Do We Have Evidence for Predictive Biomarkers for Major Depressive Disorder?: A Meta-Analysis and Systematic Review of Prospective Studies

Abstract

Background During the last century many biological hypotheses have been postulated to underlie the psychopathology of major depressive disorder (MDD). In order to gain insight in the evidence for these hypothesis from patients, a systematical search for longitudinal studies investigating biological factors before the onset of MDD was performed. Methods PubMed, PsychINFO, and Embase were used as databases. The search strategy included terms relating to (1) MDD; (2) a longitudinal design or onset/relapse/recurrence; and (3) potential biomarkers. Current leading biological models for depression were covered, including neuroimaging, neurotransmitters, neurotrophic factors, hormones and immunology. Results PRISMA guidelines were followed and 46830 articles were initially screened, indication 642 relevant articles that were screened on full text. Eventually, 90 articles fulfilled the inclusion criteria. Results were too heterogeneous or limited to perform meta-analyses for the topics: Neuroimaging (n=19), Gut (n=1), Immunology/inflammation (n=5), Neurotrophic (n=1), Neurotransmitters (n=1), Hormones (n=62), Oxidative stress (n=1). A meta analysis was performed for cortisol, which showed that higher cortisol levels predict a 44% higher chance of developing MDD (n=14, OR: 1.44 [1.12-1.84] p = 0.004), but not having a relapse or recurrent episode (n=4, OR: 1.524 [0.801 2.899] p=0.199). Conclusions Surprisingly, although a rigorous systematic search for prospective evidence for biomarkers for depression was performed, we found limited prospective studies investigating leading biological models. Only cortisol could be identified as prospective biomarker for depression onset. More prospective studies are necessary to investigate the causes (and consequences) of depression onset, relapse and recurrence. Supported By NIAS: Netherlands Institute For Advanced Study in the Humanities and Social Sciences (grant for Claudi Bocktings group)