A network meta-analysis of stress mediators in suicide behaviour

Natalie Thomas, Christopher W. Armstrong, Abdul Rahman Hudaib, Jayashri Kulkarni, Caroline Gurvich

    Research output: Contribution to journalReview articlepeer-review

    7 Citations (Scopus)


    Stress homeostatic mediators are the most consistently anomalous biomarkers observed in suicide and may therefore point to a common ‘core biology’ of stress susceptibility, and suicidal behaviour. Previously reported meta-analyses have demonstrated aberrant levels of stress cortisol and inflammatory cytokines in suicide patients compared to controls, and significant associations between the stress regulator FK506-binding protein 51 (FKBP5) gene and suicidal behaviour. Although these independent studies were investigated as separate entities in suicide, stress mediators interact in a dynamic system, collectively giving rise to system changes physiologically, and ultimately psychologically and behaviourally. It is therefore important to study the dynamic network these stress mediators. Network meta-analysis allows for the simultaneous comparison of more than two biological mediators, and for comparisons to be made between mediators that have not been directly compared before, using previously reported, pooled meta data. Such network approaches may help study the complex biological phenomena of suicide and may provide better prediction of biological risk of suicidal states. Methods: This study aimed to establish the comparative relationships between key stress mediators in suicidal patients compared to non-suicidal controls using a random-effects network meta-analysis approach. The key stress mediators included cortisol, six inflammatory markers (interleukin-6 (IL-6), interleukin-4 (IL-4), interleukin-2 (IL-2), tumour necrosis factor-a (TNF-α), interferon (IFN-y) and transforming growth factor β (TGF-β), and the FKBP5 single nucleotide polymorphism (SNP) allele. Data was derived from three previously published meta-analysis. The study population comprised of 1348 suicidal patients, defined as suicide attempters, completers, or patients with severe suicidal ideation, and 1750 non-suicidal controls, defined as healthy controls and psychiatric patients without suicidal ideation or previous attempts. Results: Pair-wise indirect effects of stress mediators in suicide compared to controls demonstrated that relative to the effect of the FKBP5 risk SNP allele on suicide risk, the magnitude of differences (suicide vs control) for the levels of IL-2 (SMD −0.72; 95% CI, −0.135 to −0.09 and IL-4 (SMD −0.71; 95% CI, −1.34 to −0.08) were significantly smaller (with 95% confidence intervals not crossing the null). The comparative relationships between stress mediators in suicidal behaviour demonstrates that the dynamic stress network relationship is dysregulated in suicide patients when compared to controls. Conclusions: This model suggests that a genetic stress susceptibility with downstream abnormal cortisol stress axis functioning, together with anomalous interactions between the inflammatory system, may be one of the neurobiological correlates of suicide behaviour. This biological state may leave the individual physiologically susceptible and unable to cope with environmental stressors, which is consistent with the stress-diathesis hypothesis of suicide behaviour.

    Original languageEnglish
    Article number100946
    JournalFrontiers in Neuroendocrinology
    Publication statusPublished - Oct 2021


    Dive into the research topics of 'A network meta-analysis of stress mediators in suicide behaviour'. Together they form a unique fingerprint.

    Cite this