Baseline cortisol and the efficacy of antiglucocorticoid treatment in mood disorders: a meta-analysis

Introduction
Hyperactivity of the Hypothalamic-Pituitary-Adrenal (HPA) axis and high cortisol levels have been widely reported in patients with mood disorders but previous clinical trials investigating the efficacy of antiglucocorticoid treatment in this population have reported inconsistent findings. These inconsistencies among these studies may be because not all patients with mood disorders have increased HPA axis activity and therefore might not benefit from antiglucocorticoid treatment. The aim of this meta-analysis was to investigate whether baseline cortisol levels influence influence the efficacy of antiglucocorticoid drugs in patients with mood disorders.

Methods
PubMed and Scopus databases were searched systematically up to October 2018. We included studies using metyrapone, ketoconazole or mifepristone in patients with major depressive disorder, bipolar disorder and major depressive disorder with psychotic symptoms. We tested for a difference in cortisol levels between responders (a reduction equal to or greater than 30% on depression scales following antiglucocorticoid treatment) and non-responders (a reduction of less than 30% on depression scales). We performed a meta-analysis to look specifically at differences in cortisol levels in the sample of patients treated with cortisol synthesis inhibitors (metyrapone and ketoconazole) and in those treated with glucocorticoid receptor (GR) antagonist (mifepristone).

Results
We were able to retrieve data from 11 of the 16 selected studies and to include 9 studies in the meta-analysis. In the overall sample (N = 846), responders had similar baseline cortisol levels compared with non-responders (standardised mean difference, SMD = -0.03, 95% CI [-0.17, 0.12], p = 0.75). In the group of patients treated with cortisol synthesis inhibitors, responders (N = 109) had significantly higher peripheral baseline cortisol levels compared with non-responders (SMD = 0.42, 95% CI [0.01, 0.83], p = 0.047). In the group of patients treated with a GR antagonist (N = 737), both responders and non-responders had similar baseline cortisol levels (SMD= -0.09, 95% CI [-0.25, 0.07], p = 0.26).

Conclusion
Our data suggest that only patients with higher cortisol levels at baseline benefit from treatment with cortisol synthesis inhibitors and support a potential role for cortisol as a predictive biomarker for treatment with cortisol synthesis inhibitors in patients with mood disorders.