Associations between reasons for vaping and current vaping and smoking status: Evidence from a UK based cohort

Methods In a UK cohort of young adults we explored the characteristics of vapers and whether different reasons for vaping were associated with later vaping and smoking status. We used multinomial logistic regression to assess the association among ever smokers and vapers. Results Ever and vapers differed in numerous demographic and behavioural characteristics. Vaping out of curiosity was associated with lower likelihood of being a current vaper (adjusted relative risk ratio [aRRR] = 0.26, 95% confidence interval [95%CI] = 0.12 to 0.57, p < 0.001) but a higher likelihood of being a neither user compared to a current smoker at 24 years (aRRR = 2.13, 95%CI = 1.44 to 3.17, p < 0.001). Vaping to quit smoking was associated with higher likelihood of being a dual user (aRRR = 3.21, 95%CI = 1.74 to 5.90, p < 0.001) or a current vaper (aRRR = 3.26, 95%CI = 1.62 to 6.54, p < 0.001) but a lower likelihood of being a neither user (aRRR = 0.53, 95%CI = 0.34 to 0.82, p = 0.004) compared to being a current smoker at 24. These results provide strong evidence for associations between reasons for e-cigarette use and later vaping and smoking status using a large UK-based young adult cohort while adjusting for a range of potential confounders. However, the study was limited by the measures included. The reasons for vaping were self-reported retrospectively. Consequently, the measure could suffer from recall bias (e.g. young people may be less likely to state that they vaped to quit smoking if they were unsuccessful in their quit attempt). Also, the reasons for vaping chosen were not exclusive for each young person due to the multiple-choice format of this question. As with many previous studies, a multiple-choice item was used to determine reasons for vaping where open questions and qualitative analysis may have uncovered other potential reasons for vaping which are omitted from this analysis. The young people were also provided with an ‘other’ option in the questionnaire where they could provide open answer responses, however, this was rarely selected (4% of the sample) or selected without another reason being selected (2%) so was not included in the analysis. cigarette initiation people

3 adults as most research either focusses on adolescents or older adults. Among young adults in the UK, curiosity is a common reason for vaping but intention to quit smoking may be necessary for young people to effectively stop smoking by vaping.
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint

INTRODUCTION
Of an estimated 2.9 million vapers in Great Britain, over half now consider themselves ex-smokers 1 .
Evidence suggests e-cigarettes can be an effective smoking cessation aid 2 , but some smokers have not tried e-cigarettes, and not all who have tried them have successfully quit smoking 3,4 . Given the popularity of e-cigarettes, and the potential they have for population wide harm reduction, it is important to know which individuals use e-cigarettes, why, and whether different reasons for use are associated with continued use of e-cigarettes and smoking cessation.
A US study showed that vapers are more likely to be male than female and have a lower income 5 .
Additionally, smokers and ex-smokers who had quit in the last 3 years were more likely to have used e-cigarettes than those who had never smoked, but a lower percentage of current and regular vapers were current smokers compared to ever vapers 5 . Studies of UK adults have produced similar results; current and non-current vapers differ in terms of socio-economic status, number of cigarettes smoked per day and quit attempts in the last year 6 . However, there is limited research which has explored whether these differences are also observable among young adults in the UK.
Among adults in Great Britain, the primary reasons for vaping are related to smoking cessation 1 .
However, young adults in Great Britain may have different reasons for vaping than older adults; evidence from the US suggests that adolescents and young adults vape primarily out of curiosity 7 or because their friends/family vape 8 . In a study of South Korean adolescents, the most common reason for vaping among infrequent users was out of curiosity, but the most common reasons given for frequent vaping were to quit smoking and being able to vape indoors 9 . Vaping to quit smoking was also associated with continued vaping among US middle and high school students 10 . Although a number of studies have explored associations between reasons for vaping and both vaping and smoking, there is limited evidence from the UK. Furthermore, young adults are an understudied subgroup in this research area with most research either focussing on adolescents or older adults.
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org /10.1101/19006007 doi: medRxiv preprint The aim of the present study was to explore the characteristics of young adults who are vaping or have previously vaped, and whether different reasons for vaping are associated with later vaping and smoking behaviours among ever smokers and ever vapers. Specifically, among a UK cohort of young adults are different reasons for vaping by 23 years associated with vaping and/or smoking one year later?

Study Population
Young people enrolled in the Avon Longitudinal Study of Parents and Children (ALSPAC) formed the study sample. The profile of the cohort has previously been described in two publications 11,12 .
ALSPAC recruited 14,541 pregnant women resident in Avon, UK with expected dates of delivery 1 st April 1991 to 31 st December 1992. When the oldest children were approximately 7 years of age, an attempt was made to bolster the initial sample with eligible cases who had failed to join the study originally. Consequently, for some variables measured at age 7 and onwards, there are data for more than the 14,541 pregnancies stated above. The number of new pregnancies not in the initial sample (known as Phase I enrolment) and that were currently enrolled in the study at the age of 24 was 913 (456, 262 and 195 recruited during Phases II, III and IV respectively). The phases of enrolment are described in more detail in the cohort profile update 13 . The total sample size for analyses using any data collected after the age of seven is therefore 15,454 pregnancies, resulting in 15,589 foetuses ( Figure 1). Of these, 14,901 were alive at 1 year of age. Please note that the study website contains details of all the data that is available through a fully searchable data dictionary and variable search tool (http://www.bristol.ac.uk/alspac/researchers/our-data/). The questionnaires on vaping and smoking at 23+ years were completed by 4,102 children (from here on in described as young people) and 3,241 of these also completed the questionnaires at 24+ years. The 23+ questionnaire was completed by the young people between the ages of 23 and 25 and the 24+ questionnaire was completed roughly a year later when the young people were . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint between 24 and 26 years. For ease, we will refer to these time points as 23 years and 24 years respectively. Figure 1 displays the recruitment process in detail.
ALSPAC study data from 22 years onwards were collected and managed using REDCap electronic data capture tools hosted at the University of Bristol 14 . REDCap (Research Electronic Data Capture) is a secure, web-based application designed to support data capture for research studies.
Ethics approval for the study was obtained from the ALSPAC Ethics and Law Committee and the Local Research Ethics Committees. Informed consent for the use of data collected via questionnaires and clinics was obtained from participants following the recommendations of the ALSPAC Ethics and Law Committee at the time.

Exposure
In a multiple-choice question at 23 years, respondents were instructed to cross all answers that applied for their reasons for vaping. Seven response options were available: "To help me quit smoking", "To help me cut down on the number of cigarettes I smoke", "To help me with cravings in situations where I cannot smoke e.g. travel, indoors", "Pleasure", "Curiosity", "Friends use them" and "Other". The "Other" option was not included in this analysis.
Supplementary Table 1 presents all questions together with possible response options.

Outcome
At 24 years, e-cigarette outcome data was collected via questionnaire on current vaping. Current vaping was self-reported by the respondent in response to the question "Do you currently use/vape e-cigarettes or other vaping devices?" To determine smoking status at 24 years, the young people were asked "Have you smoked any cigarettes in the past 30 days?" . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint Secondary analyses were also completed in which the outcome was the highest generation of vaping device they had ever used by 24 years. Information regarding this outcome and the results of this analysis can be found in the Supplementary material.

Potential confounders
Various behavioural, lifestyle and demographic factors were measured as well as factors relating to mental health. These factors have previously been shown to impact the likelihood of vaping and smoking and could potentially influence the reason given for vaping [15][16][17][18][19][20][21][22][23][24] . Sex was recorded at birth.

Differences in participant characteristics
Differences between never and ever vapers as well as never, former and current vapers were assessed using a χ 2 test.

Reasons for e-cigarette use and smoking/vaping behaviour
In a series of multinomial logistic regressions, we explored the association between reasons for vaping by 23 years and vaping and smoking status at 24 years. Each of the six reasons for vaping were analysed individually as a binary variable (indicated/not indicated as a reason for use). Vaping and smoking status was categorised into four groups: current smoker (smoking but not currently . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint vaping), dual user (both vaping and smoking currently), current vaper (vaping but not currently smoking), or neither (not currently vaping nor smoking). Model 1 was unadjusted, model 2 adjusted for demographic factors (sex, ethnicity, SEP, and age in months at 23 year questionnaire) and model 3 adjusted for all demographic, behavioural, lifestyle and mental health factors. Details of the secondary analysis exploring reasons for vaping at 23 and the type of device used by 24 are provided in the Supplementary Material.

Multiple imputation
Only 20% of the young people who completed the questionnaire on vaping at 23 years and stated they had ever vaped and ever smoked also had complete data (including all covariates), so we used multiple imputation to increase the sample size available for analysis and minimise bias due to attrition. Models 2 and 3 were repeated using multiply imputed data (models 4 and 5 respectively).
Multiple imputation is a recommended method to account for missing data 25 . The multiple imputation by chained equations procedure was completed using the ICE package in Stata 15.1 which created 100 datasets with 20 cycles. Data were imputed for the 668 young people who completed the questionnaires at 23 and 24 years and responded that they had ever smoked and ever used an e-cigarette. Further details of how the data were imputed, including the auxiliary variables used, can be found in Supplementary Table 2.

Minimum detectable effect
A power calculation revealed that we had sufficient power (90%) to detect a minimum odds ratio of 1.67. The calculation assumed a two-tailed hypothesis (α= 0.05) and was based on a logistic regression of vaping to quit smoking by 23 years and vaping and smoking status (current smoking versus dual use) at 24 years. Of 242 young people who had not vaped to quit smoking, 216 were smokers and 26 were dual users at 24 years. Therefore, the calculation was based on an 11% probability of an event in the control group.
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint

Characteristics of Vapers at Age 23
The characteristics of the young people are described in Table 1, grouped into never vapers (n = 3,013), and ever vapers (n = 981); ever vapers were also grouped into former vapers (n = 814) and current vapers (n = 167). A higher percentage of ever vapers were male, of lower SEP (parental), had a mother who smoked in pregnancy, were harmful or hazardous drinkers and had ever used drugs or gambled than never vapers. Ever vapers were also more likely to report anxiety or depressed mood, smoking by the age of 20 or 23 years, and be current, weekly or daily smokers than never vapers at 23 years. Vaping among never smokers was rare; only 5% of ALSPAC participants who had ever vaped had never smoked, and less than 1% of ever vapers defined themselves as a current vaper who had never smoked at 23 years. There were no clear differences between these groups in terms of ethnicity, BMI, unemployment or parenthood. On average, participants were 23 years old across all groups at the initial questionnaire. Supplementary Table 3 shows these characteristics for never vapers and the former and current vapers who were included in the final analysis.

Reasons for Vaping
Participants' reasons for vaping by 23 years are shown in Table 2. Current vapers were more likely than former vapers to vape for all reasons except 'out of curiosity', the most popular reason given for ever vaping (51%). Most young people (56%) selected only one reason for vaping (Supplementary Table 4).

Reasons for vaping and smoking/vaping behaviour
Due to small numbers of never smokers who had tried vaping at 23 years (n = 47), analyses were restricted to ever-smokers. The study sample consisted of 668 young people who had completed both questionnaires and had ever vaped by 23 years. The median time between questionnaire completion at 23 years and 24 years was 12 months. The age at which the young people first vaped . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint ranged from 17 to 24 years (SD = 1; median = 22) and on average occurred 2 years before questionnaire completion. Supplementary Table 5 displays vaping characteristics for vapers at 23. At 24 years, 49% of the young people (n = 330) were current smokers who were not currently vaping.
Dual use (n = 62) and current vaping (n = 47) were less common. A substantial proportion of the young people were neither users at 24 years (n = 229).
The results for the unimputed models (Models 1-3) are shown in Supplementary Table 6 and the results for the imputed models (Models 4-5) are shown in Table 3. The results were broadly consistent, with all associations in the same direction with a similar magnitude but the confidence intervals were narrower in the unadjusted and multiply imputed analyses.

Reasons for vaping and device type used
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

DISCUSSION
The results indicate that there are substantial differences in characteristics between young people who have never and ever vaped at 23. Vaping out of curiosity and to quit smoking were common reasons for vaping among young people in this UK-based sample. Five out of the six reasons for vaping by 23 years included in this study were strongly associated with vaping and smoking status at 24 years. Identifying the factors which influence vaping among young people could be key to understanding why some young people effectively use e-cigarettes to quit smoking and others do not. Additionally, different reasons for vaping could provide insight into why some young people experiment with e-cigarettes and some continue to vape.
Similar to previous findings 6 , we found that vapers were more likely to have a lower SEP than never vapers. Vapers were also more likely to report risk taking behaviours such as drug use and gambling and were more likely to report mental health issues. Smoking has previously been associated with similar characteristics 15,17,24,26 . Smoking could mediate the relationship between these factors and vaping; most vapers smoke prior to vaping and some start vaping to help them quit smoking (as shown in this study). In line with previous findings 5 , vapers were more likely to be ever, weekly or daily smokers at 23 years, and vaping among never smokers was rare. However, given that 52% of the sample did not provide a smoking-related reason for vaping and that we are unable to determine order of product use for all participants, the similarity in characteristics could be due to vaping leading some individuals to start smoking rather than leading them to quit. Alternatively, the similarity could indicate that there is a common liability for both behaviours (i.e., the same factors . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint increase the likelihood of engaging in both behaviours). This could explain why not all young people who vape go on to quit smoking.
Our results imply that vaping to quit smoking may facilitate young people's quit attempts as those who were vaping for this reason were more likely to be e-cigarette only users rather than current smokers at 24. Previous research has also found that vaping to quit was associated with continued vaping among middle and high school students 10 and college students in the US 27 . However, vaping to quit smoking could also encourage the continued use of e-cigarettes rather than quitting nicotine products entirely as these individuals were more likely to be dual users than smokers and more likely to be smokers than neither users at 24 years. If these people aim to be completely nicotine free, they may need additional help to quit vaping or reduce the levels of nicotine in their e-liquid once they have successfully quit smoking.
Although vaping to cut down smoking is similar in terms of motive, the associated behaviour was slightly different from vaping to quit smoking. These users were more likely to be dual users than current smokers but more likely to be current smokers than using neither product. There was no clear association with being a current vaper suggesting that these users are potentially less likely to quit smoking. Logically, this could be due to the users having no intention to quit smoking all together or, because this study was over a relatively short time period, these users may still be in the process of cutting down and may eventually quit completely.
Consistent with previous research conducted in the US, we found that curiosity is the most common reason for vaping among young people 7 . Interestingly, young people were more likely to be current smokers than current vapers at 24 if they had vaped out of curiosity but were less likely to be current smokers than neither users. As the sample are ever smokers who have ever vaped, this may highlight a group of 'experimenters' who try both vaping and smoking but do not continue either behaviour. For those who had smoked less than 100 cigarettes in their lifetime, curiosity was a . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint common reason for use in a study of young adults in the US 28 which supports our interpretation of this result. Additionally, this result implies that smokers who vape out of curiosity don't want to or are unable to quit smoking.
These results provide strong evidence for associations between reasons for e-cigarette use and later vaping and smoking status using a large UK-based young adult cohort while adjusting for a range of potential confounders. However, the study was limited by the measures included. The reasons for vaping were self-reported retrospectively. Consequently, the measure could suffer from recall bias (e.g. young people may be less likely to state that they vaped to quit smoking if they were unsuccessful in their quit attempt). Also, the reasons for vaping chosen were not exclusive for each The timing of this cohort study may impact the generalisability of the findings. E-cigarettes are still a relatively new product compared to cigarettes; consequently, cigarettes were available to this cohort of young adults for a considerable period of their adolescence before e-cigarettes became widely available in 2007. In 2007, the study sample were roughly 17 years old and cigarette initiation has been shown to peak at around 15-16 years of age 29 , so it is likely that these young people experimented with cigarettes prior to being exposed to e-cigarettes. Young people today are being exposed to both e-cigarettes and cigarettes during adolescence, a key period for experimentation and risk-taking behaviour, and this may have an impact on their reasons for vaping as well as their current vaping and smoking status. Although it would be interesting to observe the association . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint between reasons for vaping and current vaping and smoking status among those who were never smokers when they first tried vaping, we are unable to identify these individuals with the current available data at 23 years. Similarly, it would also be interesting to observe the associations among never smokers at 23 years but there are too few individuals who have vaped but never smoked at 23 to report any meaningful analysis.
In conclusion, vapers and never vapers differed in socio-demographic and smoking characteristics, and reasons for vaping are associated with later vaping and smoking status. Intention to quit smoking may be necessary for young adults to effectively stop smoking using e-cigarettes.
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Competing Interests
The authors have no conflicts of interest to declare.

Acknowledgements
We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses.
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint Age (initial questionnaire) 23.9 (0.5) 23.9 (0.5) 23.9 (0.5) 23.8 (0.5) *Due to missing data, the percentage of users refers to the number of participant/participants who responded. **Only current smokers were asked this question. SEP = Socioeconomic position. The highest socioeconomic position of the mother or father was coded as manual vs non-manual occupation. Harmful/hazardous alcohol use was defined as a score of 8 or more on the Alcohol Use Disorders Identification Test (AUDIT). Anxiety was defined as scores of 5 or more on the GAD-7 which indicated mild to severe anxiety. Depressed mood was defined as feeling downhearted and depressed in the last 4 weeks and was not an indicator of clinical diagnosis. Unemployment status was defined as not currently being employed or engaging in any form of full or part-time education or training. Differences between never and ever vapers as well as never, former and current vapers were assessed using χ 2 .
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint  Table 3. Associations between reasons for vaping by 23 years and vaping and smoking at 24 years among ever smokers and ever vapers.
The models were run on multiply imputed data for individuals who ever smoked and ever vaped. Model 4 adjusted for demographic factors (sex, ethnicity, socioeconomic position, and age in months at 23 year questionnaire). Model 5 adjusted for sex, ethnicity, socioeconomic position, age, maternal smoking in pregnancy, body mass index, drug use (alcohol, tobacco, marijuana and other drugs), gambling, condom use, education/employment status, parenthood status, anxiety and depressed mood in the past 4 weeks. Note: aRRR = adjusted relative risk ratio.  668 had exposure and outcome data and had ever smoked and ever vaped at age 23 years* 189 individuals had complete covariate information* . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/19006007 doi: medRxiv preprint