Evidence for genetic correlations and bidirectional, causal effects between smoking and sleep behaviours

Introduction Cigarette smokers are at increased risk of poor sleep behaviours. However, it is largely unknown whether these associations are due to shared (genetic) risk factors and/or causal effects (which may be bi-directional). Methods We obtained summary-level data of genome-wide association studies of smoking (smoking initiation ( n =74,035), cigarettes per day ( n =38,181) and smoking cessation ( n =41,278)) and sleep behaviours (sleep duration and chronotype, or ‘morningness’) ( n =128,266) and insomnia ( n =113,006)). Using LD score regression, we calculated genetic correlations between smoking and sleep behaviours. To investigate causal effects, we employed Mendelian randomization (MR), both with summary-level data and individual level data (n=333,581 UK Biobank participants). For MR with summary-level data, individual genetic variants were combined with inverse-variance weighted meta-analysis, weighted median regression and MR Egger regression methods. Results We found positive genetic correlations between insomnia and smoking initiation ( rg =0.27, 95% CI 0.06 to 0.49) and insomnia and cigarettes per day ( rg =0.15, 0.01 to 0.28), and negative genetic correlations between sleep duration and smoking initiation ( rg =-0.14, -0.26 to -0.01) and chronotype and smoking cessation ( rg =-0.18, -0.31 to -0.06). MR analyses provided strong evidence that smoking more cigarettes per day causally decreases the odds of being a morning person, and weak evidence that insomnia causally increases smoking heaviness and decreases smoking cessation odds. Conclusions Smoking and sleep behaviours show moderate genetic correlation. Heavier smoking seems to causally affect circadian rhythm and there is some indication that insomnia increases smoking heaviness and hampers cessation. Our findings point to sleep as a potentially interesting smoking treatment target.