Is the Association Between Smoking and Depression Mediated by Inflammation? A Mendelian Randomization Study


Smoking, inflammation and depression commonly co-occur and may be mechanistically linked. However, key questions remain around the direction of association and the influence of residual confounding. We aimed to characterize the association between lifetime smoking and depression, as well as to assess the role that genetically-predicted C-reactive protein (CRP) level, an archetypal inflammatory marker, as a potential mediator for this association. We performed inverse variance weighted Mendelian randomization (MR) analyses using recently published summary-level GWAS data for lifetime smoking index, CRP levels, and depression. A subset of inflammatory-related genetic variants from the lifetime smoking GWAS were also used to assess the potential inflammatory causal pathways between smoking and depression. The analysis indicated significant reciprocal relationships between lifetime smoking and both depression (OR Smk-Dep = 2.01, 95% CI 1.71-2.37, p < 0.001; O R Dep-Smk = 1.09, 95% CI 1.06-1.13, p < 0.001) and CRP levels (OR Smk-CRP = 1.40, 95% CI 1.21-1.55, p < 0.001; OR CRP-Smk = 1.03, 95% CI 1.02-1.05, p < 0.001). These significant and positive associations were also supported by the majority of the robust MR methods performed. The reciprocal relationships between CRP levels (using >500 genetic instruments for CRP) and depression were not significant (OR CRP-Dep = 1.01, 95% CI 0.99-1.04; OR Dep-CRP = 1.03, 95% CI 0.99-1.07). We observed little variation in the IVW estimates between smoking and depression when we limited the genetic variants assessed to those related to inflammation or when we adjusted the analysis by CRP-levels in multivariable analysis. Our study supports potential causal associations between lifetime smoking and depression, as well as between lifetime smoking and CRP levels, but not between CRP and depression. No evidence was found that CRP mediates the relationship between smoking and depression.