Endocrine risk factors for COVID-19: Endogenous and exogenous glucocorticoid excess

The novel coronavirus, named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), was first identified in January 2020. Since the initial reports of a cluster of pneumonia cases in Wuhan, China, in December 2019, acute infection by SARS-CoV-2, officially named as coronavirus disease-19 (COVID-19), has spread throughout the world [1]. Symptoms and clinical course of SARS-CoV-2 infection are highly variable, ranging from asymptomatic to lethal [2, 3]. The symptomatology and impairments of COVID-19 primarily affect the respiratory tract, but many other organs are also involved, such as the gastrointestinal tract, liver, kidneys, skin, endocrine organs as well as the musculoskeletal, central nervous and cardiovascular system [4‐6]. Epidemiological evidence from the first cases of hospitalized COVID-19 patients demonstrated that COVID-19 more likely affects older males with chronic underlying conditions, mainly metabolic, cardiovascular and cerebrovascular diseases [7]. In further studies, older age and a large body of comorbidities have been found to be associated with morbidity and mortality due to COVID-19 [2, 8]. But the underlying mechanisms remain largely unknown [9]. Especially patients with chronic metabolic diseases, like obesity, diabetes, hypertension, cardiovascular and kidney diseases showed a higher risk of acute respiratory distress syndrome (ARDS) and fatal outcome [10]. Administration of dexamethasone has been shown to be effective therapy in patients with respiratory complications due to COVID-19 [11, 12]. This appears to be partly contradictory, as the use of exogenous glucocorticoids (GCs) is, despite their efficiency in terms of immunomodulation, frequently associated with concomitant metabolic diseases, such as obesity, hypertension, cardiovascular diseases, hepatic steatosis, insulin resistance and type 2 diabetes [13‐15]. In patients with endogenous GC excess, a serious endocrine disease caused by chronic, autonomous, and excessive secretion of cortisol from the adrenal glands, the metabolic side-effects of GCs are often many times more pronounced. While chronic endogenous hypercortisolism, i.e. endogenous Cushing’s syndrome (CS), is a rare disorder with an annual incidence of 2-3/million [16], up to 70% of patients under GC treatment experience adverse systemic side-effects in the sense of an iatrogenic CS [15, 17]. Regarding the latter, it is estimated that 1-3% of the population of the United States, United Kingdom or Denmark take GCs for the treatment of inflammatory and autoimmune diseases [18‐20]. Beside the metabolic changes, patients with both endogenous and iatrogenic hypercortisolism also have increased risk of various bacterial, viral, fungal and parasitic infections, induced by an immunocompromised state [21‐23]. The outcomes of SARS-CoV-2 infection in patients with immune-modulatory therapies are, so far, only partly understood. Understanding the real impact of GC treatment on the frequency and severity of COVID-19 is an important inter-disciplinary issue. In this article, we will review the epidemiology and outcomes of SARS-CoV-2 infections in patients with endogenous and exogenous GC excess based on the literature available to date. Furthermore, we discuss safety and efficacy of SARS-CoV-2 mRNA vaccines in patients with GC excess as well as the potential effects of a SARS-CoV-2 infection on the hypothalamic-pituitary-adrenal (HPA) axis.

Since their discovery in the 1940s and the recognition of the immunosuppressive actions, GCs have been considered one of the most widely used and effective therapies to control inflammatory and autoimmune diseases [24]. Patients receiving chronic oral GCs due to rheumatic diseases are at higher risk for serious infections [25]. Over the past 12 months, since the outbreak of SARS-CoV-2, an unprecedented number of case reports and studies on COVID-19 has been published, and numerous risk factors for severe disease have been identified. Whether or not patients with systemic GC treatment are at an increased risk of SARS-CoV-2 infection and/or severe outcomes is so far not fully clarified. On the one hand, GCs were discussed theoretically to mitigate hyper-immune reactions responsible for severe disease progression [26]. An enhanced local innate immune defense in respiratory epithelial cells was reported in response to GC exposure, beside the known anti-inflammatory and anti-allergic actions of GCs [27]. However, chronic GC exposure might also increase the risk for infection and/or severe COVID-19 in these patients. Because of the mostly underlying immune system dysfunction, the treatment with immunosuppressive drugs and a higher level of comorbidities, these patients belong to a vulnerable population. Indeed, a meta-analysis of patients with COVID-19 and autoimmune diseases demonstrated in a meta-regression analysis a higher prevalence of COVID-19 in studies with a higher proportion of GC use in patients with autoimmune diseases (regression coefficient: 0.020, 95% CI 0.001 to 0.040, p = 0.042) [28]. Due to the heterogeneous studies with different sample sizes, variable diseases included, diagnostic criteria and geographic location, these data should be interpreted with caution. However, there are other studies of patients with autoimmune or inflammatory diseases and GC exposure that have described an increased risk of SARS-CoV-2 infection [26, 29, 30]. Table 1 shows studies with patients under GC treatment and SARS-CoV-2 infection. Concerning the prevalence of SARS-CoV-2 infection, in all three studies mentioned, higher GC doses were associated with a higher risk of contracting SARS-CoV-2. As an exception, Soldevila-Demenech and colleagues reported that women treated with prednisone-equivalent doses ≤ 10 mg per day had a decreased adjusted relative risk for symptoms of SARS-CoV-2. However, since the primary outcome of the study was the clinical diagnosis of COVID-19, and the majority of patients did not undergo a confirmatory SARS-CoV-2 test, these data should be interpreted cautiously. Overall, they also found an increased risk when prednisone-equivalent doses were > 10 mg per day [30].

The risk for hospitalization due to COVID-19 was analyzed by Gianfrancesco and colleagues in a publication from the COVID-19 Global Rheumatology Alliance that analyzed factors associated with hospitalization in 600 cases of COVID-19 in patients with rheumatic diseases [31]. They found, that the use of high-dose GCs (≥ 10mg per day of prednisone-equivalent) was associated with hospitalization (adjusted OR 2.05, 95% CI 1.06 to 3.96, p = 0.03). Compared to non-hospitalized patients, hospitalized patients due to COVID-19 were more likely to have chronic GC therapy (16% vs 7% for doses ≥ 10mg per day, p = 0.01). In multivariate regression models, age, comorbidities as well as GC therapy prior to the SARS-CoV-2 infection remained associated with the risk of hospitalization [31]. Other studies and case reports of patients with immune-mediated inflammatory diseases are in line with these results and point to similar observations (Table 1): the use of GCs is higher among hospitalized patients, and chronic oral GC use is associated with a higher risk of hospitalization due to COVID-19 in this population [32‐35]. In contrast, other studies and case reports of patients with rheumatic diseases showed no direct and significant relation between GC use and the risk of a severe course of disease [36‐38]. In general, the outcome for patients with rheumatic diseases does not seem to differ from that of the general population when demographic factors and comorbidities are taken into account [39]. However, since medication use and outcomes are most likely associated [31], GC administration per se already has many unfavorable side effects, and patients with underlying rheumatic diseases are more likely to have comorbidities, this should be considered and the data interpreted carefully.

The risk for a severe course of disease or death due to COVID-19 was as well analyzed in studies of patients with chronic inflammatory diseases (Table 1). In patients with inflammatory bowel disease, Brenner and colleagues reported a strong positive association between systemic GC use and adverse outcomes, with GC treatment as a risk factor for death due to COVID-19 (adjusted OR 11.6, 95% CI 2.09 to 64.74, p = 0.005) [40]. In line with these results, treatment with higher dosages of GC in patients with rheumatic diseases has been shown as an independent risk factor associated with COVID-19-related death [41, 42]. In contrast again, there are studies that do not see any significant correlation in multivariable adjusted analyses [43], or even one report that suggests reduced disease progression from mild to critical conditions under low- to medium-dose GC treatment (5-15 mg/day prednisone) [44]. However, the latter study consisted of only 22 patients with GC therapy.

Overall, there is a growing body of evidence with most studies showing that patients on chronic GC therapy are at higher risk for a SARS-CoV-2 infection and a severe course of disease, regardless of age and comorbidities (Table 1). A significant association between GC treatment and adverse outcomes for COVID-19 was also shown in the meta-analyses mentioned above [28]. In many studies patients with high-dose GC therapy were at particularly high risk for a severe course of disease, so it is reasonable to assume that there is a dose-dependent effect. This data supports the hypothesis, consistent with prior literature, that chronic GC exposure in auto-inflammatory conditions is associated with infectious complications including an adverse outcome after contracting SARS-CoV-2. However, since uncontrolled inflammatory disease activity also enhanced the risk of infection, and disease activity itself was already associated with an increased risk of hospitalization and adverse outcome [41], abrupt discontinuation of GC treatment must be avoided. Especially, as there is a high risk of HPA axis suppression here. Rather, the lowest possible dose is suggested [35].

Endogenous CS leads to a variety of alterations in the immune system and clinical complications such as sepsis and opportunistic infections. The prevalence of infections is increased to 21-51%, as a result of the excessive exposure to endogenous GCs [45, 46]. The combination of a generalized immunosuppression with impaired immune response and a chronic low-grade inflammatory state is suspected to be responsible for the large number of clinical complications [47]. Pre-existing chronic inflammatory conditions are known to increase the risk of COVID-19-related death [10]. Little is known about the risk of SARS-CoV-2 infection or severe disease progression in patients with endogenous CS. Table 2 summarizes the studies and case reports on patients with endogenous CS. Belaya and colleagues recently reported on three patients with ACTH-dependent CS and confirmed SARS-CoV-2 infection [48]. The course of these three patients ranged from asymptomatic to lethal. As the course was particularly severe in a patient with newly diagnosed and florid CS, the authors assumed that the clinical course of COVID-19 might be dependent on severity of hypercortisolism in patients with CS, and that florid CS and COVID-19 are more likely to require emergency care. Two additional manuscripts report single case studies (Table 2): the first one was a 71-year-old man with florid hypercortisolism and past metyrapone therapy, who recovered after one week of isolation [49], the second a 27-year-old woman who was scheduled for pituitary surgery due to Cushing’s disease when contracting SARS-CoV-2 [50]. The latter developed COVID-19 pneumonia with a respiratory deterioration and the need for 7 L/min oxygen supply by mask, without further risk factors associated with severe COVID-19. After controlling hypercortisolism by a ‘block and replace’ regime with steroidogenesis inhibitors and hydrocortisone, and a further supportive therapy, the patient improved and was successfully operated one month later. The authors concluded from their case that a multi-disciplinary management with prompt treatment strategies was essential, and that endogenous GC excess could have enhanced the severity of SARS-CoV-2 infection [50].

Table 3 shows published data on serum cortisol concentrations in patients with SARS-CoV-2 infection. Tan and colleagues reported plasma cortisol concentrations in patients hospitalized for COVID-19 within 48 hours after admission [51]. Patients with confirmed COVID-19 had higher plasma cortisol concentrations compared to patients without COVID-19, and cortisol concentrations were associated with a higher mortality in patients with COVID-19 (Table 3) [51, 52]. There is also the hypothesis that endogenous cortisol acts through the activation of unprotected mineralocorticoid receptors and exhibits deleterious effects in COVID-19 disease via this mechanism [53, 54]. As dexamethasone, unlike cortisol, binds more selectively to the glucocorticoid receptor, suppresses the secretion of endogenous cortisol and exhibits only a weak mineralocorticoid effect, it has been suggested that this could be an additional effective action of dexamethasone in critically ill COVID-19 patients, beside its anti-inflammatory action. Simultaneous medication with mineralocorticoid receptor antagonists was suggested to be evaluated in these patients [53, 55]. In line with that, glucocorticoid resistance has been hypothesized to contribute to morbidity and mortality in COVID-19, mechanistically via a reduced ability to inhibit the inflammation triggered by SARS-CoV-2 [56], or via activation of mineralocorticoid receptors, as known in familial glucocorticoid resistance syndromes [57].

Little is known about the effects of COVID-19 on the HPA axis. In general, endogenous GCs are essential for survival under conditions of stress [79]. GCs from the adrenal glands have profound metabolic, cardiovascular and immunological roles in an adequate stress response, elevated cortisol concentrations in critically ill patients are triggered by an activation of the HPA axis [80, 81]. Studies on the impact of COVID-19 on endocrine organs are limited. However, because of the expression of angiotensin-converting enzyme 2 (ACE2) in many endocrine glands [82, 83], endocrine organs are likely to be involved in COVID-19 [6, 84]. Alzahrani and colleagues reported on an impaired adrenocortical response in 28 patients with COVID-19 with plasma cortisol and ACTH concentrations indicating central adrenal insufficiency [85]. The authors found an inverse correlation between disease severity and cortisol or ACTH concentrations. In accordance with these data, compared to non-COVID-19 critically ill patients, lower cortisol concentrations were reported in critically ill patients with COVID-19, among whom 67% met the criteria for the diagnosis of critical illness-related corticosteroid insufficiency [83]. Contradicting with these findings is the above-mentioned report of Tan and colleagues on higher cortisol concentrations in patients with COVID-19 [51].

Autopsy studies showed frequent microscopic adrenal lesions in patients with severe COVID-19 [86, 87]. However, whether these led to altered cortisol dynamics or an insufficiency is questionable. Already in the context of the previous outbreak of SARS 2003, it has been hypothesized that SARS-CoV inhibits the adrenal stress responses causing a relative adrenocortical insufficiency, via molecular mimicry of ACTH and an immune response that cross reacts with ACTH, and/or direct hypothalamic and pituitary effects [84, 88]. Data from the previous SARS outbreak even suggested that a substantial proportion of patients (39%) who recovered from the infection had evidence of central adrenal insufficiency 3 months after recovery. As the HPA axis function of the majority recovered within a year, the authors suggested a reversible hypophysitis or direct hypothalamic injury induced by the virus [89]. Whether the beneficial effects of dexamethasone in the treatment of COVID-19 is due to general supportive benefits of GCs in critical ill patients, mitigation of the cytokine release syndrome and ARDS [12, 90], the treatment of undiagnosed adrenal insufficiency [83, 85], or the suppression of endogenous cortisol secretion with reduced effects on mineralocorticoid receptors [53], has to be clarified in the future.

Whether patients with endogenous hypercortisolism or GC treatment have a blunted messenger RNA (mRNA) vaccine response is not known. Systemic GC therapy affects both the innate and adaptive immunity. Because of the immunosuppressive actions and the known glucocorticoid receptor expression on all immune cells, the primary concern regarding SARS-CoV-2 vaccines in the setting of GC excess is efficacy. GCs are known to suppress the ability of antigen presenting cells to process antigen and to impair T cell activation [91]. In animal models, GC exposure was also shown to influence B cell function [92]. Furthermore, there is some evidence suggesting that chronic GC treatment may decrease B-cell counts and specific antibody responses in human [93]. Studies on the functional vaccine outcomes in patients with chronic high-dose GC therapy point toward a potentially impaired vaccine-based immunity with decreased serologic responsiveness in individuals with exogenous GCs [94‐96]. However, the decrease of vaccine efficacy in these setting was small, and it has been rather established that patients with chronic GC treatment generate an adequate humoral response to vaccines [97‐100].

The use of intraarticular corticosteroids was previously associated with an increased risk for developing influenza despite vaccination [101]. Due to the long-lasting systemic effects of intraarticular corticosteroid injections and the time to establish an effective response of the adaptive immune system, it is suggested to conduct an elective corticosteroid injection no less than two weeks prior and one week following mRNA vaccine dose, whenever possible [102, 103]. However, on the other hand, there are also studies that do not see any connection between vaccine responsiveness and short-term systemic GC administration [100, 104, 105]. To the best of our knowledge, there are no available studies on vaccine response in patients with endogenous GC excess. In dogs with hyperadrenocorticism treated with trilostane, the immune response to canine parvovirus vaccination was comparable with that of healthy dogs [106].

Patients with chronic immunosuppressive therapies have been excluded from the large studies of SARS-CoV-2 mRNA vaccines to date. In a recently published prospective study with 436 transplant recipients and immunosuppressive therapy including tacrolimus (83%), corticosteroids (54%), mycophenolate (66%), azathioprine (9%), sirolimus (4%), and everolimus (2%), antibodies were detectable in only 76 of 436 participants (17%, 95% CI 14-21%) at a median of 20 days after the first dose of SARS-CoV-2 mRNA vaccine [107]. This is in contrast with the robust early immune response observed in the SARS-CoV-2 mRNA vaccine trials with 100% seroconversion within 2-3 weeks after immunization with mRNA-1273 and BNT162b2 [108, 109]. Despite limitations, such as antibody measurements after only one dose, this study shows that immunocompromised individuals may show a reduced response to mRNA vaccination [107], and could therefore be at a higher risk. Both the mRNA-1273 and BNT162b2 vaccine trials allowed GC use under an oral prednisone equivalent dose of 20 mg per day for up to 14 days, or 280 mg of prednisone equivalent in total [108, 109]. However, so far, no subgroup analysis was provided with participants who were exposed to GCs. Further studies are required for exogenous GCs and other immunosuppressive agents to determine whether these patients can build up a serum titer of anti-SARS-CoV-2 antibodies comparable with individuals not taking immunosuppressive agents. To answer this question, several studies on efficacy and safety of SARS-CoV-2 mRNA vaccination in specific patient populations with immunomodulatory therapies are currently recruiting or planned (ClinicalTrials.gov: amongst others, NCT04848493, NCT04839315, NCT04845997, NCT04824651, NCT04798625).

Noteworthy, all vaccines approved or currently in late-stage development, including vaccines based on mRNA, are considered inactivated vaccines. With regard to live vaccination, there are safety concern with systemic GC doses equivalent to 2 mg/kg or a dose of 20 mg per day of prednisone equivalents for 2 or more weeks [100, 110, 111]. However, non-live vaccines can be used without restrictions in patients under immunosuppressive therapy [112].

A potentially decreased response to vaccination should not lead to a decision not to vaccinate, and vaccines should not be deferred in these vulnerable patients. As patients with endogenous or exogenous GC excess might be at higher risk for a severe course of COVID-19, immunization is recommended. Thus, it can be assumed that in patients with a higher risk, vaccination even with the result of a lower titer of antibodies is better than no vaccination. Depending on further studies on the immune response, repeated immunization after the period of altered immunocompetence, such as florid endogenous CS or high-dose GC treatment, may be necessary after achieving remission.

There is growing evidence that patients under especially high doses of exogenous GC or during endogenous GC excess exhibit an increased risk for SARS-CoV-2 infection and for a severe course of COVID-19. Therefore, in patients with higher doses of exogenous GCs, the lowest possible dose is recommended to control the underlying inflammatory disease. In patients with endogenous CS, adequate control of hypercortisolism is recommended, as increased risk has been associated with the level of cortisol excess. Moreover, good monitoring of metabolic and cardiovascular comorbidities seems particularly important to ameliorate the risk profile of patients with GC excess.

Many questions about the relation between SARS-CoV-2 and endogenous or exogenous GC excess remain unanswered. Further studies are required to understand the risk of these patients as well as the long-term effects of COVID-19 in this population. The immediate collection of data in the SARS-CoV-2 pandemic is important to answer key pressing questions. International collaboration between specialists appears to be of great importance for certain patient groups in order to obtain valid data [113]. However, current limitations should be taken into account in the interpretation of studies and case reports. In particular selection bias and confounding are important to consider [39]. We want to acknowledge that our recommendations are based on the aforementioned and currently available data and evidence, and that it has to be updated based on ongoing data.