Tobacco Use and the Risk of COVID-19

Overview

Contrasting hypotheses, including that of a protective role of nicotine, have been generated concerning the association between smoking and the occurrence of COVID-19 infection. The question has attracted a lively scientific and public debate. However, the studies conducted so far are based on clinical samples, with a majority of hospital case series, thus most likely suffering from bias due to selection. The investigators propose to conduct an analysis of the potential causal association between smoking or the use of the Swedish smokeless tobacco snus and the occurrence of COVID-19 using data from a newly identified retrospective cohort in Sweden. Information on tobacco use will be extracted by public dental clinic records in Stockholm Region between October 2015 and January 2020. Information on diagnoses of COVID-19 will be obtained through record linkage with health care registers of inpatient and outpatient care during the period February 2020-August 2021. Socio-demographic information will be accrued from the register of the total population in Stockholm Region (Statistics Sweden). The risk of COVID-19 for tobacco users compared to non-tobacco users will be calculated as a measure of association adjusting for potential confounders.

Background The role of tobacco use in the incidence and in the prognosis of the Severe acute respiratory syndrome-Cov2 disease (COVID-19 pandemic) has raised much international interest, due to contrasting findings reported so far in the scientific literature. On one side smoking may increase the risk of adverse outcomes in clinically overt COVID-19 requiring hospital admission. Several cross-sectional studies (case series) indicated a higher risk of negative disease progression among smokers compared to non-smokers in cases hospitalized with severe symptoms of the disease. On the other side, studies also suggest also that smoking is associated with a decreased risk of hospital admission with a diagnosis of COVID-19, or a decreased risk of occurrence of the disease in the community. These studies have noted 4 to 5 times lower proportions of smokers among patients hospitalized for COVID-19 compared with the underlying source population. A metanalysis of hospital case series confirmed this gap. Given the public health importance of tobacco use as a risk factor for morbidity and mortality, it is urgent to provide both the scientific and the broad lay community with sound information from large population studies, something that the World Health Organization has also recommended. Aim The present study aims to elucidate the potential causal association between tobacco use (smoking and snus) and COVID-19 using a population-based cohort study. Methods In Sweden, the public dental clinics collect information at each visit on lifestyle and co-morbidities with relevance for oral health (health declaration), with a uniform instrument in use from October 2015. Smoking and snus use are ascertained as past use, current use and categories of intensity of current use. During February 2020 the usual routine activity of the clinics was disrupted by the pandemic, and the so-called oral health check-up was discontinued. About 450 000 adult clients of the public dentistry in Stockholm Region were identified in the period October 2015 to January 2020. Among these about 120000 were assessed within 7 months from the onset of the pandemic, therefore represent recent experiences. The investigators expect about 9% to be smokers and about 11% to be snus users. With the help of the national personal numbers assigned to every resident in Sweden at birth or at immigration, the investigators will link the reports of smoking among these clients with diagnoses of COVID-19 and other comorbidities registered in the regional database of inpatient and outpatient health care. The information in the dental clinic registers will be linked to the total population of the region of Stockholm, in order to extract socio-demographic information. Statistical analysis plan The investigators will examine the role of tobacco use on the risk of the following events: any diagnosis of COVID-19, hospitalization, receiving intensive care, and death because of COVID-19 using generalized linear models, with and without possible confounders (age, gender, country of birth, type of employment, education and cohabitation), and within subgroups to identify potential effect modification (age group, gender \[in analyses of snus use\], assumed infection risk associated to occupation, pandemic period, and strategies of virologic testing, very restricted in the first 3 months, available on individual request thereafter).