Biomarkers of inflammation as predictors of treatment response in depression

People suffering with depression often do not respond sufficiently to current treatments. A wealth of research suggests that mood disorders are associated with higher levels of inflammation, yet this literature currently provides an inadequate description of the nature of the depressive disorders studied and involves extensive heterogeneity between patients and study methodologies.This thesis comprises three studies examining a broad spectrum of circulating pro-inflammatory proteins before and after interventions for depression, to test specific hypotheses that raised inflammatory markers predict a poor response to treatment, and to detect inflammatory changes in relation to clinical response. Firstly, a systematic review and meta-analysis of inflammatory markers and treatment-response was undertaken (Study 1). Subsequently, investigations of inpatients with treatment-resistant depression (Study 2) and outpatients undertaking psychological therapy (Study 3) assessed 33 biomarkers alongside treatment and clinical outcome.Interleukin-6 (IL-6), Tumour necrosis factor (TNFα) and c-reactive protein (CRP) represent the most studied biomarkers. In meta-analyses, these were higher in depressed participants than controls and showed decreases with antidepressant treatment. In both subsequent studies, numerous biomarkers were higher in depressed patients than a non-depressed control group; few were lower in depression. IL-6, CRP and TNFα were predictive of treatment outcomes. Both inpatient and outpatient studies indicate increases in inflammatory levels with treatment. Results suggest that overall, inflammation is more closely linked with refractory or somatic depression but that age and other sociodemographic constructs also affect inflammatory activity.These results support previous evidence that inflammatory responses are dysregulated in mood disorders and provide novel candidates for future study. Reports of psychological interventions and those for treatment-resistant populations are scarce. Our findings provide promising indications that, if validated in future research, measures of inflammation may be useful in identifying patients less likely to respond to standard treatments or defining more homogenous subpopulations within affective disorders