Examination of peripheral basal and reactive cortisol levels in major depressive disorder and the burnout syndrome: A systematic review
Introduction
Above all, the human organism tries to maintain a state of homeostasis, which is constantly challenged by a great variety of internal and external stressors. Therefore, crucial physiological systems – stress systems – have evolved to counterbalance the disruptions caused by stressors (Chrousos, 2009). Stress systems are required to work within an optimal range during basal activity and in response to stress in order to maintain a sense of well-being, successful performance of tasks, and appropriate social interactions. A failure to do so results in excessive or inadequate basal activity and responsiveness to stress associated with a wide variety of health consequences including fatigue, worse mood, and impaired cognition (Chrousos, 2009). Thus, dysregulations of different stress systems have often been linked to stress-related mental conditions such as major depressive disorder (MDD) or the burnout syndrome.
Neuroendocrine hormones are crucially integrated in the regulation of both basal activity and the response to stress. One of the central stress-related mechanisms is the release of cortisol upon the perception of a stressor through the activation of the hypothalamus-pituitary-adrenal (HPA) axis. Cortisol binds to glucocorticoid receptors in target cells and is highly relevant for its metabolic properties providing immediate energy to the organism to overcome acute stress situations and finally to restore homeostasis (Kudielka and Wüst, 2010). Typical symptoms such as fatigue, exhaustion, depressed mood, or impaired cognition associated with a dysregulated HPA axis are also cardinal symptoms for depressive syndromes as well as the burnout syndrome.
Research has therefore extensively investigated patients suffering from major depressive disorder (MDD) – which is often associated with stressful periods prior development – with regard to potentially altered HPA axis activity (Knorr et al., 2010; Stetler and Miller, 2011). The burnout syndrome, is considered a syndrome characterized by an excess of work-related stress and in comparison insufficient coping resources in the individual to adequately counter the stress. Thus, it has also been suggested that the HPA axis may be particularly affected in individuals suffering from the burnout syndrome (Jonsdottir and Sjörs Dahlman, 2019; Rohleder, 2018). Yet, after more than 30 years of research on the matter, no clear picture emerges. An extensive body of literature has developed, which is difficult to integrate due to the many inconsistencies reported as well as the several different layers on the relation between the HPA axis and MDD and the burnout syndrome.
Furthermore, due to a large symptom overlap of the conditions, there is an ongoing debate whether the burnout syndrome is a special depression subtype, or whether they are two different entities with similar symptom presentation (Bianchi et al., 2015). Based on the strong relatedness of both conditions to stress, a systematic examination of differences and similarities of basal HPA-related measures and in response to stress in the two conditions might add a biologically informed basis for this ongoing debate. At first, we will characterize MDD and the burnout syndrome in detail and follow with a description of the function of basal and reactive measures of the HPA axis from different specimens. Following, we describe and discuss differences and similarities in HPA-related measures in these conditions.
MDD is a well characterized disorder, according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), and is defined and diagnosed on the basis of its core symptoms of depressed mood or anhedonia along with a combination of four of nine other symptoms, such as changes in appetite, sleep, fatigue, concentration, feelings of worthlessness and suicidal ideations persisting throughout the majority of the day for a minimum of a two week period (American Psychiatric Association, 2013). MDD shows a lifetime prevalence for different countries between 8–12% (Andrade et al., 2003), while in western countries 12-month and lifetime prevalence mount as high as 10.4 % and 20.6 %, respectively (Hasin et al., 2018). MDD is the leading cause of disability worldwide (World Health Organization, 2017), and there are major difficulties in improving the treatment of MDD and reducing the high prevalence rates (Vos et al., 2017). Diagnosing and treating MDD has been shown to be complicated due to its heterogeneous illness presentation exemplified by anhedonia, which is one of two cardinal symptoms, but only being truly present in up to 50 % of patients (Bogdan et al., 2013). Furthermore, the subjective nature of the patients’ report of symptoms, the requirement of the exercise of a clinical judgement based on a person’s history and cultural norms (Akil et al., 2018; Wager and Woo, 2017), and high rates of comorbidity significantly contribute to low diagnostic reliability and the potential for misdiagnosis (Regier et al., 2013; Sanislow et al., 2010). Psychological research further shows a large heterogeneity in the measurement of MDD or depressed mood with open, semi-, and completely standardized interviews and many different questionnaires often used to measure depressive symptoms, which hinders uniform recording and comparability.
Burnout, a syndrome defined by emotional and physical exhaustion, negative attitudes towards work, and a negative evaluation of one’s work performance (Maslach et al., 2001; Schaufeli et al., 2009), has increasingly been discussed as a public health concern since it was reported to be present in more than half of practicing physicians (Shanafelt et al., 2015). Burnout negatively influences physicians’ effectiveness, professionalism, patient safety and satisfaction (Panagioti et al., 2018; Shanafelt et al., 2015), while it is also prevalent in other professions at high percentages (Leiter and Schaufeli, 1996; Pietarinen et al., 2013). However, due to large heterogeneity in studies, estimations about the prevalence of the burnout syndrome seem difficult and vary for specific professions between 0–80.5% (Rotenstein et al., 2018). Burnout can affect workers of all fields, but has been particularly reported in those whose work involves an intense interaction with people (Shanafelt et al., 2017). However, burnout is not listed as an independent mental disorder in the DSM-5, while the International Classification of Diseases 10 (ICD-10) mentions burnout within the residual category Z 73 – problems related to life management difficulty – (World Health Organization, 1992) and plans to include burnout syndrome as a more detailed occupational phenomenon in the ICD-11 (World Health Organization, 2019). Importantly, currently the burnout syndrome is listed within this residual category unrelated to symptoms or time criteria not providing sufficient basis for an independent diagnosis. This results in different descriptions and operationalizations of this syndrome. Therefore, it is not surprising that different measurements (Korczak et al., 2010), as the Maslach Burnout Inventory (MBI; Maslach et al., 1986; Schutte et al., 2000) or the Shirom-Melamed Burnout Questionnaire (SMBQ; Lundgren-Nilsson et al., 2012) were used for the diagnosis of the burnout syndrome. In Sweden in clinical practice the burnout syndrome is defined as `exhaustion disorder´ whereas in the Netherlands criteria for work-related `neurasthenia´ (Schaufeli et al., 2009) were used to define the burnout syndrome. Burnout has been introduced as a concept being strictly related to insufficient coping capacities to increased work stress and not necessarily by signs of a depressed mood, which is a core symptom of MDD. However, while the risk of developing a MDD is comparably low when presenting only a mild burnout syndrome (odds ratio [OR] for comorbid MDD = 2.99), the relative risk dramatically increases as a function of burnout severity (severe burnout associated OR for comorbid MDD = 46.84; Wurm et al., 2016).
As mentioned above, there is much debate as to whether MDD and the burnout syndrome represent different pathologies with overlapping symptoms, or if they describe different aspects of the same syndrome. While some argue that the conceptual and clinically represented overlap of MDD and burnout is too high to reasonably differentiate between them (Bianchi et al., 2017), others claim to take a more differentiated look at the two conditions and consistently pronounce the conceptual differences between MDD and burnout and stress the importance of the concept of the burnout syndrome for more effective treatments in affected individuals (Ahola et al., 2005; Kakiashvili et al., 2013; Maslach et al., 2001; Schaufeli et al., 2009). Dissecting the pathophysiology of the conditions and describing differences and similarities with regard to stress physiological systems might further clarify whether underlying etiological models of these syndromes differ.
Neuroendocrine hormones are crucially integrated in the regulation of both basal homeostasis and responses to stress. The HPA axis is one of the primary stress systems in the human organism. After perceiving a stressor, the hypothalamus releases corticotropic-releasing hormone (CRH), which activates the pituitary, which subsequently releases the adrenocorticotropic hormone (ACTH). ACTH reaches the adrenal via the blood stream and thereby cortisol – the primary effector of the HPA axis – is released. Cortisol induces metabolic actions, increasing the organisms energy level necessary for adequately dealing with stressful situations (Kirschbaum, 2008). This is achieved by cortisol’s function to increase blood sugar via its effect to stimulate gluconeogenesis (the formation of glucose) and its role in liver and muscle glycogenolysis (the breakdown of glycogen to glucose-1-phosphate and glucose; Quax et al., 2013). The negative feedback loop of the HPA axis plays a further crucial role for stress regulation because cortisol starts to act as a suppressor of the HPA axis as soon as it is transported back to the primary structures of the HPA axis cascade (hypothalamus and pituitary). There, it binds to glucocorticoid receptors (GRs) and mineralocorticoid (MRs) and down-regulates the release of CRH and ACTH and eventually the cortisol levels will also decline due to the lack of releasing hormones until a state of homeostasis is reached again (Kirschbaum, 2008). Although both receptors, GRs and MRs, display mediating effects in the negative feedback loop of the HPA, there are two major differences to be considered. At first, cortisol shows a 6–10 times higher affinity for MRs compared to GRs. Thus, under stress with increased cortisol output, available MRs are completely occupied, while GRs are only occupied about 70 % (Lupien et al., 2007). It has further been suggested, that MRs contribute to the initial phase of the stress reaction, while GRs come in to play at a later stage and are responsible for termination of the physiological stress response (Joëls et al., 2008). Secondly, the distributions of these receptors in the brain are different. MRs are located in the limbic system (predominantly in the hippocampus, parahippocampal gyrus, entorhinal, and insular cortices), whereas GRs are present in subcortical areas as the limbic system, but also in cortical brain structures, with a preferential distribution in the prefrontal cortex (Lupien et al., 2007).
Chronically high or low basal cortisol levels due to a medical condition such as in Cushing’s disease (caused by a pituitary adenoma leading to an overproduction of cortisol) or Addison's disease (caused by a malfunction of the adrenals leading to insufficient cortisol production) are associated with severe symptom burden that require surgical or pharmacological treatment (Nieman and Chanco Turner, 2006; Pivonello et al., 2015). However, if basal cortisol levels are only moderately elevated or decreased, not due to any medical factor but instead due to prolonged stress, this may also lead to consequences in various areas of life particularly with regard to fatigue, mood and cognition.
An adequate cortisol response to a short-term stressor with a subsequent recovery to baseline cortisol secretion levels is essential. Chronic HPA axis activation due to ongoing long-term stress accompanied by a constantly increased cortisol output failing to return to homeostatic levels of cortisol secretion has been suggested to initiate a vicious cycle due to alterations of the glucocorticoid receptor function and distribution called glucocorticoid resistance (Quax et al., 2013). The integrated negative feedback of the HPA axis via GRs on the hypothalamic and pituitary level responsible for the downregulation of the cortisol secretion increasingly fails due to a reduction of the amount of expressed receptors as well as due to a decreased sensitivity of the glucocorticoid receptors (Pariante, 2017). This ultimately leads to constantly increased basal cortisol levels, but reduced glucocorticoid signaling, which has been causally related to fatigue, worse mood (Pariante, 2017) or impaired cognitive function (Lupien et al., 2018). However, there are also theoretical models which link a blunted cortisol response with psychopathology. Extreme stressors, after an initial extreme hyperactivation of the HPA axis leading to molecular changes, cause a permanent dysregulation of the HPA axis rendering it unable to respond adequately to subsequent stressors (Steudte-Schmiedgen et al., 2016). While this model has more been associated with conditions such as post-traumatic stress disorder, a continuously hyperactive HPA axis might ultimately reach a point of exhaustion and adapt from a hyper-responsive stress system to a hypo-responsive one (Wichmann et al., 2017).
As outlined above and illustrated in Fig. 1, different forms of basal HPA axis activity and response to stress might be associated with the conditions MDD and the burnout syndrome and shall be investigated with regard to the available peripheral HPA-related measures.
Cortisol has commonly been measured in a large number of studies in blood serum, saliva, or urine. The adequate specimen for cortisol extraction depends on the research-question that is to be answered. Cortisol in blood serum is mostly bound to serum proteins such as corticoisteroid-binding globulin (CBG) and albumin leaving only between 1–15 % of the serum cortisol free and biologically active. Thus, the measurement of non-protein-bound cortisol in serum requires sophisticated techniques often not suitable for routine use (Tunn et al., 1992). Salivary cortisol has been shown to closely reflect the concentration of non-protein-bound cortisol in blood and thus presents a valid alternative for quantification (Kirschbaum and Hellhammer, 1994). In addition, salivary cortisol measurements have increasingly been applied in psychophysiological research due to advantages such as lower costs, laboratory independence, or stress-free sampling. Urinary cortisol, however, is decreasingly being used due to the relative latency till being sampled and further metabolization processes by the organism. Yet, saliva, blood serum, and urine are well established specimens to measure cortisol, with one shared shortcoming: they provide a measurement of cortisol concentration at a single time point up to several hours. As a consequence, they are susceptible to circadian fluctuations, for instance early morning peaks or gradual daily decreases (Russell et al., 2012). The cortisol awakening response (CAR) further presents an important physiological process of the human organism. Every day between waking up and after one hour of being awake, cortisol levels are described as an inverted U-shaped curve. This HPA axis activation with waking up causes a minor peak in cortisol release in the morning in order to fuel the organism with sufficient energy to commence the day (Stalder et al., 2016). Most studies investigating the CAR, or diurnal cortisol secretion used saliva or blood sampling.
As a novelty, cortisol extraction from human scalp hair was introduced in the last decade and the long-term integrated hair cortisol concentration (HCC) emerged as a new biological marker for chronic stress exposure (Raul et al., 2004). Considering an average hair growth of 1 cm per month (Wennig, 2000), the first 3 cm of hair most proximal to the scalp represent a cumulative and retrospective marker of cortisol comprising three months prior assessment. A meta-analysis by Stalder et al. (2017) revealed small significant correlations of HCC with a single time point and mean diurnal cortisol measurements. Still, unlike other matrices, HCC represents a long-term average exposure to the hormone but is unsuitable to reflect acute levels (Sauvé et al., 2007). Due to the retrospective reflection of integrated cortisol secretion over periods of several months (Stalder and Kirschbaum, 2012), HCC represents a powerful biomarker for long-term stress exposure. Therefore, when interpreting basal cortisol levels, it is of paramount importance to consider the specimen’s characteristics and qualities. While timing is an aspect that needs to be thoroughly considered (Rohleder, 2018), the metabolism of cortisol up until being secreted via urine (Jerjes et al., 2006), its bioactivity in different specimens due to available binding-proteins or minor elevations due to sampling methods (e.g. venipuncture) are further relevant aspects causing confusion in the field if not properly interpreted (Weckesser et al., 2014).
In addition to basal short- and long-term HPA-related measures, the reactivity of the HPA axis plays a key role in the characterization of the biological reaction to stress. A crucial feature of the HPA axis is its capability to dynamically respond to a given stressor. Therefore, the HPA axis response to a standardized laboratory stressor provides an important source of information on the capacity of the organism to counter stress (Kirschbaum, 2008). For humans, on the one hand, a standardized psychosocial laboratory stressor is the Trier Social Stress Test (TSST; Kirschbaum et al., 1993). The TSST has often been used to examine differences in the HPA axis response between patient groups and healthy subjects as well as between different pathologies and can be considered the most widely applied standardized psychosocial stress test (Foley and Kirschbaum, 2010). The TSST consists of an anticipatory phase, whereby the participant has to prepare for a mock-job interview, which is to be held in front of an expert panel. Subsequently, the interview starts and after completion, the participant is asked to perform a mental arithmetic task in front of the expert panel (counting backwards in steps of 17 starting from 2023). The expert panel is instructed to reduce their face mimics as far as possible, to keep a neutral attitude (e.g. no smiling or nodding), and to not interact with the participants except for giving them instructions during these tasks. In this psychosocial stress paradigm, the so-called Mason Factors (novelty, unpredictability, uncontrollability, ego-involvement) are perceived by the participant (Mason, 1975), consistently leading to an uncontrollable social-evaluative threat experience, eliciting consistently strong HPA axis responses (Dickerson and Kemeny, 2004).
Pharmacological challenge tests, on the other hand, further also indicate GR function (e.g. hyper- vs. hyposensitivity; Menke et al., 2016). A commonly used test to investigate the sensitivity of the HPA axis is the dexamethasone-suppression test (DEX), which has been discussed for decades as potential biological test for the characterization of stress-related psychiatric disorders (Arana et al., 1985). Dexamethasone is an artificial glucocorticoid which binds to the GR in the hypothalamus and the pituitary. When, for example, 1.5 mg dexamethasone is orally administered at 06:00 pm, a significant decrease in cortisol secretion is identified at 09:00 pm due to the negative feedback of the HPA axis initiated by dexamethasone binding to GR in the hypothalamus and the pituitary. Glucocorticoid non-suppression of individuals will subsequently emerge by showing no further suppression of cortisol levels but instead a slight increase until 03:00 pm the following day. There are different variants of HPA axis challenge tests, such as the DEX test, the CRH test (where synthetic CRH is administered, causing a receptor occupancy only at the hypothalamic level examining the releasing capacity of down-stream HPA axis structures), or the combined DEX-CRH test, which is considered the most powerful variant to examine HPA axis function due to its simultaneous occupancy of CRH and GR receptors (Mokhtari et al., 2013).
It becomes evident from the above that there are many ways to investigate the HPA axis in terms of associations with MDD and the burnout syndrome. Therefore, we provide a systematic review encompassing all peripherally measured HPA axis activity areas with its final effector cortisol potentially linked to MDD or the burnout syndrome and will outline similarities and differences between the conditions.
Section snippets
Method
The systematic search was based on peer-reviewed articles, which are publicly available. Only articles published in English were included. The systematic search was carried out on 14 June 2018. Candidate studies were identified via PubMed/Medline, EMBASE and PsycINFO by using the following search strategy.
Results
In the next chapter we will summarize the studies on findings of HPA-related markers in MDD and the burnout syndrome. Starting with studies on basal HPA axis levels, we then describe studies on HPA axis responses to stress.
Discussion
In the following, we summarize and discuss findings on the identified HPA-related similarities and differences between individuals with MDD or the burnout syndrome compared to healthy controls from 190 studies identified by our systematic search. Importantly, no study was detected that directly compared pure MDD cases to pure burnout syndrome cases with regard to any examined cortisol measure highlighting the need for further studies with parallel group comparisons of the conditions. Based on
Conclusion
For MDD and the burnout syndrome, a differential diagnostic classification is debated. However, due to the large overlap in symptom presentation, the psychiatric differentiation poses a challenge to the allied mental health fields. The conducted systematic review identified similarities and differences between MDD and the burnout syndrome with regard to specific cortisol secretion parameters supporting the notion of the burnout syndrome as a mental syndrome with a unique pathophysiology.
Funding
N.R. was supported by scholarship program for the Promotion of Early-Career Female Scientists of TU Dresden. Funders were not involved in design, collection, analysis, interpretation, and publication decisions of the data.
Author contributions
AW, CK and NR designed the structure of the article, conducted the initial literature search and wrote the introduction, results and discussion part. NR further summarized the findings of the systematic search, conducted the quality assessment and prepared the tables, guided the revision process and responded to reviewers. AW further conducted the quality assessment, edited revised versions of the manuscript and responded to reviewers. JS supported the design of the structure of the article,
Declaration of Competing Interest
None declared.
Acknowledgments
We thank our interns Annette Panzlaff, Gail Baker, Isabella Zimmermann, Franziska Ruwe, Richard Künzel, Stefanie Möhlmann, Elias Bopp, and Josefine Suske for screening the studies, supporting the quality assessment and the support for preparing the tables. We also thank Julia Elmers for the support in creating the artwork.
© 2020 Elsevier Ltd. All rights reserved.