E-cigarette use and smoking reduction or cessation in the 2010/2011 TUS-CPS longitudinal cohort

This is the first longitudinal study to use a nationally representative sample of U.S. smokers to provide evidence regarding the impact of e-cigarettes on smoking behavior at the U.S. population level. This study, conducted before e-cigarettes were widely used, did not find evidence that the ever-use of e-cigarettes as a cessation aid among these early adopters was associated with increased 30+ day smoking cessation at one-year follow-up. Further, reported ever use of e-cigarettes by heavier smokers was not associated with reduced cigarette consumption at follow-up. However, among lighter smokers, those who had ever used e-cigarettes for cessation had stable consumption levels, whereas all other light smokers showed increased consumption at follow-up, although these differences were not statistically significant.

In this study, those who had ever used e-cigarettes to quit or had used pharmaceutical aids on their most recent quit attempt equally appeared to be heavier, more dependent smokers. Such a finding is well known for pharmaceutical aids [28]. Thus, it was expected that success rates for e-cigarette users and for pharmaceutical aid users might be less than for non-users in this population study, as we found to be the case in both adjusted and unadjusted analysis. However, the data also suggest that cessation rates among those who had used e-cigarettes for quitting may be lower than cessation rates seen with use of a pharmaceutical aid.

While our study is the first which can provide nationally representative estimates for the U.S and focuses on early adopters who used first generation e-cigarettes, it supports results from the other published U.S. population-based longitudinal cohort study [19], as well as the summary results in meta-analysis across all 20 studies published as of this writing [8]. Most recently, a Canadian study of smokers enrolled in cessation assistance in the primary care setting showed association of e-cigarette use with poor cessation outcomes [29], consistent with our results. In contrast, relying on samples in two U.S. metropolitan areas Biener et al. [23] reported a strong association of smoking cessation with daily use of e-cigarettes for 30+ days, as compared to non-users. However, most (66 %) of those reporting daily e-cigarette use had adopted e-cigarettes because they wanted to quit smoking, whereas non-users did not report the same motivation, and the resulting confounding by reverse causation can account for the reported association. Indeed, to control for such confounding in our study, we compared use and non-use of cessation aids only among those who had tried to quit, as in prior work [26, 27, 30, 31]; without this careful choice of comparison group, we are able to replicate a spurious association which is similar in magnitude to that reported by Biener et al. [23] (data available upon request). However, our study is unable to address whether extensive and heavy use of e-cigarettes might facilitate cessation, as we were unable to find such users in the national population. As e-cigarette use moves beyond early adopters into the population and product innovation continues, such heavy use may become more common allowing the association with cessation to be further tested [6, 32]. Our observation that use of e-cigarettes by light smokers appeared to prevent increased consumption by follow-up was not hypothesized prior to this study. Thus, it is a hypothesis-generating observation in need of replication.

Our study has a number of limitations. The survey was conducted when e-cigarette use was uncommon but rapidly increasing in the US. Accordingly, the e-cigarette question at follow-up only sought information on ever-use, as well as whether the use was for cessation. Ideally, the usage report would have been linked directly to the most recent quit attempt. However, in 2009 ever use of e-cigarettes among smokers was less than 2 %, and 40 % of ever-users also reported past 30-day use [4, 33]. Thus we expect that most reported use fell within the past year, as did the quit attempt. However, these findings need to be replicated in other nationally representative longitudinal studies. Although the diversity of e-cigarette products was much smaller in 2010–2011 than at present [32], the TUS-CPS treated e-cigarettes as a single homogeneous product and there was no information on intensity of use, whereas others have noted substantial variability in nicotine levels, patterns of use and customer experiences [34–36]. In the future it will be important to investigate patterns of cigarette use among sub-populations of e-cigarette users.