Association of e-Cigarette Use With Discontinuation of Cigarette Smoking Among Adult Smokers Who Were Initially Never Planning to Quit

IMPORTANCE Cigarette smokers not planning to quit are often overlooked in population studies evaluating the risk-benefit potential of electronic nicotine delivery products (e-cigarettes). OBJECTIVE To evaluate whether e-cigarette use is associated with discontinuing cigarette smoking among smokers who were initially never planning to quit. DESIGN, SETTING, AND PARTICIPANTS This cohort study used US nationally representative data from the longitudinal Population Assessment of Tobacco and Health Study (waves 2-5 conducted between October 2014 and November 2019), with participants evaluated in 3 pairs of interviews. Adult daily cigarette smokers initially not using e-cigarettes and with no plans to ever quit smoking for good (2489 observations from 1600 individuals) were included.


Introduction
Most published studies that investigate whether use of electronic nicotine delivery products (e-cigarettes) can help cigarette smokers quit smoking have been restricted to smokers who are planning to quit, or they have not considered smokers' quit intentions. 1 However, a recent study using data from the US nationally representative Population Assessment of Tobacco and Health (PATH) Study looked at daily cigarette smokers who were not planning to ever quit smoking and found that subsequent daily use of e-cigarettes was positively associated with change in intentions to quit cigarette use. 2 A long-standing theory suggests that taking even a first step toward contemplating quitting smoking can have a positive impact on net cigarette cessation rates 3 ; thus, evaluation of factors associated with cigarette discontinuation among smokers not planning to quit is important to understanding the range of potential impacts of e-cigarette use on net cigarette cessation. In this study, we assessed whether use of e-cigarettes was associated with cigarette discontinuation among adult daily cigarette smokers who were initially not planning to ever quit smoking and were initially not using e-cigarettes. We

Methods Participants
The PATH Study is an ongoing, nationally representative, longitudinal cohort study of youth and adults in the US that collects self-reported information on tobacco-use behaviors, attitudes and beliefs, and health outcomes. Data were collected using audio computer-assisted self-interviews Westat and approved by the Westat Institutional Review Board. All adult respondents aged 18 years or older provided written informed consent. For the adult interview, the overall weighted response rate at wave 1 was 74.0%, at wave 2 was 83.2%, at wave 3 was 78.4%, at wave 4 was 73.5%, and at wave 5 was 69.4%. 4 Further details regarding the PATH Study design and methods [5][6][7] and overall demographic and tobacco use distributions 8 are published elsewhere. Details on interviewing procedures, questionnaires, sampling, weighting, response rates, and accessing the data are also published elsewhere. 4 This report followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.
We conducted analyses among the 17.0% (95% CI, 16.0%-18.0%) of adult (aged 18 years or older) cigarette smokers in the US who were daily cigarette smokers, not using e-cigarettes (though they may have been using other tobacco products), had no plans to ever quit smoking for good, and had follow-up data on e-cigarette use status and cigarette discontinuation (2489 observations contributed by 1600 individuals). A total of 908 individuals were lost to follow-up over the course of the study period. Those with missing data on educational attainment or annual household income were included in analyses as valid unknown groups; those with missing data on other covariates were excluded from this study using listwise deletion (35 individuals). To adjust for complex study design characteristics (eg, oversampling) and attrition, all estimates were weighted and represent the resident population of the US aged 18 years or older at the time of data collection who were in the civilian, noninstitutionalized population in 2013-2014. 8

Sample-Defining Measures and Key Variable Measure
At each interview, respondents were asked separately whether they currently smoke cigarettes, whether they currently use e-cigarettes (ie, any electronic nicotine product), and whether they plan to ever quit cigarettes or tobacco for good. Exact item wordings are provided in Table 1. We restricted our analysis to those who at baseline assessment were currently smoking cigarettes every day, were currently using e-cigarettes not at all, and currently did not plan to ever quit cigarettes or tobacco for good. Our key variable measure was e-cigarette use at the follow-up wave using a 3-level e-cigarette use variable at follow-up: (1) no use, (2) nondaily use, or (3) daily use ( Table 1).

Outcome Measures
We defined cigarette discontinuation at follow-up assessment as having not smoked cigarettes in the past 12 months or currently smoking not at all. We defined discontinuing daily cigarette smoking at follow-up assessment as having not smoked in the past 12 months, currently smoking not at all, or currently smoking some days. Exact item wordings and variable categories are provided in Table 1. other non-e-cigarette tobacco product and had ever used that tobacco product fairly regularly were asked about intending to quit using tobacco rather than specifically about intending to quit smoking cigarettes.
b e-Cigarettes refers to all e-products (ie, e-cigarettes, e-cigars, e-pipes, and e-hookah).

Statistical Analysis
We used generalized estimating equations to evaluate the association between time-varying e-cigarette use and time-varying cigarette discontinuation and discontinuation of daily cigarette smoking (so as to assess both the transition away from smoking altogether and the transition away from daily smoking in particular) using 3 wave pairs: waves 2-3, waves 3-4, and waves 4-5. and calculate adjusted odd ratios (aORs) and 95% CIs. 13 Analyses were run on the W2-W5 Restricted Use Files. 4

Results
The composition of the US population of adult daily cigarette smokers, based on data from 1600 adult participants, who were not using e-cigarettes and had no plans to ever quit smoking for good is based on a denominator sample size less than 50, or the coefficient of variation of the estimate or its complement is larger than 30%.
In the study population overall, 6.2% of adult daily cigarette smokers who were not using e-cigarettes and had no plans to ever quit smoking for good were not smoking cigarettes at all at follow-up (95% CI, 5.0%-7.5%). The odds of cigarette discontinuation were significantly higher among those who used e-cigarettes daily (28.0%; 95% CI, 15.2%-45.9%) compared with those who did not use e-cigarettes at all (5.8%; 95% CI, 4.6-7.2; aOR, 8.11; 95% CI, 3.14-20.97), while the odds of cigarette discontinuation among those who used e-cigarettes nondaily did not statistically differ from those who did not use e-cigarettes at all (   b Analyses were adjusted for biological sex, race and ethnicity, age group, educational attainment, annual household income, cigarettes smoked per day, and wave pair; all covariates were assessed at baseline wave of each wave pair; generalized estimating equation models were fitted specifying the unstructured covariance and within-person correlation matrices, Wald χ 2 22 of 95.4 (P < .001) for the cigarette discontinuation model; Wald χ 2 22 of 170.6 (P < .001) for the daily cigarette discontinuation model. c Estimate should be interpreted with caution because it has low statistical precision. It is based on a denominator sample size of less than 50, or the coefficient of variation of the estimate or its complement is larger than 30%.

Discussion
This study focused on a group of daily smokers who were not initially planning to ever quit smoking and who did not use e-cigarettes; results showed that those who subsequently used e-cigarettes every day experienced an 8-fold higher odds of cigarette discontinuation compared with those who did not use e-cigarettes at all. Smokers with no plans to ever quit smoking tend to smoke more cigarettes per day and have lower educational attainment and household income compared with their counterparts who do plan to quit 2,14 and are also often overlooked in the population-based e-cigarette use and cigarette discontinuation literature. 1 Our findings here suggest that such smokers should be specifically considered when evaluating the risk-benefit potential of e-cigarettes for smoking cessation in the population.
Prior population-based work has shown that intentions to quit cigarettes change alongside uptake of e-cigarette use, 2 suggesting a possible mechanism underlying our findings. Specifically, the association between e-cigarette use and change in quit intentions could expand the pool of smokers engaged in cessation efforts and ultimately increase overall cessation rates. Indeed, an experimental study among smokers not planning to quit found use of e-cigarettes to be associated with increased plans to quit, 15 and an ecological momentary assessment found that giving e-cigarettes to smokers with no quit intentions was associated with reductions in cigarette smoking when e-cigarettes were used frequently. 16 Further, our findings are consistent with clinical trials showing that giving nicotine replacement therapy (another type of nicotine delivery product) to smokers not planning to quit is associated with increases in overall smoking cessation rates. [17][18][19] It is also possible, however, that change in quit intentions may not explain the association found between uptake of e-cigarette use and cigarette discontinuation; research from Hughes et al 20 on the natural history of smokers' efforts to quit smoking shows that intentions to quit smoking can change quickly and repeatedly and may not be a good indicator of behavior.

Limitations
Limitations of our study include the lack of ability to evaluate whether quit intentions changed after uptake of e-cigarettes, and we did not assess whether any changes in quit intentions mediated the association between e-cigarette uptake and cigarette discontinuation; thus, future work is important to understand the causal mechanisms underlying our findings. Further, the collection of qualitative data can be useful to informing possible processes that may contribute to discontinuing smoking among this group of smokers. We also note that there may be self-selection differences between those who subsequently used e-cigarettes and those who did not, which were not assessed here.
Another limitation is that we had relatively small sample sizes, although weighting ensured the sample was representative of the US population. We also did not assess long-term patterns of subsequent relapse and requitting. The e-cigarette marketplace has also changed since our study period of 2014 through 2019.

Conclusions
Although controlled experimental studies have shown associations between e-cigarette use and smoking cessation among smokers without plans to quit, it is important to identify whether similar associations exist in a real-world context. This cohort study found an association between daily e-cigarette use and cigarette discontinuation among daily smokers in the US population who initially had no plans to ever quit smoking in their lifetimes. These findings call for consideration of smokers who are not planning to quit when evaluating the risk-benefit potential of e-cigarettes for smoking cessation in the population. Further, given the growing popularity of e-cigarettes in the US, and considering smokers' very low interest in using traditional smoking cessation medications, future research into the comparative reach and effectiveness of different types of products will be

JAMA Network Open | Public Health
Association of e-Cigarette Use With Discontinuation of Cigarette Smoking