Association of naturalistic e-cigarette use and smoking cessation in US adults

Introduction

Prior trends of increasing smoking cessation rates in the United States have recently stalled.^1,2 Although data from the national study (2013-2021) indicate a modest increase in cessation in recent years,³ national surveillance data show that annual cessation rates have remained below 10% during 2020-2024. In 2022, 8.8% of adults who smoked successfully quit—slightly higher than in previous years—yet the majority of smokers continue to attempt quitting without success.^3,4 FDA-approved smoking cessation medications remain underutilized,⁴ perhaps due to limited accessibility and smokers' perceptions that these medications cannot suitably replace the nicotine and sensorimotor stimulation provided by cigarettes.⁵

E-cigarettes have demonstrated efficacy for smoking cessation in randomized trials.^6-9 However, some clinicians and public health officials remain hesitant to recommend e-cigarettes as cessation aids due to conflicting evidence from population-based observational studies regarding whether vaping provides real-world benefits or harms for quitting smoking.^10-17 Recent analyses of the nationally representative US Population Assessment of Tobacco and Health cohort similarly report mixed and modest effectiveness of e-cigarettes for cessation at the population level: one study reported that daily and non-daily e-cigarette users were no more likely to quit smoking than non-users, with non-daily use linked to reduced abstinence,¹⁸ whereas another found that e-cigarette users were more likely to discontinue cigarettes compared with non-users.³ Interpretation of these findings is complicated due to methodological challenges. Previous nationally representative longitudinal studies have relied on long inter-assessment intervals (eg, annual), which fail to capture the rapid cycle of quit attempts and relapses that characterize smokers' real-world experiences.^3,14,15,17 Consequently, existing observational studies on e-cigarette use and smoking cessation, though informative, remain vulnerable to several sources of confounding. Individuals with more severe, relapsing patterns of nicotine dependence and a preexisting vulnerability to cessation failure may be more likely to adopt any available cessation aid, including e-cigarettes. Conversely, those who are highly motivated to quit smoking may be more inclined to use e-cigarettes as a means to support cessation.^10-13,19

Building on these limitations, this observational, naturalistic study of US adult cigarette smokers estimated associations of e-cigarette vaping with two clinically distinct smoking cessation milestones—attaining abstinence and preventing relapse. A novel, nationally representative data source with (semi-)monthly repeated within-person assessments provided temporal precision and enabled a unique application of multilevel longitudinal analyses. This approach accounts for between-person stable (ie, time-invariant) confounders that may contribute to inter-individual variance in both e-cigarette use and smoking cessation outcomes.

Methods

Results

Discussion

In this naturalistic observational study of US adult smokers, associations of within-person changes in e-cigarette use with smoking outcomes diverged across unique smoking cessation milestones. Transitioning to daily vaping (vs no vaping) increased the odds of attaining smoking abstinence 2 weeks later. Conversely, among individuals who had achieved sustained abstinence, timepoints with no vaping were associated with lower relapse hazards than timepoints involving daily or non-daily vaping.

This study’s design addresses methodological concerns that might have contributed to inconsistent results in prior naturalistic studies, which have found positive,^10,16,17,19 null,^11,14 and inverse^12,13,15 associations between vaping and smoking cessation. Annual inter-assessment intervals in most nationally-presentative datasets lack the temporal precision required to capture cases in which vaping promotes quitting smoking but fails to sustain abstinence. Instances of e-cigarette adoption, smoking cessation, and relapse occurring within a few months may be entirely missed between annual assessments.^10-13,19 Additionally, previous observational studies are prone to confounders that contribute to between-person variability. Individuals at high risk for chronic nicotine dependence are more likely to adopt vaping, while those most committed to quitting may seek out any product to help them stop smoking.³⁷ High exposure to cigarette and e-cigarette marketing in certain neighborhoods may enable access to e-cigarettes while simultaneously tempting individuals to relapse back to smoking.^38,39 Although some studies have applied propensity score matching to mitigate confounding, matching approaches cannot fully eliminate bias.¹⁴

This study implemented multiple strategies to address confounding variables. Specifically, to account for the potential effects of smoking dependence severity—a factor closely linked to both smoking abstinence and relapse—we adjusted for participants’ past-week smoking frequency in our analyses. For smoking abstinence analyses, this adjustment was applied as a time-varying, time-lagged covariate across exposure timepoints. For smoking relapse analyses, it was incorporated as a baseline covariate measured prior to establishing smoking abstinence. Our models additionally included key factors influencing smoking cessation, such as demographics that are proxies for exposure to social determinants and cultural norms (eg, ethnicity/race), developmental stage (eg, age), familial support (eg, marital status), and behavioral health (eg, depression symptoms, drug use).^2,27 Chronological time was also adjusted for, accounting for the dynamic health behavior trends during the first two years of the COVID-19 pandemic.⁴⁰ While covariate adjustment partially attenuated estimates, associations between vaping and cessation outcomes remained statistically significant. Importantly, multilevel longitudinal analysis controlled for between-person variance in vaping to account for unmeasured stable confounders.⁴¹ Statistically significant within-person associations with abstinence and relapse, combined with null between-person estimates, suggest these findings were not driven by inherent differences between vapers and non-vapers. The associations are unlikely explained by time-varying confounders influencing short-term vaping and smoking decisions, as they persisted after adjusting for time-varying smoking frequency. Sensitivity analyses showed no evidence of effect modification by smoking frequency, indicating the associations are consistent across circumstances and populations that vary along the continuum of smoking.

Although time-varying residual confounding cannot be ruled out, it is plausible that e-cigarettes may support initial quit attempts but hinder long-term success when used after achieving sustained abstinence. Approximately half of US adult smokers do not attempt to quit each year.⁴² Some of these individuals might be motivated to stop but demoralized by past cessation failures or the perception that FDA-approved cessation medications are unappealing, ineffective, burdensome, or inaccessible.⁴³ Others might enjoy inhaling nicotine too much to consider quitting smoking.⁴³ For both groups, FDA-approved nicotine replacement therapies may fail to replicate the sensorimotor experience and rapid pulmonary nicotine delivery of combustible cigarettes.⁴⁴ E-cigarettes, when used daily, may provide enough satisfaction to motivate quit attempts in individuals who might not otherwise try. Similar to this study, some prior research has shown that daily (but not non-daily) vaping is associated with smoking abstinence.^10,13,16,19 A randomized trial also found that vaping significantly increases quit attempts, particularly in those initially less motivated to quit.⁸

Vaping after extended smoking abstinence may undermine long-term cessation by sustaining addiction mechanisms that typically diminish when all tobacco products are discontinued. Unlike nicotine replacement therapies, which deliver nicotine slowly via buccal mucosa or transdermally,⁴⁵ e-cigarettes replicate cigarettes’ addictive pharmacokinetics by rapidly delivering nicotine to the brain through pulmonary absorption.⁴⁶ This rapid delivery may prolong cycles of fluctuating blood nicotine levels that perpetuate tolerance, withdrawal, and compulsive nicotine-seeking behaviors.⁴⁷ E-cigarettes are often designed to mimic cigarettes, with aerosols resembling smoke and inhalation sounds imitating the crackling of drawing puffs from lit cigarettes.⁴⁸ These sensory attributes may maintain conditioned associations between smoking-related stimuli and nicotine reward, which typically fade during extended cessation. Post-quit vaping could prolong the conditioned urge to smoke and increase relapse risk when encountering smoking-related cues. Some observational studies suggest post-cessation vaping is associated with higher residual cigarette dependence symptoms⁴⁸ and long-term smoking relapse risk.⁴⁹ Although randomized clinical trials with intensive behavioral support show favorable long-term cessation outcomes with e-cigarettes,⁹ a “real-world” trial without such support appeared to find that the cessation-promoting effect of vaping weakened over time.⁸ Taken together, these mixed findings underscore the complexity of evaluating the public health impact of e-cigarettes. While our study separately examined their associations with abstinence and relapse, future research that integrates both initiation of abstinence and risk of relapse within a unified analytic framework will be important to more directly assess the overall net impact of e-cigarette use on smoking cessation outcomes.

Randomized trials with random assignments, free e-cigarette provisions, and narrow participant samples, do not replicate the self-motivated adoption of e-cigarettes in real-world commercial settings. The FDA Center for Tobacco Products has cited insufficient rigorous, nationally representative observational evidence of vaping’s cessation-promoting effects as a reason for denying e-cigarette marketing applications.⁵⁰ This study’s findings suggest that e-cigarettes—when used naturalistically within the unrestricted commercial market—may increase quit attempts among US adults. Clinicians and healthcare agencies, previously hesitant to promote vaping due to the limited ecological validity of randomized trials, may find this real-world evidence compelling.^51,52 This study offers new real-world evidence supporting e-cigarettes as a potential cessation aid for individuals who have not succeeded with established treatments. However, clinicians should advise patients that vaping after prolonged smoking abstinence may increase the risk of unfavorable long-term outcomes compared to complete discontinuation of all tobacco products.

Limitations

This study has limitations. First, it did not collect detailed data on smoking history, intensity, dependence, quit intentions, e-cigarette product characteristics (eg, flavors, nicotine concentration, liquid volume), or use of other cessation aids (eg, nicotine replacement therapy, other pharmacotherapies for smoking cessation). The potential influence of these unmeasured factors on our findings remains unclear and may introduce residual confounding. For example, successful quitting associated with daily e-cigarette use could partly reflect unmeasured individual characteristics, such as stronger motivation or commitment to cessation, rather than causal effects of vaping itself. Relatedly, because e-cigarette users were more likely to achieve short-term quitting, the relapse sample may have been systematically different, including a greater proportion of individuals with higher relapse risk.¹⁵ Furthermore, restricting the relapse analysis to individuals who had achieved sustained abstinence could induce selection bias (ie, collider stratification) if abstinence is a common effect of both e-cigarette use and another unmeasured variable that is a cause of relapse.⁵³ While our analytic approach adjusted for a number of variables associated with subsequent relapse (eg, mental health, other drug use), collider stratification bias may still occur due to unmeasured common causes of abstinence and relapse. In addition, without measures of smoking intensity, variations in cigarettes per day among daily smokers were also not addressed, and it remains uncertain whether some instances of cessation were unintentional. Second, some baseline covariates such as non-nicotine substance use and mental health may vary over time, and residual time-varying effects of these factors cannot be fully ruled out. Future research should further explore their dynamic roles in shaping the association between vaping and smoking cessation. Third, the within-participant association does not clarify the sequence of vaping onset and offset relative to smoking cessation. Moreover, the dataset did not capture discrete vaping cessation events; vaping was modeled solely as a past-week, time-varying exposure at each wave. As a result, our relapse analyses compared relapse risk across observed vaping states rather than the effect of a specific discontinuation event. Future research should distinguish between outcomes associated with post-quit vaping onset and those associated with sustained vaping initiated prior to quit attempts, ideally using designs that more precisely capture the timing of vaping initiation and cessation. Fourth, while the UAS data source uses probability sampling and weights to address biases, non-response may still limit generalizability. Finally, the COVID-19 pandemic and unique e-cigarette market and policy conditions during the study period may have influenced the results, although chronological time did not moderate results suggesting consistency of findings while the pandemic began to abate.

Conclusion

This observational naturalistic study in US adults who smoke found that daily vaping was associated with achieving smoking abstinence. However, among individuals who attained smoking abstinence, vaping during the post-abstinence period was associated with an increased risk of returning to smoking. Policymakers and clinicians should consider that vaping might be associated with different quit smoking outcomes depending on the stage of the smoking cessation process.

Supplementary Material

Supplementary material

Supplementary material is available at the American Journal of Epidemiology online.

Funding

This research was supported by the National Institute on Drug Abuse (NIDA) and the US Food and Drug Administration (FDA) under award R21DA058343; by the National Cancer Institute (NCI) and FDA under award U54180905; and by NIDA under awards R01DA058992, K24DA048160, K99DA058241, and K01DA058084. The funding agencies had no role in the design, conduct, analysis, interpretation, or writing of the study.

Conflict of interest

The authors have no conflicts of interest relevant to this article to disclose.

Disclaimer

The views expressed in this article are solely those of the authors and do not represent the views of the funding agencies or the Understanding America Study.

Data availability

Data from the Understanding America Study (UAS) are available to researchers upon request through the UAS data portal, subject to registration and approval of the UAS data use agreement.