (15). Yet, this trial was confounded by the concomitant use of etomidate, a drug known to suppress adrenocortical cortisol synthesis (15). Moreover, the 2008 CORTICUS RCT did not confirm survival benefit in the subgroup of ACTH nonresponding patients with septic shock (16). Also, larger RCTs that were not restricted to ACTH nonresponders gave conflicting results (17, 18). Although a rise in blood pressure and less need of vasopressors over a time course of a few days was a reproducible finding in most studies (15-18), survival benefit was only observed in 1 subsequent RCT by Annane et al (17). Yet, in contrast to the earlier hypothesis, increased survival in this RCT was observed in the subgroup of ACTH responders and not in the ACTH nonresponders (17).

In 2013, an alternative interpretation of the results of the above-mentioned studies was suggested by new insights from a paper published in the New England Journal of Medicine (19). This paper reported the results of a series of human studies which provided evidence against the concept of “relative adrenal insufficiency” as well as against the textbook assumption that critically ill patients with a normal HPA axis produce—and would need—around 200 mg of cortisol per day (20-22).

Indeed, with use of tracer technology, it was shown that critically ill patients requiring vital organ support for 7 to 10 days in the ICU produced only between 30 mg and 60 mg of cortisol per day (assessed during the morning hours and extrapolated to 24 hours), depending on the severity of the hyperinflammation, whereas they all had adequately elevated plasma cortisol concentrations (19). Instead, the data revealed that the breakdown of cortisol was substantially and uniformly reduced in all patients, due to suppressed expression and activity of the cortisol metabolizing enzymes, A-ring reductases in the liver and 11β-hydroxysteroid dehydrogenase 2 in the kidney. This reduction in cortisol metabolism coincided with a rapid decline in circulating levels of the main cortisol carrier proteins albumin and transcortin (CBG), the latter of which the binding affinity was also markedly reduced in response to inflammation and activation of neutrophils (23, 24). These alterations result in a decrease in cortisol binding and thus increase the unbound, biologically active free fraction of cortisol. Together, suppressed cortisol metabolism in liver and kidney and suppressed levels cortisol carrier proteins explain the substantial increase in systemic total and even more so free cortisol availability, as was observed in this study and in others (19, 25). The data also revealed that ICU patients do not have increased plasma ACTH concentrations, whereas such an increase would be expected in cases of “relative adrenal insufficiency,” with—at least throughout the first week in the ICU—plasma ACTH concentrations that are in fact lower than those of matched healthy subjects. A subsequent study published in 2014 further showed that pulsatile nocturnal ACTH and cortisol secretion in ICU patients, as determined by deconvolution analysis of hormonal time series and taking the prolonged cortisol half-life into account, was lower than in matched healthy subjects (26). Taken together, these studies have provided evidence for a cortisol production rate in response to critical illness that is hardly, if at all, higher than in healthy subjects, and for an increased systemic cortisol availability for target tissues that is determined predominantly by peripheral mechanisms rather than by a continued central HPA axis activation (27).

Subsequent studies of prolonged critically ill patients requiring intensive care for weeks and longer, published in 2018 (28, 29), and a 2021 study performed in a mouse model of sepsis-induced critical illness (30), provided more data to support the concept of low plasma ACTH concentrations that are explained by negative feedback inhibition exerted at the pituitary level. The mouse study revealed that sepsis-induced central activation of the HPA axis, initially via upregulation of hypothalamic CRH and vasopressin expression, followed by upregulation of the pituitary expression of their receptors, indeed activated pituitary gene expression of the ACTH precursor, proopiomelanocortin (POMC) (30). Concomitantly, activation of the pituitary glucocorticoid receptor, as part of an intact negative feedback loop, resulted in an impaired processing of POMC into ACTH and suppression of secretion of mature ACTH. In turn, this resulted in low plasma ACTH concentrations and elevated plasma POMC levels (30). Indeed, unprocessed POMC likely accumulates within the pituitary and subsequently leaches into the circulation via the constitutive secretory pathway (31). Such elevated plasma POMC in the face of low or normal ACTH was confirmed in 2 studies of human ICU patients, possibly contributing to ongoing cortisol synthesis in the adrenal cortex despite lack of increased ACTH (30, 32-34). The negative feedback inhibition at the pituitary level could be explained by glucocorticoid receptor binding by free cortisol, of which the circulating availability is increased by peripheral mechanisms as mentioned earlier, or by other glucocorticoid receptor ligands such as bile acids (30, 32, 35-37). Drugs such as opioids may further suppress ACTH and cortisol (38, 39). In addition, it was found that lower than normal total incremental cortisol responses to the ACTH stimulation test are present in virtually every ICU patient, explained by the substantially increased cortisol distribution volume, which is predominantly due to the low plasma binding of cortisol (28). These findings do not contradict the early Annane studies that found a low incremental cortisol response to 250 µg of synthetic ACTH to be highly predictive of mortality (12, 15). Indeed, plasma CBG and albumin are most suppressed in the sickest patients with the highest risk of death. Thus, the lowest incremental total cortisol responses to the ACTH stimulation test are expected for nonsurvivors. Yet, the findings from more recent studies do invalidate the usefulness of an ACTH stimulation test to assess adrenocortical integrity and function in the ICU context (19, 26, 28-30)

Recent studies also suggested that patients in the ICU for 4 weeks or longer may develop a central suppression of the HPA system, a condition that mimics central (secondary) adrenal insufficiency of patients outside the ICU context who have been treated with glucocorticoid doses slightly higher than replacement doses for 4 weeks or longer (40-42). Indeed, long-stay patients in the ICU beyond 4 weeks no longer have elevated plasma total and free cortisol; only upon recovery 1 week later on a regular ward were both ACTH and cortisol above normal (28). Another 2018 study further supported this possibility, as it showed that incremental ACTH responses to a CRH stimulation test were robustly suppressed in ICU patients who had been critically ill for more than 5 days (29). Such a condition, specifically in long-stay ICU patients, would represent a central (secondary) adrenal insufficiency. Likewise, a human postmortem study reported adrenocortical atrophy and suppressed expression of ACTH-stimulated steroidogenic genes in patients who died after prolonged, but not after brief, critical illness (43). Such central (secondary) adrenal insufficiency may result clinically in lingering otherwise unexplained vasopressor dependency and associated organ failure, encephalopathy, delirium, and fatigue, together hampering or delaying recovery. For treatment of this condition in the ICU context, where cortisol breakdown and plasma protein binding are suppressed, a daily dose of 60 mg of hydrocortisone—40 mg in the morning and 20 mg in the evening—likely suffices. Higher doses of hydrocortisone may increase the risk of further suppression of plasma ACTH, aggravating the negative impact on the integrity of the adrenal cortex, as recently shown in a mouse study (32). In contrast, in prolonged critically ill mice treated with a continuous CRH infusion from early in the disease onwards, plasma ACTH was normalized, though without an overt effect on adrenocortical structure and function or plasma corticosterone (32). It remains to be investigated in the human setting whether treatment with CRH infusion could serve as a safer endocrine alternative to increase systemic glucocorticoid availability, via stimulation of endogenous production.

However, some authors have advocated a different viewpoint and argued that higher stress doses of hydrocortisone are essential to treat patients with septic shock. This viewpoint was based on results of studies that showed reduced expression of the active glucocorticoid receptor alpha (GRα) or increased expression of the dominant negative glucocorticoid receptor beta (GRβ) in peripheral blood, which was interpreted as “general glucocorticoid resistance” (44-47). The authors therefore advocated 200 to 400 mg of hydrocortisone to treat critically ill patients, high doses that were deemed necessary to overcome such glucocorticoid resistance in target tissues (48, 49). However, suppressed glucocorticoid receptor expression documented in a mixture of peripheral blood cells harvested from critically ill patients can also be interpreted as a beneficial, adaptive response to protect the innate immune cells against the immune-suppressive effects of high cortisol while allowing the increased cortisol availability to exert its desired effects in other target tissues for an adequate stress response. Indeed, glucocorticoid receptor expression and signaling could be regulated in a tissue specific manner during sepsis and other inflammation-induced critical illnesses to avoid side effects (6). Currently, ongoing preclinical and clinical studies are further investigating the possibility of such tissue-specific regulation of glucocorticoid action in the critically ill (ISRCTN registry ISRCTN17621057)

What Is CIRCI Likely Not? Back to the Case Presentations …

First, the term critical illness–related corticosteroid insufficiency or CIRCI should not be used to describe impaired function of the HPA axis already present prior to the onset of a critical illness which is exacerbated by an acute triggering event. In case 1, the patient was likely suffering from a pre-existing (undiagnosed/untreated/latent) intrinsic disease affecting the HPA axis (eg, Addison’s disease) which led to a full-blown adrenal crisis triggered by the elective nephrectomy, which may have included ipsilateral adrenalectomy. People with latent diseases affecting the HPA axis should always be monitored carefully in the peri-operative phase after elective surgery but also when admitted in an urgent setting (eg, trauma or burn victims). Appropriate treatment should be given promptly when an adrenal crisis is suspected (clinical practice guidelines (50); primary, secondary, and tertiary adrenal insufficiency is reviewed in (9); Fig. 1).

Figure 1.

Diagnostic approach and steroid treatment for suspected adrenal crisis, refractory septic shock and critical illness–related corticosteroid insufficiency.

In case 2, we briefly described a prototypical phenotype of an emergency ICU admission of a patient with septic shock. According to the in 2017 established guidelines for the diagnosis and management of CIRCI (48), despite lack of consensus on a single test to reliably diagnose CIRCI, the patient’s increment in plasma total cortisol of ≤9 µg/dL after ACTH stimulation could be interpreted as relative adrenal insufficiency or CIRCI, as it was rebranded back then (14, 48) (Fig. 1). However, as explained in the introduction and reviewed in (5, 6), the low increment in plasma total cortisol is likely the result of a substantial increase in cortisol distribution volume, rather than of an already maximally stimulated adrenal cortex that is unable to produce more glucocorticoids. Hence, a low increment in total cortisol in response to ACTH is thus merely indicative of the high severity of illness. The observed hemodynamic response to the administered stress doses of glucocorticoids, most clearly vasoconstriction allowing reducing of the dose of vasopressors, is a pharmacological effect and this does not necessarily mean that the endogenously produced amount of glucocorticoids would be insufficient. The lack of a uniform mortality benefit across several RCTs (15-18) questions whether the observed pharmacological hemodynamic effects of adjunctive glucocorticoid therapy in patients suffering from septic shock are also clinically relevant for patient-centered outcomes. In line with these findings, most recent surviving sepsis guidelines suggest as a weak recommendation to only initiate hydrocortisone treatment in septic shock patients with ongoing and high vasopressor need, without performing an ACTH stimulation test (51).

Redefining CIRCI: The Neglected Long-stay Patient

The third case illustrates what CIRCI could be. Instead of an exacerbation of a preexisting intrinsic disease of the HPA axis (case 1), or the presence of septic shock as a condition of potential glucocorticoid resistance in some tissues (case 2), the term “critical illness–related corticosteroid insufficiency” better fits the description of “an acquired state of central adrenal insufficiency in prolonged critically ill patients.” The patient in case 3 who had been critically ill for several weeks may have developed central adrenal suppression. This could be due to the sustained increase in systemic glucocorticoid availability (ie, ongoing elevated plasma free cortisol) brought about by peripheral mechanisms, which exerted negative feedback inhibition at the central level of the HPA axis for several weeks, while high levels of other ligands targeting the hypothalamic and/or pituitary glucocorticoid receptor, such as bile acids, and the prolonged use of opioids may have contributed to lowering of plasma ACTH. Sustained low circulating levels of ACTH may have impaired the adrenocortical structure and function due to loss of ACTH-mediated stimulation. In a patient in the ICU for several weeks who remains dependent on vasopressor support and who has other potential symptoms/signs of adrenal insufficiency (impaired consciousness and mild hypercalcemia in this case), the progressive decrease in systemic glucocorticoid availability may be maladaptive, in which case it would require treatment.

How to Recognize and Diagnose a Prolonged Critically Ill Patient With CIRCI?

In analogy with the risk of noncritically ill hospitalized patients and outpatients who are treated with glucocorticoids, even in doses only slightly higher than the normal physiological range (ie, prednisolone 5 mg per day or an equivalent dose of another glucocorticoid) for 4 weeks or longer (40, 50), prolonged critically ill patients in the ICU for 4 weeks or longer should be considered at risk for developing acquired central adrenal insufficiency or CIRCI. In addition to the prolonged duration of illness, with prolonged mechanical and pharmacological support, typical critical illness–induced cholestasis and exposure to drugs that are known to negatively interfere centrally with the HPA axis, such as glucocorticoids in the earlier phase(s) of critical illness and opioids, should raise suspicion of CIRCI (Fig. 2). The CIRCI phenotype can be further aggravated by drugs that directly suppress cortisol synthesis in the adrenal cortex, such as azoles and etomidate (Fig. 2).

Figure 2.

Schematic overview of the pathophysiology and possible symptoms and signs of CIRCI. Created with Biorender.com.

Unfortunately, the symptoms and signs of CIRCI are often nonspecific and generally regarded as part of the clinical phenotype of a prolonged critically ill patient with ongoing vasopressor requirement (Fig. 2). These signs and symptoms most typically comprise poor neurological recovery after stopping of sedation, encephalopathy, continued and otherwise unexplained need of vasopressors, mild hypercalcemia, hyponatremia, and sometimes eosinophilia. Patients with CIRCI have normal or low plasma total cortisol levels, low levels of albumin and CBG, and normal-to-low plasma ACTH, whereas bile acids and bilirubin are usually elevated (6, 28, 35). However, currently, there is no single diagnostic test or imaging technique to reliably diagnose CIRCI, although the absence of increased plasma ACTH and (free) cortisol is suggestive hereof. The 250 µg of ACTH stimulation test with assessment of the incremental total cortisol response is not appropriate to diagnose CIRCI as the result is confounded by the increased cortisol distribution volume. Potentially, repeated ACTH stimulation tests with assessment of the increment in plasma free cortisol over time in ICU could be informative (Fig. 1). However, a specific threshold for an adequate increment in plasma free cortisol has not been established. In addition, a 100 µg of CRH stimulation test and assessment of the incremental response in plasma ACTH can indicate whether central HPA axis suppression is present or not (Fig. 1). However, as is the case for the ACTH stimulation test, a single threshold level for the ACTH response to CRH to allow or refute CIRCI diagnosis remains unknown. Imaging techniques, such as bedside ultrasound, computed tomography, or magnetic resonance imaging, may help differentiating with rare complications of critical illnesses such as adrenal hemorrhage (47).

How to Treat a Prolonged Critically Ill Patient With Suspected CIRCI and What About Prognosis?

Upon clinical suspicion of CIRCI in a prolonged critically ill patient and after considering the possibility of stopping any drugs that may suppress the HPA axis, given the lack of tests to accurately diagnose it with certainty, treatment can be recommended. For reasons described above (19), we recommend treatment with hydrocortisone using a daily dose of 60 mg and dividing it in 2 bolus injections, 40 mg IV in the early morning and 20 mg IV in the evening, to mimic at least to some extent the normal diurnal variation which is disrupted in critically ill patients (52) (Fig. 1). Some diagnostic tests as proposed above may be informative (Fig. 1), but initiating hydrocortisone treatment without further laboratory investigations may also be reasonable provided clinical responses are defined a priori and are carefully documented in the hours and days following treatment initiation (Fig. 2). In case such clear clinical response is absent, it is advisable to stop the treatment. In any case, after initiation of treatment with hydrocortisone for CIRCI, the dose should be tapered as soon as possible to the lowest possible maintenance dose, and attempts to fully stop treatment should be done upon recovery.

Given that the results of 2 follow-up studies have suggested that any central hypothalamic-pituitary-adrenocortical axis suppression or CIRCI present in the ICU is likely reversible upon recovery in survivors (28, 53), it may be wise to schedule follow-up by an endocrinologist for further and more accurate diagnosis of any residual pathology beyond the confounded phase of the critical illness and its management.

Open Research Questions

It should be clear from the above that many uncertainties remain which require detailed future studies. First, regarding the use of the (high) stress doses of hydrocortisone that have been advocated for septic shock, future studies should focus on possible long-term harmful effects, in particular on the role of such treatment in the pathogenesis of the long-term physical and neurocognitive legacy of critical illnesses. The COVID19 pandemic has drawn much attention to this problem, which was labeled “long-COVID,” although similar long-term sequelae are also present in survivors of other critical illnesses. Indeed, 5-year mortality was higher in ICU survivors who had a prolonged ICU stay compared with those with a shorter ICU stay. Furthermore, the use of glucocorticoids during critical illness was identified as an independent ICU risk factors (54). In addition, patients who suffered from respiratory muscle weakness while in ICU, which may be caused or worsened by the use of glucocorticoids, had worse physical function 5 years after discharge (55). Second, to allow better understanding of CIRCI, here narrowed down to the HPA axis suppression of the prolonged critically ill patient, and to more reliably diagnose this condition, more preclinical and clinical research is needed, also RCTs that investigate the patient-centered short- and long-term outcomes of the proposed treatment strategy for CIRCI. Third, to optimize treatment of CIRCI, further studies should address the most optimal dosing and tapering strategies to prevent side effects that can be expected with any degree of overdosing, perhaps not so much in the acute critical care phase, but all the more so in the long-term.