The study is the first of its kind to pool observational and genetic data on smoking and COVID-19 to strengthen the evidence base.
The evidence on whether smoking is associated with a greater likelihood of more severe COVID-19 has been inconsistent. Several studies carried out early on in the pandemic reported a lower prevalence of active smokers among those with COVID-19 than in the general population, but other population-based studies have suggested that smoking is a risk factor for the infection and its severity.
Most of the research to date has been observational in nature and unable to determine cause and effect. So, the researchers combined observational and Mendelian randomisation analyses to better understand the relationship.
Mendelian randomisation is an analytical method that uses genetic variants as proxies for a particular risk factor – in this case genetic variants that make someone more likely to smoke or to smoke more heavily. The method overcomes some of the limitations of observational analyses to obtain evidence in support of a causal relationship.
The researchers drew on anonymised linked primary care records, COVID-19 test results, hospital admissions data and death certificates to look for associations between smoking and COVID-19 infection severity from January to August 2020 in 421,469 UK Biobank participants. Of these, 281,105 participants had data regarding their genetic make-up available for the main analyses.
During the study period, 13,446 (3.2%) people took a COVID-19 swab (PCR) test, 1,649 (0.4%) of whom tested positive; 968 (0.2%) required admission to hospital; and 444 (0.1%) died with their infection.
The majority (59%) of participants had never smoked; over one-third (37%) were former smokers, and only 4% were current smokers. Among current smokers, most (71%) were light/moderate smokers; only 29% were heavy smokers.
In the observational analyses, current smokers were 80% more likely than never-smokers to be admitted to hospital with COVID-19 and significantly more likely to die. Moderate to heavy smokers were about twice as likely to be hospitalised, and six times more likely to die than never-smokers, and these effects were not explained by age, sex, or other factors.
The genetic analyses showed a predisposition to smoking was also associated with a 45% higher risk of infection and 60% higher risk of hospital admission for COVID-19. The risks went up with the intensity of smoking; the risk of infection more than doubled; there was a five-fold increase in the risk of hospital admission; and a 10-fold increase in the risk of death with the virus.
Dr Jemma Hopewell, Professor of Precision Medicine & Epidemiology in NDPH, British Heart Foundation IBSR Fellow, and joint senior author of the paper, said ‘The serious health risks associated with smoking are well-established. By combining insights from genetic and observational data we have now also been able to conclude that smoking is harmful for COVID-19 outcomes. Our study not only provides reliable evidence supporting a causal link between smoking behaviours and COVID-19, but also highlights the value of electronic healthcare records and importance of resources such as UK Biobank for undertaking timely research into public health priorities.’
Adam Von Ende, Medical Statistician in NDPH and joint first author commented ‘There are many potential biases that can influence COVID-19 research but by addressing our question using different analytical approaches we were able to make more robust conclusions.’
Dr Ashley Clift, Clinical Research Fellow in the Nuffield Department of Primary Care Health Sciences and joint first author added ‘The consistency across all our analyses was striking, and provided clear evidence that smoking is linked to increased risks of hospitalisation and of death with COVID-19. The results provide another compelling reason for smokers to consider quitting.’