Depression and cortisol responses to psychological stress: A meta-analysis
Introduction
Psychosocial stressors are associated with the onset (Daley et al., 2000, Kendler et al., 1999, Lewinsohn et al., 1999), symptom severity (Hammen et al., 1992), and course of major depressive disorder (MDD) (Kendler et al., 1997). One possible mechanism linking stress and MDD is altered hypothalamic-pituitary-adrenal (HPA) functioning (Carroll et al., 1981, Holsboer et al., 1984, Halbreich et al., 1985, Pfohl et al., 1985, Gold et al., 1986, Gold et al., 1988a, Gold et al., 1988b, Young et al., 1993). The entire HPA system is designed to allow organisms to adapt to physical and psychosocial changes in their environments. In humans, perceived stress activates the central nervous system (CNS), causing the release of corticotropin releasing hormone (CRH) from the hypothalamus, adrenal corticotrophic hormone (ACTH) from the anterior pituitary, and cortisol from the adrenal cortex. In turn, elevations in cortisol levels typically inhibit the HPA system via negative feedback mechanisms in the hippocampus (Jacobson and Sapolsky, 1991, Munck et al., 1984, Sapolsky et al., 1986).
The temporal dynamics of HPA responses to stressors typically consist of three phases: (1) basal activity, which reflects unstimulated, non-stressed HPA activity, (2) a ‘stress reactivity’ phase in which cortisol increases from baseline (i.e. pre-stressor) levels following the onset of a stressor, and a ‘stress recovery’ phase in which cortisol levels return to baseline levels following the offset of the stressor (McEwen, 1998). Each of these phases reflects different physiological processes, with mineralocorticoid receptors (MCRs) regulating cortisol levels during periods of low HPA activity (e.g. evening), and GCRs regulating cortisol responses to stress and cortisol levels during periods of high HPA activity (e.g. morning).
A substantial literature has documented the link between MDD and abnormalities in HPA activity (Carroll et al., 1981, Holsboer et al., 1984, Halbreich et al., 1985, Pfohl et al., 1985, Gold et al., 1986, Gold et al., 1988a, Gold et al., 1988b, Young et al., 1993). Most of what is known about HPA functioning in depressed individuals is based on studies of basal HPA activity (Halbreich et al., 1985, Asnis et al., 1987) and pharmacological challenge studies designed to test the negative feedback mechanisms in the HPA axis, such as the dexamethasone suppression test (Carroll et al., 1981). Some findings have suggested that a subset of individuals with MDD have elevated basal cortisol levels coupled with blunted ACTH responses and cortisol elevations in response to dexamathasone administration (Holsboer et al., 1984, Gold et al., 1986, Young et al., 1993).
Although the pharmacological literature on HPA activity in depressed patients has yielded important findings for the field of psychoneuroendocrinology, it is not without limitations. First, dexamethasone is a synthetic glucocorticoid, and the levels of dexsamethasone used in pharmacological challenge studies are specifically designed to mimic the highest extreme of glucocorticoid functioning in order to suppress subsequent endogenous cortisol release (Carroll et al., 1981). Such high levels of glucocortcoids may not accurately reflect the magnitude of endogenous HPA responses to psychosocial stressors. Further, dexamethasone specifically probes glucocorticoid but not mineralocorticoid receptor function, and poorly crosses the blood-brain-barrier (De Kloet et al., 1998). Finally, in contrast to psychological stressors, many pharmacological and neuroendocrine challenge tests ignore suprahypothalamic (e.g. limbic) input. Thus, psychological stress challenges offer the advantage of reflecting endogenous activity of the entire HPA system.
There is also evidence that HPA activity may differ between individuals, as well as between situations. Not all MDD patients are cortisol non-suppressors, suggesting that there may exist significant variability in HPA response patterns amongst MDD patients. Potential sources of variation include: (1) individual characteristics such as age and gender (Seeman and Robbins, 1994, Kudielka et al., 2004, Otte et al., 2005), (2) depression characteristics such as subtype (Gold et al., 1995), severity (Carroll et al., 1981, Meador Woodruff et al., 1990), hospitalization status (Maes et al., 1994), and early life stress or PTSD comorbidity (Heim et al., 2000, Yehuda et al., 2004), and (3) stressor characteristics such as type of stressor (e.g. public speaking vs. cognitive task) and duration (Dickerson and Kemeny, 2004). Finally, the effect of depression on HPA responses to stress may depend on the time of day or HPA phase examined (e.g. basal activity, stress reactivity, stress recovery).
The primary purpose of this meta-analysis is to quantify the association between MDD and HPA responses to and recovery from psychosocial stressors. A secondary goal is to identify potential moderators of the depression and HPA stress response association. The large majority of empirical studies have used cortisol rather than ACTH. Therefore, only studies using cortisol responses to stress are included in this meta-analysis.
Section snippets
Studies
Several factors were considered when selecting studies for review. First, only studies comparing MDD using structured diagnostic interviewing based on a clinical diagnosis (e.g. DSM or RDC criteria) to ND individuals were included to ensure construct validity. Second, because pediatric and adolescent MDD may be biochemically distinct from adult forms (Kaufman et al., 2001), only studies with adult samples were included. Third, since physical illness may produce abnormal neuroendocrine stress
Characteristics of selected studies
Seven laboratory stress studies with 98 MDD and 98 ND individuals were included in the meta-analyses (Breier, 1989, Trestman et al., 1991, Croes et al., 1993, Gotthardt et al., 1995, Ravindran et al., 1996, Heim et al., 2000, Young et al., 2000) (Table 1). Though not included in the formal meta-analyses, an additional study examining stress responses of 47 MDD and 39 ND participants in the field study will be discussed separately (Peeters et al., 2003). Studies reviewed in this meta-analysis
Discussion
The purpose of this meta-analysis was to review and quantify the findings of studies examining the pattern of cortisol responses to psychological stressors in depressed individuals. We compared responses between MDD and ND individuals at two phases of stress: stress reactivity and stress recovery. Our results suggest that MDD and ND individuals differ in their cortisol response patterns to stress. Specifically, in MDD individuals, cortisol activity is characterized by blunted stress reactivity
Acknowledgements
This work was supported by grants F31 AG5850 and T32 MH19391 to the first author.
References (58)
- et al.
The relationship of the dexamethasone suppression test (1 and 2 mg) to basal plasma cortisol levels in endogenous depression
Psychoneuroendocrinology
(1987) - et al.
Cortisol reaction in success and failure condition in endogenous depressed patients and controls
Psychoneuroendocrinology
(1993) - et al.
Individual differences in cortisol responses to a laboratory speech task and their relationship to responses to stressful daily events
Biol. Psychol.
(1996) - et al.
Neuroendocrine measures and lymphocyte subsets in depressive illness: influence of a clinical interview concerning life experiences
Psychoneuroendocrinology
(1997) - et al.
Physiological correlates of burnout among women
J. Psychosom. Res.
(2003) - et al.
Depression in unemployed Swedish women
Soc. Sci. Med.
(1988) - et al.
Are child-, adolescent-, and adult-onset depression one and the same disorder?
Biol. Psychiatr.
(2001) - et al.
HPA axis responses to laboratory psychosocial stress in healthy elderly adults, younger adults, and children: impact of age and gender
Psychoneuroendocrinology
(2004) - et al.
The relevance of the in- versus outpatient status for studies on HPA-axis in depression: spontaneous hypercortisolism is a feature of major depressed inpatients and not of major depression per se
Prog. Neuropsychopharmcol. Biol. Psychiatr.
(1994) - et al.
A meta-analysis of glucocorticoid reactivity in human aging: implications for Alzheimer's disease and depression
Psychoneuroendocrinology
(2005)