Vaping during the COVID-19 lockdown period in Belgium

Background Due to the Corona Virus Disease 2019 (COVID-19), the Belgian government set out a range of measures to prevent the spread of the virus. One measure included closing all non-food shops, including vape shops. Methods A retrospective online questionnaire was used to investigate the impact of closing the vape shops on the vaping and/or smoking behavior of current vapers. Results The sample (n = 202) reached consisted of 70% exclusive vapers, 29% dual users and 1% no-product users. Over half (55%, 112/202) of participants needed to buy e-liquid during the lockdown, with a small majority being able to buy e-liquids – mostly with their usual nicotine concentrations, flavor or brand –, but as much as 39% (44/112) of them ran out of e-liquid. Those buying e-liquid mainly did so by making purchases via foreign online webshops. A similar pattern was observed with respect to purchasing hardware, with about half (47%, 95/202) of participants reporting hardware availability and with a small majority (53%, 107/202) reporting hardware unavailability. Of those indicating that hardware was not available, 38% (41/107) ran out of a properly functioning e-cigarette. A non-trivial minority had to consume e-liquids with another nicotine concentration, flavor or brand than usual. One seventh of exclusive vapers before lockdown relapsed partly or completely to smoking during the lockdown. The main reasons for changing vaping and/or smoking behavior included the unavailability of e-liquid with nicotine, the unavailability of hardware, and stress/worries about COVID-19. Conclusions The majority of vapers succeeded in maintaining their vaping behavior as usual, highly likely due to (illegally) buying consumables online. Nevertheless, for a minority the lockdown period resulted in unintended consequences and these vapers relapsed (completely) to smoking. Even during periods of lockdown, smokers and vapers should be able to purchase low(er)-risk alternatives to smoking, for example e-cigarettes. Supplementary Information The online version contains supplementary material available at 10.1186/s12889-021-11637-4.


Background
Early March 2020, the first confirmed infections with the Corona Virus Disease 2019  in Belgium were recorded. In order to prevent the spread of the virus, the government set out a range of measures including closing stores (exceptions: food stores, newsagents and pharmacies) for a period of almost two months (March 18th until May 10th). Consequently, all vape shops were closed and people who vaped (i.e. vapers) could no longer purchase consumables (e.g., eliquids, batteries, coils) for their electronic cigarettes (ecigarettes) in brick-and-mortar stores. Some newsagents and some food stores, however, offered a limited product range of e-liquids and e-cigarettes, or temporarily expanded their range (sometimes in cooperation with closed brick-and-mortar stores). Buying consumables online is complicated in Belgium due to the fact that the online sale of e-cigarettes and e-liquids is prohibited [1].
More specifically this prohibition implies that Belgian vape shop owners cannot sell their consumables online (nor do they do so), that foreign webshops cannot sell to Belgian vapers as this is a criminal offence, and that customs can (and occasionally do) seize consumables ordered abroad. The online purchase of consumables itself is not punishable, however. In contrast, people who smoke (i.e. smokers) could still buy their cigarettes in food stores or from newsagents during the lockdown period. Unfortunately, smoking cigarettes is responsible for a very high rate of preventable premature deaths worldwide [2]. The availability of low(er)-risk alternatives to smoking, like e-cigarettes, could be one way to curb the burden of smoking on our society. Currently, it is generally accepted that e-cigarettes are low(er)-risk alternatives to smoking as they are much less harmful to health [3]. E-cigarettes also have the potential to help smokers to quit smoking and to remain smoke-free [4]. The severely restricted availability of low(er)-risk alternatives (i.e. e-cigarettes) during lockdown, however, could have had negative consequences on vaping behavior, including increasing the risk of (re)lapse to smoking.
The purpose of this study was to investigate to what extent the lockdown (i.e. closed vape shops) in 2020 had an impact on vaping and/or smoking behavior among people who were currently vaping (including also dual users of e-cigarettes and tobacco cigarettes). We used a retrospective online questionnaire to investigate the impact of closed vape shops on changes in vaping and/or smoking behavior among current Belgian vapers.

Recruitment
The targeted sample of participants consisted of current (adult + 18 years) Dutch or French speaking Belgian vapers, including exclusive vapers and dual users of ecigarettes and tobacco cigarettes. In order to recruit participants, advertisements with the link to the online questionnaire were distributed through social media (Twitter, Facebook, "Belgische Damp Bond / Union Belge pour La Vape"). A total of 241 participants started filling out the online questionnaire, with only two (1%) not giving their consent for participation. The questionnaire was fully completed by 202 participants (response rate of 84%).

Study design
The cross-sectional survey was available in Qualtrics from May 25th until June 8th 2020. After participants clicked on the link, they firstly received general information about the study and were asked to give their consent to participate. When they did not give their consent, participants were thanked for their interest and the questionnaire was finished. Participants who gave their consent, were provided with the questionnaire, thanked at the end and finally were given the opportunity to enter their e-mail address if they wanted to receive information about the results of the study. The majority of the questions were retrospective in nature, meaning that the questions referred to the lockdown period, so from March 18th until May 10th 2020.
The current study is in accordance with the General Data Protection Regulation (GDPR) and was approved by the Societal and Social Ethics Committee of the University of Leuven (G-2020-1960).

Materials and outcome measures Questionnaire
The questionnaire (see Additional file 1) started with assessing sociodemographic variables like age, gender, highest obtained degree, occupation, marital status, net income per month and nationality (all predefined categories, except age). Next, participants reported on their smoking and vaping status before the lockdown period (i.e. before March 18th 2020) by indicating how frequently (daily, weekly, monthly, not) they were vaping and/or smoking. Subsequently, participants were asked about their stock of consumables (e-liquids and service/ hardware), their need to purchase consumables, their ability to purchase these consumables, and how and where they had made such purchases during the lockdown period (i.e. from March 18th until May 10th 2020). This was followed by questions that focused on the type of e-liquid (nicotine concentration) they were using and their vaping frequency (more, less, same as usual, quit vaping) during the lockdown period. Finally, the questionnaire investigated whether any changes had occurred during the lockdown period regarding the vaping and smoking behavior, separately for exclusive vapers and dual users, and what the reasons were for these changes.

Statistical analyses
We used descriptive analyses, including frequencies and proportions, to describe the obtained results. Crosstabulations were used to get a more detailed picture of specific subgroups (e.g., not being able to purchase supplies) and their changes in vaping and/or smoking behavior during the lockdown. Additionally, logistic regressions were carried out to predict smoking initiation among exclusive vapers (starting to smoke vs. remaining smoking abstinent during the lockdown) and to predict a rise in smoking behavior among all participants (smoking more vs. smoking the same/less during the lockdown). Four models were included using three different predictors (i.e. sufficient e-liquid in stock, no vs. yes; ability to purchase e-liquid, no vs. yes; and ability to purchase hardware, no vs. yes). With respect to the ability to purchase e-liquid and hardware, the original response categories (i.e. yes, but not needed due to sufficient stock; yes, and made purchases; no, but not needed due to sufficient stock; no, and fell without) were recoded into simple no vs. yes responses. Only those who explicitly reported that they were not able to purchase e-liquid/hardware and fell without were recoded as those who were not able to make purchases. Some analyses were conducted using Statistica, version 13. Logistic regressions were conducted using SPSS, version 26.0.

Results
A total of 241 participants started filling out the online questionnaire with 202 participants (response rate of 84%) fully completing the questionnaire. Participants were on average 39 years old (SD = 9.89) and were mainly middle-class, full-time working males (all details are presented in Table 1). Before the lockdown period, only one participant (1%) was not smoking nor vaping, 70% of the sample were exclusive vapers (n = 142) and 29% were dual users (n = 59). Among those dual users, the majority (66%, 39/59) smoked daily and the others smoked weekly (19%, 11/59) or monthly (15%, 9/59). Vapers (including dual users) were vaping daily (96%).
We also asked participants to indicate their ability to receive service or purchase any hardware (e.g., ecigarettes, batteries, coils, atomizers, …) and their need for doing so during the lockdown period (see Table 2). About half of participants (47%, 95/202) indicated that hardware could be purchased, and nearly half (40%, 38/ 95) of these participants did so. Among the other half of participants (53%, 107/202) who reported that hardware was not available, less than half (38%, 41/107) ran out of properly functioning hardware and the majority (62%, 66/ 107) were not in need of hardware. Among those who were able to, and did buy some hardware (n = 38), the majority (63%, 24/38) did so via an online webshop (outside Belgium) or in a brick-and-mortar vape shop on demand (24%, 9/38) or from a newsagent (18%, 7/38), see Fig. 1.

Logistic regressions
Four logistic regression models were used to predict smoking initiation during the lockdown period among exclusive vapers (see Table 5) and to predict a rise in smoking behavior among all participants (see Table 6). The first model included sufficient e-liquid in stock (no vs. yes) during lockdown as the sole predictor. The odds to start or increase smoking were 22 and 17 times higher, respectively, in participants who did not have sufficient e-liquid in stock than in those who did. Second, when including the sole predictor "ability to purchase e-liquid" (no vs. yes), the odds of starting to smoke vs. remaining smoking abstinent among exclusive vapers, were 77 times higher in participants who were not able to make a purchase than in those who were able to purchase e-liquid. The odds of an increase in smoking behavior were 30 times higher in participants who were not able to buy e-liquid. The third model included the ability to purchase hardware (no vs. yes) as the predictor. The odds to initiate or to increase smoking, were 18 times and eight times higher, respectively, in participants who were not able to purchase hardware than in those who were able to make a purchase. Finally, the fourth model included all three predictors combined from the previous models. With regard to smoking initiation, only one predictor proved to be significant: the odds of smoking initiation were 48 times higher when e-liquid could not be purchased than when it could be purchased. Considering increased smoking behavior as dependent variable, having sufficient e-liquid in stock and the ability to purchase e-liquid were significant predictors. The odds to increase smoking were five and 18 times higher in participants who did not have sufficient e- Fig. 1 Locations where e-liquid and/or hardware (and service) was purchased. Note: E-liquid n = 92; Service/hardware n = 38

Discussion
One measure taken by the Belgian government to prevent the spread of COVID-19 in 2020 was closing all non-food shops, including the vape shops. The current online retrospective questionnaire study shows that, during this lockdown period, most vapers continued vaping and that even one fifth of dual users became exclusive vapers. Thus, although the brick-and-mortar vape shops were closed and there was a ban on the online sale of ecigarettes in Belgium, most vapers were able to purchase consumables (e-liquid, hardware) as usual and when needed. Nevertheless, 11, 35 and 33% had to buy eliquids with a different nicotine concentration, a different flavor, or a different brand than usual, respectively. Although we cannot determine if the vapers reached were already buying consumables online before the lockdown period (as this was not asked), it is not unlikely that this was the case as we made use of a convenience sample of web-savvy vapers. By doing so, it is likely that we reached vapers who were familiar with making online purchases. As the online sale of consumables in Belgium was prohibited, but vapers were still able to make online purchases, this implies two things: 1) that foreign webshops have sold illegally to Belgian vapers while taking the risk to be punished for doing so, and 2) that the ban on the import of consumables purchased online is not enforced.
Being able to purchase consumables online served, for many vapers, as a kind of protective factor against relapse to smoking. If the legislation had been enforced, these vapers might not have been able to obtain consumables and relapse to smoking would probably have been higher. Vapers who are not familiar with buying online or who not prefer to take the risk of buying consumables online, closing vape shops during a lockdown can result in a decrease in the availability of consumables. The limited availability of low(er)-risk alternatives for smoking could consequently result in an increase in smoking behavior. Recent data from the US shows indeed that increasing taxes on e-cigarettes result in an increase in smoking behavior and vice versa [5]. These observations highlight that e-cigarettes act as economic substitutes for regular cigarettes.
It should be stressed that for a non-trivial minority of participants the lockdown period resulted in combining vaping with (more) smoking, or even quitting vaping completely and relapsing to smoking. One seventh of exclusive vapers started smoking (with 60% running out of e-liquid, but having a properly functioning e-cigarette) and or switched completely to exclusive cigarette smoking (with all of them running out of e-liquid and 90% running out of a properly functioning e-cigarette), and the majority of dual users remained using both products while smoking more than usual. Only a minority of dual users switched completely to vaping. Sciensano [6], a  federal scientific institution and federal research centre for public health, animal health and food safety in Belgium, also found an increase in smoking prevalence (19.8% in 2018 vs. 21.9% during lockdown 2020). Among smokers, around one third also reported an increase in the amount of cigarettes smoked [6]. The main predictor for smoking initiation or increasing smoking behavior was the inability to purchase e-liquid. Although only a minority of the current sample relapsed to smoking, extrapolation to the Belgian population (approximately 9,500,000 citizens aged 15 year and older) [7] and translating this to absolute numbers may be instructive to estimate the number of vapers that were potentially negatively affected by the lockdown measure in Belgium.
If we were to apply our results to the most recent Belgian vaping prevalence data [8], this is what we could expect: According to IPSOS, 10% (950,000 people) of the population vapes at least once in a while, with 3% (285,000 people) vaping daily. Our sample mainly consisted of daily vapers of whom 56% did not have enough e-liquid during the lockdown and of whom 39% ran out of e-liquid. This would amount, in absolute numbers, to 159,600 and 62,244 Belgian daily vapers, respectively. A total of 6% quit vaping, which would translate to 17,100 daily vapers. IPSOS (2019) further reports that 41% of all vapers are exclusive vapers (389,500 people). In our  In contrast to the relapse rates that we observed, Caponnetto and colleagues [9] found that Italian exclusive smokers decreased their cigarette consumption and exclusive vapers showed a similar vaping behavior as before the lockdown. It is not clear what factors could explain these differences. A likely explanation is that Italy is one of the countries that, despite original plans to do so, in the end did not close the vape shops, such that vapers (more) easily could access their consumables during the lockdown period compared to Belgian vapers. There are differences between EU-countries with regard the opening or closing of vape shops during the lockdown periods, for example, vape shops were open in Finland, France, Switzerland, Denmark, Italy, Romania, Norway, Netherlands, Cyprus, and Sweden. In the countries where brick-and-mortar vape shops were closed (UK, Ireland, Spain, and Germany), the online sale of consumables was allowed, in contrast to Belgium. In the longer term, research is needed to investigate the effects of these different regulatory decisions on smoking and/or vaping behavior.
The current results should be read in light of the following limitations. Firstly, we used a convenience sample of current vapers, which may not be representative of the general population of vapers. As a consequence, it can be expected that people who suffered negative consequences of the lockdown on their ability to continue vaping were more likely to participate in the study (i.e. self-selection bias). Secondly, retrospective crosssectional surveys are prone to recall bias, meaning that participants may experience difficulties to correctly recall specific details about their past behaviors. Thirdly, due to the cross-sectional nature of the study, any potential causal relationships needs to be interpreted with caution. Fourthly, although we distributed the online questionnaire on social media, the response rate was relatively low resulting in a small, mainly male sample. Fifthly, as is true for most questionnaire-based studies, there was no biological verification of self-reported smoking and/or vaping status.

Conclusions
Carrying out lockdown measures to curb a pandemic can (negatively) impact specific behaviors of specific groups of people. Related to vaping, it is critical to rethink the potential unintended consequences of closing vape shops while there is also a ban on the online sale of e-cigarettes. Smokers and vapers should be provided with the possibility to purchase low(er)-risk alternatives for smoking, even during a lockdown. Denying this opportunity can result in an increase in smoking prevalence, especially when tobacco cigarettes are easily available and the lack of available e-liquids/hardware, and exposure to toxic substances due to the lack of maintenance of e-cigarettes.