Smoking as a Risk Factor for Cardiovascular Disease in Females and Males: Observational and Mendelian Randomisation Analyses in the UK Biobank

Introduction: Observational studies have shown that smoking is more strongly associated with cardiovascular disease (CVD) in females than in males. It remains unclear whether these observed sex differences reflect differences in the causal effects of smoking between the sexes. Methods: This study investigated sex-specific associations between ever smoking, smoking continuation, and the number of cigarettes smoked per day and CVD outcomes by conducting sex-stratified observational and Mendelian randomisation (MR) analyses in the UK Biobank. Results: In observational analyses, we found a greater excess risk of ever smoking, smoking continuation, and number of cigarettes smoked per day for CVD in females than in males with female-to-male ratios of hazard ratios (HRs) of 1.08 (95% confidence interval CI 1.04, 1.12), 1.15 (1.07, 1.22), 1.05 (1.02, 1.08), respectively. Results were similar for CHD, and we found no sex differences for stroke. Results from MR analyses were directionally similar; however, we were not able to detect statistically significant sex differences in the effect of smoking exposures on any CVD outcome. For subarachnoid haemorrhage (SAH), we found indications for a stronger causal effect of ever smoking in females as compared to males (female-to-male ratio of ORs 2.61 95%CI 1.06, 6.42). Conclusion: This study shows that both smoking initiation and higher smoking intensity are observationally and causally related to a higher CVD risk in both females and males. Observed sex differences in the association between smoking and CVD were directionally similar to sex differences in the causal effects of smoking on CVD. In general, MR estimates were more uncertain, and the causal effects of smoking on CVD may be similar in females and males.