HPA Axis Function as Biomarker for Atypical and Melancholic Depression

• ¹ Institute of Psychiatry, Psychology & Neuroscience (IoPPN), Department of Psychological Medicine, United Kingdom

Introduction: Major depression is one of the most crucial issues in modern healthcare. According to the latest WHO report, it has been ranked as the single largest contributor of years lived with disability. The abnormalities of the hypothalamo-pituitary-adrenal (HPA) axis have been consistently linked to depressive conditions over decades of research. However, the precise role of the alterations in the HPA-axis regulation in depression is not clear yet. In particular, recent studies have been suggesting that the heterogeneity of depressive disorders may determine the degree and the direction of HPA-axis abnormalities. Namely, a few studies have demonstrated that increased cortisol levels and blunted response to challenge tests are characteristic of depression with melancholic features, such as insomnia and weight loss. On the other hand, in atypical depression, characterized by symptoms like hypersomnia, hyperphagia and fatigue, as well as comorbidity with PTSD, borderline personality disorder and history of childhood trauma, decreased basal and post-Dexamethasone suppression test (DST) cortisol levels and increased sensitivity of glucocorticoid receptors (GR) have been implicated. However, whether there indeed exists such a dichotomy and which factors contribute to it, has not been thoroughly evaluated yet. Objective: To systematically analyse the evidence concerning the abnormalities of the HPA axis in melancholic and atypical depression subtypes available by March 2017. Materials and Methods: A systematic search was performed in MEDLINE, PsycINFO and Embase databases using the following search items: “hypothalamic-pituitary-adrenal” or “HPA” or “cortisol” or “corticotropin releasing hormone” or “ACTH” AND “atypical depression” OR “non-atypical depression” OR “melancholic depression” OR “non-melancholic depression” OR “endogenous depression” OR “non-endogenous depression”. Filters were set to include studies in adult humans, published in the English, German, or Russian languages, from the first available article until March 2017. Results: A total number of 56 articles were selected for the review. These were subdivided into three groups: 35 studies which compared the “melancholic” subtype of depression with nonspecified depression subtype and/or control subjects; 14 studies which compared “atypical depression” (including “atypical traits/reversed neurovegetative symptoms”) with nonspecified depression subtype and/or control subjects; and 7 studies which compared “melancholic” subtype with “atypical” subtype. The studies varied sufficiently in the measures used to identify HPA-axis activity (i.e. basal/post-challenge cortisol/ACTH/CRF in blood, urine, saliva or the CSF), as well as in the criteria used to define depression subtypes. In the first group, the majority (n=20) of studies addressed the response to DST in melancholic or endogenous depressed patients. The key conclusions drawn from the analysis of papers were that: DST in Melancholic depression is characterized by high specificity (above 90%), yet low sensitivity (below 50%). Besides, non-suppression rates in depressed patients varied dramatically depending on the criteria used for evaluating depression subtypes, e.g. clinican-rated “endogenicity” and Research Diagnostic Criteria (RDC) “definite endogenous” phenotypes showed a stronger association with increased post-challenge cortisol than RDC “probable endogenous depression” and NEDDI-defined phenotype. At the same time, studies focusing on particular symptoms such as weight loss and insomnia showed a more consistent association with DST non-suppression than those focusing on phenotypes in general. At the same time, two of the studies reported a stronger association of non-suppression with psychotic subtype and psychotic symptoms (delusions) rather than melancholic features. The association of elevated basal cortisol with melancholic depression was not consistent across all studies. Moreover, one study questioned the validity of basal cortisol as a measure of HPA-axis activity since it was prone to situational alterations including admission stress. Regarding the second group, the definitions also varied, with some authors employing DSM criteria, others adhering to Atypical Depression Diagnostic Scale or the presence of the so-called “reversed vegetative symptoms” such as increased appetite, weight gain and hypersomnia. A few studies demonstrated significantly decreased basal cortisol levels in atypical depression vs. depressed patients not matching subtype criteria or controls; others did not find any difference between the groups (notably, no elevated HPA-axis function). Two studies using desipramine challenge test demonstrated blunted response in atypical depressive patients compared to controls. Distinct characteristics such as weight loss, hypersomnia, and carbohydrate craving, as well as a subset of early/chronic atypical depression, showed a more robust association with decreased basal or post-DST cortisol than the broadly defined subtype. Notably, two studies demonstrated no effect of the symptom of mood reactivity on the difference in DST suppression rates. The direct comparison between atypical and melancholic depression subtypes on some parameters failed to yield consistent results. While two of the studies reported decreased basal serum cortisol in AD patients vs. both controls and melancholic patients (one of the cohorts included chronically depressed patients only), the rest failed to show a significant difference between the subtypes. Two of the studies also reported elevated serum ACTH levels in atypical depression. The limitations of studies include sufficient differences in methodology, small sample sizes, and, as well as in the previous groups, lack of consistency in diagnostic criteria and selected outcome measures. Conclusion: Although quite a large number of studies on the issue exist, varying approaches to diagnosing depression subtypes and varying outcome measures make it difficult to compare results reliably. The review confirms that there is a generally evident difference in HPA-axis function between melancholic and atypical depressive subtypes, although a more precise evaluation of the dichotomy would require focusing on biological symptom clusters rather than subtypes in general. Challenge tests, in particular DST, accessing just one subtype of GR, remain insufficient as diagnostic tools even for melancholic depression due to low sensitivity. Further research may benefit from applying unified criteria to subtype identification, with emphasis on biological symptoms and the most reliable measure of HPA-axis activity. The emergence of newer methods such as measuring hair and nail cortisol, as well as of novel challenge tests, may help facilitate this process. Finally, the development of novel approaches to categorizing mental disorders, may result in the identification of a more precise and specific phenotype characterized by HPA-axis abnormalities, with account of genetic, epigenetic factors, early life stress, cognition and life events.