Daily E-cigarette Use and the Surge in JUUL Sales: 2017–2019

Data source: The PATH Study is a large, nationally representative cohort of the US noninstitutionalized population aged 12-plus years.¹⁴  The study started with a screener survey of an address-based sample of US households, and then oversampled tobacco users, adolescents, young adults (aged 18–24 years), and African Americans. Separate adolescent (aged 12–17 years) and adult (aged 18-plus years) questionnaires in English and Spanish were used. At the first survey wave in 2013 to 2014 (labeled 2014 in this paper), 32 320 adults and 13 651 adolescents completed surveys. The cohort was reinterviewed each year through 2017, when a refreshment sample was recruited to reset the cohort size, adding 6064 adults and 5435 adolescents. Survey wave 5 was completed in 2019. The response rate for 2-year follow-up of the 2014 cohort was 78.2%, and for the 2017 cohort, it was 85.6%. Our population of interest consists of PATH respondents in each cohort who were aged 14 to 34 years (2014 cohort = 24 409; 2017 cohort = 20 161). We use restricted-use data available from the National Addiction and HIV Data Archive.¹⁵ 

Tobacco use measures: PATH respondents were shown images for each tobacco product and asked if they had ever used the product even 1 or 2 puffs/times. Adult daily use was assessed from the question, “Do you now [smoke/use product] every day, some days, or not at all?” On the adolescent questionnaire, respondents were asked “on how many of the last 30 days” they had used each tobacco product. We classified reported use on all 30 days as daily use. To obtain the e-cigarette brand used by daily vapers, on each survey, respondents were asked, “What brand of electronic nicotine product do you own?” with open-ended responses coded.

Dependence scores: The PATH Study uses the Wisconsin Inventory of Smoking Dependence Motives framework³  and, in random order, queries 7 domains on both the adolescent and the adult questionnaires separately for cigarettes and e-cigarettes: “I find myself reaching for [product] without thinking about;” “I frequently crave [product];” “My [product] use is out of control;” “I usually want to use [product] right after I wake up;” “Using [product] really helps me feel better if I’ve been feeling down;” “Using [product] helps me think better;” “I would feel alone without my [product].” The 5-point response scale ranged from 1 = “Not true of me at all” to 5 = “Extremely true of me.” Scores were added across the questions giving a range of 7 to 35, where 7 indicated no evidence of dependence.

Statistical analyses: All analyses were conducted in R (version 4.1.1). We used available single-wave weights for data from an individual survey and longitudinal weights for data involving multiple waves (such as incidence estimates).¹⁴  Variance estimates for confidence intervals (CIs) were calculated considering both population weights and replicate weights, with the latter constructed from a balanced repeated replications procedure with Fay adjustment (ρ = 0.3).¹⁶  Sample characteristics were explored using weighted proportions with 95% confidence limits.

We stratified each cohort by baseline age group (ages 14–17, 18–21, 22–25, 26–29, and 30–34). Within each age group, we calculated the prevalence of ever-use and daily use of any tobacco, cigarettes, and e-cigarettes. Then, for each cohort, we calculated the 2-year incidence of ever-use, among those who were baseline never tobacco users. We also calculated the rate of progression (over 2 years) to daily tobacco use, daily cigarette smoking, and daily e-cigarette vaping, among baseline nondaily ever tobacco users. We use nonoverlapping 95% CIs to identify significant differences in the above statistics.

There were no differences between the weighted estimates from the 2 cohorts on sex or race/ethnicity distribution at baseline (Table 1). The 2017 cohort had fewer respondents aged 14 to 17 years (14.9%, 95% CI: 14.6%–15.1%) than the 2014 cohort (18.5%, 95% CI: 18.3%–18.7%), but there were no other age group differences across the cohorts. The 2017 cohort had fewer never tobacco users (34.1% 95% CI: 32.4%–35.8%) than the 2014 cohort (38.9%, 95% CI: 37.3%–40.4%), and correspondingly more ever, but not daily, tobacco users (50.2%, 95% CI: 48.6%–51.9%), than the 2014 cohort (44.5%, 95% CI: 43.2%–45.8%). Approximately 16% in each cohort were current daily tobacco users. We examined e-cigarette use, and none of the 2014 cohort were daily JUUL e-cigarette users in 2016 (survey wave 3).

In the 2014 cohort, the proportion of baseline never-users who tried cigarette smoking over the following 2 years was similar for those aged 14 to 17 years and 18 to 21 years, at 10% (Fig 1A). Estimates for older age groups were imprecise (with large confidence limits) but were less than half that of the 14- to 21-year-old age groups. In the 2017 cohort, among those aged 14 to 17 years, the proportion who tried cigarette smoking was 33% lower than the 2014 cohort at 6.7% (95% CI: 6.0%–7.4%) and, for those aged 18 to 21 years, it was 49% lower at 5.3% (95% CI: 4.3%–6.5%).

First use of e-cigarettes was highest among the youngest age group (14–17 years) for each cohort, at 22.0% (95% CI: 20.9%–23.1%) in the 2014 cohort and 24.7% (95% CI: 23.5%–25.9%) in the 2017 cohort, an increase of 12.3% (Fig 1B). Compared with those aged 14 to 17 years, first experimentation with e-cigarettes was much lower among those aged 18 to 21 years at 15.8% (95% CI: 13.4%–18.5%) in the 2014 cohort and even lower still in the 2017 cohort at 12.2% (95% CI: 10.5%–14.0%).

Considering incidence of first experimentation with any tobacco product, there was no between-cohort difference among those aged 14 to 17 years: 2014 = 29.8% (95% CI: 28.5%–31.1%); 2017 = 28.7% (95% CI: 27.5%–30.1%). However, a major decline was noted among those aged 18 to 21 years: 2014 = 27.7% (95% CI: 24.4%–31.3%); 2017 = 19.0% (95% CI: 16.8%–21.4%).

Over the 2-year follow-up, 4.9% (95% CI: 4.5%–5.4%) of baseline nondaily tobacco users in the 2014 cohort progressed to daily use, compared with 5.3% (95% CI: 4.9%–5.8%) in the 2017 cohort, with 64.6% of this increase coming from those aged 14 to 17 years (Table 2). Indeed, only those who were aged 14 to 17 years at baseline had a higher progression to daily use in the 2017 cohort compared with the 2014 cohort (2014 cohort = 4.8%, 95% CI: 4.3%–5.5%; 2017 cohort = 6.3%, 95% CI: 5.8%–7.0%): thus, there were 1 030 957 US adolescents aged 14 to 17 years in 2017 who became new daily tobacco users by 2019. In the 2019 survey, there were 1 667 826 new daily tobacco users who were aged <21 years. Assuming an equal distribution across the year, this translates to 2284 new daily tobacco users each day between 2017 and 2019, which is approximating the estimate of new established smokers in 1989 to 1993 (Supplemental Information).

In the 2014 cohort, 2.5% (95% CI: 2.1%–3.0%) of baseline ever-users aged 14 to 17 years progressed to new daily cigarette smoking over the 2-year follow-up, and the progression rate was 3.9% (95% CI: 3.1%–4.8%) for those aged 18 to 29 years (Fig 2A). For the 2017 cohort, progression to daily cigarette smoking was 55% lower than the 2014 cohort for both those aged 14 to 17 years (1.1%, 95% CI: 0.9%–1.5%) and for those aged 18 to 21 years (1.7%, 95% CI: 1.54%–2.2%).

Although there was little progression to daily e-cigarette use in the 2014 cohort (all groups <2%), this progression rate was much higher in the 2017 cohort (Fig 2B). Among those aged 14 to 17 years, the rate was 3.6-fold higher at 4.8% (95% CI: 4.4%–5.2%) and among those aged 18 to 21 years, the progression rate was 3.4 times higher at 4.1% (95% CI: 3.5%–4.8%). Among those aged 22 to 25 years, the progression rate was 2.1% (95% CI: 1.5%–2.7%).

Almost half (48.9%) of all the new daily tobacco users in the 2017 cohort were daily e-cigarette vapers (Table 3), with the distribution skewed to the younger age groups. For those aged 14 to 17 years at baseline, 75.4% of all new daily tobacco users were e-cigarette vapers; for those aged 18 to 21 years at baseline, 64.9% were e-cigarette vapers. Both the e-cigarette dependence score (Cronbach’s α = 0.89, SE = 0.23) and the cigarette dependence score (Cronbach’s α = 0.86, SE = 0.14) had good internal consistency. Among daily vapers, in 2019, the mean e-cigarette dependence score for those aged 14 to 17 years in 2017 was 17.4 (95% CI:16.5–18.3), with the lower-bound confidence limit more than double the score, indicating no dependence. The mean dependence score in this youngest age group was similar to adult scores in all under age 35 years age groups, and these scores were similar to the dependence scores for those of the same age who became daily smokers (Supplemental Information).

In 2019, there were >1 million daily vapers among US residents under the age of 21 years. The majority (56.3%, 95% CI: 50.4%–62.0%) used JUUL products, which translates to 600 000 daily JUUL users (Table 4). The proportion of daily e-cigarette vapers using JUUL was lower among those aged 21 to 24 years (42.9%, 95% CI: 36.0%–50.1%) and those who were aged >25 years (22.6%, 95% CI: 17.5%–28.7%). In 2019, daily e-cigarette vapers aged <21 years were 2.5 times more likely to be JUUL users than young adults aged 25-plus years.

We identified 2 PATH Study cohorts aged <35 years, before and after the surge in US e-cigarette sales that started in 2017. There were no new daily JUUL e-cigarette vapers in the 2-year follow-up of our comparison group: the 2014 cohort. There were an additional 395 597 new daily tobacco users in the 2-year follow-up of the 2017 cohort compared with the 2014 cohort, two-thirds of whom were in the youngest age group (14–17 years at baseline). Indeed, it was only this youngest age group that had a significant increase in progression to daily tobacco use between the 2 cohorts. In the 2 years to 2019, there were 2284 per day, <21-year-old new daily tobacco users, a rate that approaches the levels observed in the 1990s before the 1998 additional restrictions on cigarette advertising and promotions.¹⁷  However, the vast majority of young new daily tobacco users were daily vapers, a clear difference from the 2014 cohort. In 2019, there were 600 000 daily JUUL vapers aged <21 years.

Previously, adolescent JUUL users have been reported to have some symptoms of nicotine dependence.^18,19  The PATH Study used the Wisconsin Inventory of Smoking Dependence Motives framework³  and questioned 7 dependence domains for each tobacco product. We report that <21-year-old, recent daily e-cigarette vapers in 2019 had similar dependence scores to older recent daily vapers. Additionally, they had dependence scores that were similar to those for new daily cigarette smokers.

There are at least 3 major influencers on this substantial increase in daily e-cigarette vaping seen between our 2 cohorts:

1. there was an ongoing innovative social media marketing campaign that was attractive to youth;²⁰ 

2. the new normal e-cigarettes used nicotine salt technology that made higher nicotine concentration aerosols less aversive, thus allowing users to self-administer larger nicotine doses with each puff;²¹  and

3. experimentation with e-cigarettes typically occurred with products with fruit/candy flavors.²² 

Data collection for our study was completed in 2019. In October 2019, JUUL voluntarily suspended sales of all e-cigarette flavors except for menthol and tobacco,²³  just before the Food and Drug Administration’s (FDA’s) announcement on its enforcement policies for unauthorized flavored, cartridge-based e-cigarettes.¹³  However, the FDA’s announcement did not include disposable e-cigarettes,²⁴  e-liquids, or refillable devices.²⁵  As a result, there was another major market shift to e-cigarettes that had these flavored options. In 2021, there were an estimated 1.7 million high school vapers, with 85% using flavored products. Over half these vapers used disposable devices and JUUL was no longer the market leader.²⁶ 

Our data suggest that there may be a new generation of daily tobacco users who did not become dependent on combustible tobacco. Hopefully, they may never switch to cigarette smoking, thus avoiding many of the health consequences, although e-cigarettes themselves are not harmless.²⁷  However, tobacco industry giants have copied the JUUL social media model, going further to invest large sums on social media influencers who can produce their own viral videos that market tobacco.²⁸  Further, there are recent reports that smoking is making a comeback among trendy youth²⁹  and, in 2020, annual cigarette sales increased for the first time in 2 decades.³⁰  All this suggests the possibility that dependent e-cigarette users could switch to cigarette smoking as young adults. Further follow-up of the PATH Study is needed to document whether dependent e-cigarette vaping becomes a stable adult behavior.

We confirmed results from earlier reports of the cross-sectional National Youth Tobacco Survey,^26,31  that the follow-up in 2019 showed evidence of a considerable decline in experimentation with e-cigarettes for those aged 18 to 21 years at baseline, possibly linked to concerns relating to the 2019 E-cigarette or Vaping Use-Associated Lung Injury epidemic.³²  However, there was no such decline seen among those aged 14 to 17 years in 2017. Compared with the 2014 cohort, the 2 youngest age groups in the 2017 cohort had much lower experimentation rates with cigarettes. However, only the 14- to 17-year-olds increased their experimentation with e-cigarettes.

A strength of this paper is the use of 2 cohorts from the PATH Study that allows nationally representative estimates of tobacco use. We targeted daily tobacco use and demonstrated that this daily e-cigarette use was associated with significant dependence levels. We identified 2 cohorts and validated 1 as occurring before the JUUL-related surge in e-cigarette sales. Each cohort had large samples who were in the age window for initiation of tobacco use.³³  The PATH Study includes state-of-the-art measurement of product use on each survey, as well as a theory-driven, psychometrically validated dependence scale for each product used. It also collects and analyzes biological samples on a large subsample that have been used previously to demonstrate the validity of self-reported smoking behavior,³⁴  as well as the PATH-dependence scales.³⁵  A limitation is that these cohorts were not contemporaneous but were followed in sequential time periods, thus, we were unable to rule out the influence of other time-based variables. Additionally, other unmeasured confounders could have influenced the population estimates.

The surge in sales of JUUL e-cigarette from 2017 to 2019 was accompanied by increased daily e-cigarette use that was most marked among adolescents aged 14 to 17 years at baseline. Members of this youngest age group who became daily e-cigarette users had dependence scores that were not different to older adults nor different from adult daily cigarette smokers. This recent upswing in tobacco use in the young has resulted in initiation levels that are somewhat similar to those seen in the early 1990s, before the introduction of many effective tobacco control measures. In 2019, an estimated 600 000 Americans aged <21 years used high-nicotine JUUL products daily and these would now be below the legal age to purchase tobacco in the United States.³⁶  The FDA has responded to this e-cigarette epidemic by reducing the availability of flavored additives to e-cigarettes, presumably focused on reducing experimentation. It has not limited the amount of nicotine allowed in e-cigarettes, as other countries have done.³⁷  The large increase in daily use among US adolescents could presage future health consequences and needs urgent additional action from the regulatory body.

CI

confidence interval

FDA

Food and Drug Administration

PATH

Population Assessment of Tobacco and Health