Since the onset of the COVID-19 pandemic, the effects of active smoking and exposure to second-hand tobacco smoke (SHS) on the risk of SARS-CoV-2 infection and the severity and mortality from COVID-19 have remained undetermined.1–7 An early review and meta-analysis of data1 from studies conducted all over the world was controversial and produced inconclusive results, reporting no association between smoking and risk of COVID-19, but pointing to a link between smoking and the risk of severity, hospitalization, and mortality. In Europe, the studies contributing to the above-mentioned results were from France,8 Italy,9 the United Kingdom,10–12 and Spain.13

A subsequent meta-analysis14 found a significant association between current smoking and COVID-19 mortality. Moreover, a multicenter study from Italy reported an association of COVID-19 progression among people exposed to SHS.15 Evidence indicates that COVID-19 mortality is associated with old age, being male, and having several comorbidities, particularly hypertension, kidney disease, diabetes mellitus, obesity, and cancer.13 In light of this evidence, it is particularly important to investigate the relationship between smoking and SHS with COVID-19, as stated since the onset of the pandemic,16,17 particularly among people who smoke.

The International Tobacco Control Six European Countries (ITC 6E) Survey, part of the ITC EUREST-PLUS Project,18 aimed to prospectively evaluate the psychosocial and behavioral impact of the European Union (EU) Tobacco Products Directive (TPD) and the World Health Organization (WHO) Framework Convention on Tobacco Control (FCTC) on several tobacco-related domains. To this end, a cohort of adults who smoke in Germany, Greece, Hungary, Poland, Romania, and Spain19 was established in 2016 (ITC 6E1, Wave 1), just before the transposition of the TPD, and respondents were recontacted in 2018 (ITC 6E2, Wave 2), following TPD transposition. In 2021, a third wave was conducted only in Spain (ITC ES2.5, Wave 3), as part of the ITC EUREST-PLUS Spain Project, which coincided with the initiation of the fifth wave of the COVID-19 pandemic in Spain.

In Spain, these prospective cohort data provide an opportunity to assess the COVID-19 risk among a nationally representative sample of adults who smoke. The purpose of this study was to estimate the cumulative incidence of self-reported COVID-19 and its determinants among Spanish adults who smoke.

In our study, 5.9% of adults who smoke or had recently quit smoking reported having had COVID-19 within the 1.5-year period since the onset of the COVID-19 pandemic in Spain. Since then, Spain suffered five consecutive waves (with peaks in March 2020, October 2020, February 2021, April 2021, and July 2021). A cumulative incidence of 5.9% over the 1.5-year study period is equivalent to a cumulative incidence of 136 cases/100,000 person-14 days, i.e., close to the peak of 200 cases/100,000 person-14 days reached during the first wave in Spain and well below the peaks of the subsequent waves (between 600 and 1000 cases/100,000 inhabitants-14 days).24 Although these figures should be compared with caution, given that the cumulative incidence assumes a constant rate of disease occurrence over time, they could suggest that the 1.5-year cumulative incidence among people who smoke may be lower than the 14-day cumulative incidence in the general population at the time of the survey. Other studies have found a lower percentage of people who smoke among COVID-19 diagnosed cases than among the percentage of people who smoke in the general population,25 a phenomenon that has been named the “smoking paradox”.26

The only factor positively associated with having had COVID-19 was moderate nicotine dependence. In contrast, the factor inversely associated with the cumulative incidence of COVID-19 was having a partner or friends who smoked. This finding is difficult to explain, but one possible hypothesis is that respondents who had partners who smoked increased their efforts to protect themselves from COVID-19. Similarly, respondents who reported only having friends who smoked may have had an increased awareness of their own risk and protected themselves from infection, for instance, by means of non-pharmacological practices, such as maintaining a safe social distance, wearing face masks, increasing indoor ventilation, or receiving COVID-19 vaccinations. Furthermore, no association was found between SHS exposure at home and COVID-19 incidence.

Considering the differences in SHS exposure in cars with children (male respondents reported higher exposure than females), we believe that developing intervention programs that bear these differences in mind is important to minimize the negative effects of the pandemic.27

Several potential limitations should be mentioned. First, since the COVID-19 incidence estimates are based on self-reported information, we cannot disregard a reporting bias from respondents who said they were not infected when they may have had mild symptoms or respondents who were more inclined to get tested because of their serious symptoms. For this reason, we performed a sensitivity analysis using three different definitions of a case (first, having had a test+physician diagnosis+belief in having had COVID-19; second, having had a test+physician's diagnosis; third, having had a test only). The analysis showed that the estimated RR of having had COVID-19 did not change, regardless of the lower number of cases when more restrictive criteria were applied. Furthermore, we believe that the risk of information bias is low, given the general awareness of the population of COVID-19 derived from the heavy impact of the pandemic in Spain and worldwide. Second, given that we do not know exactly when the respondents had COVID-19, we had to assume that the risk of infection was constant during the study period, although it clearly was not, given its seasonality, the existence of different variants of the virus, and vaccination uptake. Third, it is difficult to determine how self-reported SHS exposures during the period before the COVID-19 pandemic could be related to self-reported COVID-19 during the 1.5-year period after the onset of the COVID-19 pandemic. There is a possibility that smoking behavior changed between the surveys, and we cannot ignore the fact that some respondents who smoked in 2018 may have quit since then, either before or after having COVID-19. Given that the follow-up period includes 5 COVID-19 pandemic waves, such a bias would have a non-differential misclassification effect, thus biasing the RR estimates toward the null hypothesis. Fourth, attrition of the cohort could have biased the results if losses to follow-up were related with COVID-19. We do not know why some respondents were lost to follow-up, but assuming higher COVID-19 rates among Spanish residents, only 1 or 2 could have been infected. Given that the characteristics of those who were lost to follow-up did not differ from those who completed follow-up (Supplementary Table S1), selection bias is unlikely. Fifth, it is impossible to determine the extent of a respondent's exposure to SHS in different venues (bars, pubs, discos and restaurants) based on self-reported SHS exposure, since the respondent must have visited the particular venue and have seen someone smoking inside. Sixth, vaccine rollout had already begun by the time data were collected. Early on, the vaccine may have provided some protection from the variant circulating at the time, which could have influenced the results. Finally, the limited numbers of COVID-19 cases made it impossible to analyze other correlates of infection in depth.

Our focus on the incidence of COVID-19 among people who smoke is the main strength of this investigation. While other studies have analyzed information from subsets of cases detected in healthcare centers, we were able to study the cumulative incidence of COVID-19 among a nationally representative sample of people who smoke in Spain. Assessing the role of SHS exposure on COVID-19 was also considered very pertinent due to growing evidence28 of virus transmission through the exhalation of respiratory droplets. The potential of SARS-CoV-2 transmission in aerosols was the basis for Spanish COVID-19 “ad hoc” regulations29 and a ban on smoking in outdoor public places was instated that included outdoor terraces of bars and restaurants when a safe distance (set at 1.5m) could not be maintained. It should be noted, however, that this ban was only implemented in certain regions of Spain, with scant enforcement, and at a late stage of the pandemic.

In conclusion, the correlates of having had COVID-19 among people who smoke, i.e., who have moderate nicotine dependence (positive association) and among people with a partner who smoked or only had friends who smoked (negative association) should be considered when tailoring information and non-pharmacological preventive measures for people who smoke. In line with previous evidence,30 we endorse the potential of the COVID-19 pandemic as an opportunity to defeat nicotine addiction and promote national tobacco control programs to achieve a tobacco-free future. SHS exposure surveillance should be supported by proper coordination of tobacco control leaders and policy makers, in order to understand the impact of tobacco smoking in future outbreaks and epidemics.31

The EUREST-PLUS Spain Project is partially funded by the Instituto de Salud Carlos III (grant PI17/01338, co-funded by European Regional Development Fund (ERDF), a way to build Europe) and the Canadian Institutes of Health Research (Foundation grant FDN-148477). D. Carnicer-Pont, Y. Castellano, E. Fernández, M. Fu, and O. Tigova are partly supported by the Ministry of Universities and Research, Government of Catalonia (2021SGR00906). Additional support is provided by the Canadian Institutes of Health Research (FDN-148477) to A.C.K. Quah, P. Driezen, S.C. Kaai, and G.T. Fong for the work on this manuscript. Additional support to G.T. Fong is also provided by a Senior Investigator Grant from the Ontario Institute for Cancer Research (AI-004).

DCP: Conceptualization; methodology, writing – original draft, writing – review & editing, visualization.

MF: Conceptualization; methodology, writing – original draft, writing – review & editing, visualization, supervision, project administration, funding acquisition.

YC: Conceptualization; methodology, validation, formal analysis, writing – review & editing, visualization.

OT: Conceptualization; writing – review & editing, project administration.

PD: Methodology, validation, formal analysis, data curation, writing – review & editing.

ACKQ: Methodology, software, validation, investigation, resources, data curation, writing – review & editing, supervision, project administration.

SCK: Validation, investigation, writing – review & editing, supervision, project administration.

JBS: Conceptualization, writing – review & editing.

CIV: Conceptualization; investigation, writing – review & editing.

GTF: Conceptualization; methodology, validation, resources, data curation, writing – review & editing, supervision, project administration, funding acquisition.

EF: Conceptualization; methodology, formal analysis, writing – original draft, writing – review & editing, visualization, supervision, project administration, funding acquisition.

G.T. Fong has been an expert witness or consultant for governments defending their country's policies or regulations in litigation. All other authors have no conflicts of interest to declare.

In each country participating in the international Tobacco Control Policy Evaluation (ITC) Project, the data are jointly owned by the lead researcher(s) in that country and the ITC Project at the University of Waterloo. Requests for data used in this manuscript can be sent to: itc@uwaterloo.ca.

No artificial intelligence was used in the production of any part of the manuscript.