Background Smoking and depression and anxiety are highly comorbid. Whilst there are plausible biological mechanisms through which smoking may itself lead to depression and anxiety, causal links are often difficult to infer from observational data. This is due to the well-known problems of confounding and reverse causality. These issues can be overcome in Mendelian randomisation analyses, by using genetic variants as proxies for smoking behaviour. Using a variant in the CHRNA5-CHRNA3-CHRNB4 nicotinic acetylcholine receptor gene cluster, which demonstrates robust associations with smoking heaviness in smokers, we aimed to establish whether smoking causes depression or anxiety.

Methods We performed a Mendelian randomisation meta-analysis using data on 127,948 individuals from 26 studies in the Consortium for Causal Analysis Research in Tobacco and Alcohol (CARTA). We used logistic regression to look at the associations of the smoking heaviness related variant (single nucleotide polymorphism numbers: rs1051730/rs16969968) with binary measures of depression and anxiety and psychological distress. Case definitions of these outcome measures were based upon clinical diagnosis, clinical criteria or established cut points on symptom scales. All analyses were stratified by smoking status (never, former, current and ever [former + current]). We also investigated observational associations between smoking status and smoking heaviness (cigarettes per day) and depression, anxiety and psychological distress.

Results In observational analyses, current smokers were 1.84 times (95% CI: 1.64, 2.06) more likely to be depressed, 1.70 times (95% CI: 1.53, 1.89) more likely to be anxious and 1.68 times (95% CI: 1.55, 1.83) more likely to be psychologically distressed than never smokers. Former smokers were also more likely to be depressed, anxious, or psychologically distressed than never smokers. There was evidence for positive associations between smoking heaviness and depression, anxiety and psychological distress in current smokers (odds ratios [OR] per cigarette per day: 1.03 [95% CI: 1.02, 1.04], 1.03 [95% CI: 1.02, 1.04] and 1.02 [95% CI: 1.02, 1.03] respectively). In Mendelian randomisation analyses, there was no strong evidence that the CHRNA5-CHRNA3-CHRNB4 variant was associated with depression (OR per minor allele 1.00, 95% CI: 0.95, 1.05), anxiety (OR 1.02, 95% CI: 0.97, 1.07) or psychological distress (OR 1.02, 95% CI: 0.97, 1.06) in current smokers. Results were similar for former smokers.

Conclusion Although these findings demonstrate associations between smoking and depression and anxiety, they do not support a causal role of smoking in the development of depression and anxiety.