Abstract
AimsThe aim of this study was to use Mendelian randomization (MR) to determine the causality of the association between smoking and 14 different cardiovascular diseases (CVDs).Methods and resultsOur primary genetic instrument comprised 361 single-nucleotide polymorphisms (SNPs) associated with smoking initiation (ever smoked regularly) at genome-wide significance. Data on the associations between the SNPs and 14 CVDs were obtained from the UK Biobank study (N = 367 643 individuals), CARDIoGRAMplusC4D consortium (N = 184 305 individuals), Atrial Fibrillation Consortium (2017 dataset; N = 154 432 individuals), and Million Veteran Program (MVP; N = 190 266 individuals). The main analyses were conducted using the random-effects inverse-variance weighted method and complemented with multivariable MR analyses and the weighted median and MR-Egger approaches. Genetic predisposition to smoking initiation was most strongly and consistently associated with higher odds of coronary artery disease, heart failure, abdominal aortic aneurysm, ischaemic stroke, transient ischaemic attack, peripheral arterial disease, and arterial hypertension. Genetic predisposition to smoking initiation was additionally associated with higher odds of deep vein thrombosis and pulmonary embolism in the UK Biobank but not with venous thromboembolism in the MVP. There was limited evidence of causal associations of smoking initiation with atrial fibrillation, aortic valve stenosis, thoracic aortic aneurysm, and intracerebral and subarachnoid haemorrhage.ConclusionThis MR study supports a causal association between smoking and a broad range of CVDs, in particular, coronary artery disease, heart failure, abdominal aortic aneurysm, ischaemic stroke, transient ischaemic attack, peripheral arterial disease, and arterial hypertension.
Highlights
A large body of evidence from prospective observational studies indicates that smoking is a risk factor for several cardiovascular diseases (CVDs),[1,2,3,4,5,6] but the role of smoking for certain CVDs, such as aortic valve stenosis,[7,8,9] atrial fibrillation,[10] thoracic aortic aneurysm,[11,12] and haemorrhagic stroke[3] are limited or inconsistent
In the multivariable IVW models adjusting for genetically correlated phenotypes, the associations remained statistically significant for coronary artery disease, heart failure, abdominal aortic aneurysm, deep vein thrombosis, pulmonary embolism, peripheral arterial disease, and arterial hypertension; suggestive evidence of associations remained for ischaemic stroke, transient ischaemic attack, and atrial fibrillation (Supplementary material online, Table S3)
Genetic predisposition to smoking initiation was positively associated with coronary artery disease and ischaemic stroke in the CARDIoGRAMplusC4D consortium and MEGASTROKE,[25] respectively, but the associations were weaker than in UK Biobank (Supplementary material online, Table S3)
Summary
A large body of evidence from prospective observational studies indicates that smoking is a risk factor for several cardiovascular diseases (CVDs),[1,2,3,4,5,6] but the role of smoking for certain CVDs, such as aortic valve stenosis,[7,8,9] atrial fibrillation,[10] thoracic aortic aneurysm,[11,12] and haemorrhagic stroke[3] are limited or inconsistent. In contrast to prospective cohort studies,[1] shortterm and long-term follow-up of randomized smoking cessation studies have not revealed a significant effect on fatal coronary heart disease events.[13,14] Given that much of available data on smoking and CVD derive from observational studies, which are unable to fully account for confounding and reverse causality, and the lack of significant effect of smoking cessation on cardiovascular events in interventional trials, the causal nature of the association between smoking and different CVDs remains to be established. Mendelian randomization (MR) is an analytical method that uses genetic variants, generally single-nucleotide polymorphisms (SNPs), as unbiased proxy indicators for the modifiable risk factor to determine whether the risk factor is a cause of the disease.[15,16] Given that allelic variants are randomly allocated and fixed at conception, MR studies evade reverse causality and are less susceptible to confounding compared with conventional observational studies. We applied the MR design to determine the association between smoking and 14 CVD outcomes
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
