Investigating evidence for a causal association between inflammation and self-harm: a multivariable Mendelian Randomisation study

Background: The causal role of inflammatory markers on self-harm and suicidal risk has been studied using observational data, with conflicting results. Confounding and reverse causation can lead to bias, so we appraised question from a genetic perspective to protect against these biases. We measured associations between genetic liability for high levels of inflammatory markers Interleukin-6 (IL-6) and C-reactive protein (CRP) on self-harm, and conducted a secondary analysis restricted to self-harm with suicidal intent. Methods: We conducted two sample and multivariable Mendelian randomisation (MR) to assess the effects of IL-6 and CRP on self-harm utilising existing data and conducting new genome wide association studies to instrument IL-6 and CRP, and for the outcome of self-harm. Results. No single nucleotide polymorphisms (SNPs) reached genome-wide significance for self-harm, however 193 SNPs met suggestive significance levels (p<5x10-6). We found no evidence of an association between our instruments for IL-6 and self-harm in the two-sample MR, however we found an inverse association between instruments for CRP and self-harm, indicating that higher levels of circulating CRP may protect against self-harm (inverse variance weighted OR 0.92, 95%CI 0.84, 1.01, p=0.08; MR Egger OR 0.86, 95% CI 0.74, 1.00, p=0.05). The direct effect estimate for IL-6 was slightly smaller in the multivariable MR than in the two sample MR, while the CRP effect estimates were consistent with the two sample MR (OR 0.92, SE 1.05, p=0.09). Conclusions. Our findings are conflicting and indicate that IL-6 and CRP are not robust etiological markers of increased self-harm or suicide risk.