First E-Cigarette Flavor and Device Type Used: Associations With Vaping Persistence, Frequency, and Dependence in Young Adults

Abstract

Introduction

Type of e-cigarette flavoring and device during first use might differentiate later e-cigarette use and dependence. This retrospective cross-sectional study examined associations of recalled first nicotine vaping device and flavor used with current vaping frequency/dependence.

Aims and Methods

A young adult cohort from Los Angeles, California, USA completed web-based surveys (N = 2553). Using cross-sectional data from 971 reporting ever vaping nicotine, multivariable hurdle regressions tested associations between recalled first flavor (fruit/sweet, menthol/mint, other) and device (Juul, disposable, mod, box, pod, pen, other) vaped with past-30-day vaping status (yes/no) and frequency (1–30 days), and with any vaping dependence symptoms (yes/no) and count (1–10 symptoms).

Results

The most common first-flavor was sweet (71%); the most common first-device was a vape pen (37%), then Juul (22%). First-flavor of mint/menthol (vs. other; adjusted odds ratio [AOR]: 2.22[95% CI = 1.16 to 4.25]), and first-device mod (AOR = 2.40[95% CI = 1.34 to 4.31]) and non-Juul pod (2.64[95% CI = 1.41 to 4.92]) (vs. pen) were associated with past-30-day vaping, and twice as many vaping days (adjusted rate ratios [ARRs] range: 1.96–2.12; ps < .05). First flavor of mint/menthol (vs. other; AOR: 1.95[95% CI = 1.003 to 3.79) and first device mod, box, non-Juul pod, and other (AORs range: 2.36–4.01; ps < .05) were associated with nicotine dependence. First device Juul, mod, box, and non-Juul pod were also associated with more dependence symptoms (ARRs range:1.38–1.59; ps < .05).

Conclusions

Exposure to mint/menthol and certain devices (mod, box, Juul, and non-Juul pods) at first e-cigarette use may be associated with more frequent e-cigarette use and nicotine dependence symptoms in young adulthood. Mint/menthol and certain devices warrant consideration in regulation of e-cigarettes based on product characteristics.

Implications

Characteristics (flavor and device type) of first e-cigarette product used were associated with higher usage and more dependence. Pending replication with prospective designs, the findings suggest certain flavors (mint/menthol) and devices (pods, mods) merit consideration in regulation because of their possible link with continued use and dependence among young people.

Introduction

The prevalence of US youth and young adult e-cigarette use has increased over the past decade,^1,2 with a significant proportion of young people who try e-cigarettes becoming daily users.^1,3 This trend has prompted concerns over the development of nicotine dependence and other adverse outcomes associated with nicotine exposure via e-cigarettes, such as nicotine-mediated neurodevelopmental problems.⁴

In many countries across the globe, national public health agencies have the authority to regulate e-cigarette product characteristics, such as device type and flavors, if they are deemed to increase the risk for persistent vaping and addiction among young people. E-cigarette device types that deliver more nicotine could hasten the development of nicotine dependence, as nicotine is the primary reinforcing compound that sustains tobacco product use.⁵ Nicotine delivery in e-cigarettes is a function of the electrical power of the device and the nicotine concentration of the solution used in the device.⁶ Larger mod and tank-style devices tend to have high power output and can be loaded with low or high nicotine concentration solutions. Small-sized pod-style e-cigarette devices, such as Juul, are widely used among young people⁷ and often include high nicotine concentration solutions with low power.⁸ Pod-style devices are predominately available in nicotine salt formations. Evidence suggests that e-cigarettes with protonated salt (vs. standard freebase) nicotine reduce the harshness of aerosol for young adult never smokers,⁹ which may make it easier for nicotine-naive users to inhale nicotine more deeply. A cross-sectional study found high levels of nicotine dependence among young pod users, but the directionality of these associations is unclear.¹⁰

Flavors are another e-cigarette product characteristic that could accelerate vaping patterns and nicotine dependence risk. There is evidence that menthol and certain fruit flavors may increase nicotine’s bioavailability or pharmacological reinforcing effects^11,12 and suppress nicotine’s aversive sensory effects in e-cigarette aerosol among young adults.¹³ One prospective study has found an association of use of fruit and candy flavors with both continued use and a higher number of puffs per vaping session¹⁴; another found the initial use of any flavor to be associated with progression to current vaping and higher frequency of vaping.¹⁵ These studies did not examine dependence as an outcome and utilized a broad binary flavor definition, leaving unclear whether exposure to certain flavored e-cigarettes confer risk and ultimately predict dependence as a distal endpoint.

The present retrospective cross-sectional study examined a subset of a young adult cohort who had tried e-cigarettes at least once in their life. Respondents were asked to remember what type of device and flavor they used during their first vaping experience. We determined persistence of use by assessing respondents’ current e-cigarette use status at the time of the study, as well as current e-cigarette use frequency and dependence. The study aimed to estimate associations of first-flavor and device vaped with current e-cigarette use and dependence.

Methods

Sample Derivation

Data were drawn from the Happiness and Health Study,¹⁶ a prospective cohort survey of individuals originally recruited during ninth grade in 2013 as students enrolled in one of 10 participating high schools in Los Angeles County, CA in (n = 3396). Participants provided assent and parental consent to be surveyed (semi)annually about health behaviors and were reconsented as adults. The University of Southern California Institutional Review Board approved this study.

This study uses data from the Wave 9 survey conducted between October 2018 and November 2019 when the cohort were young adults. During this wave, 2553 (75%) of all respondents that originally enrolled in cohort were surveyed. Of these, 1164 reported ever use of >1 electronic vaping devices not for cannabis (i.e., e-cigarette, Juul, other device) and received a follow-up question assessing self-report of specific device type (see below), of whom 971 reported using one of the seven device types and constituted this study’s analytic sample.

Outcomes

Past 30-day Vaping

Adolescents were asked to report the number of past-month days they used (1) an e-cigarette with nicotine, (2) Juul, and (3) another unspecified electronic vaping device, in three separate items each with seven response categories. We converted responses into a count as in previous work (0, 1–2 [recoded:1], 3–5 [recoded:4], 6–9 [recoded:7], 10–19 [recoded:14], 20–29 [recoded:24], all 30 days)¹⁷ and took the maximum score across the three items (range: 0–30).

Vaping Dependence

Vaping dependence was assessed using a variant of the Hooked on Nicotine Checklist (HONC),¹⁸ adapted to assess e-cigarette dependence, which has shown good psychometric properties and predictive validity in previous work.¹⁹ HONC (yes/no) items assess craving (six items, e.g., “Ever tried to quit [Vaping/Juuling], but couldn’t?”) and withdrawal symptoms (four items, e.g., “Did you feel more irritable because you couldn’t [Vape/Juul]?”). Because youth use different terminology to describe nicotine vaping (e-cigarettes or Juul), subjects were asked to report dependence symptoms from items with language keyed to represent the type of vaping device they self-reported using—either e-cigarettes or Juul. Respondents who reported using both e-cigarettes and Juul received both HONC versions. We computed a pro-rated sum (0–10), given that a subject responded to at least 4 of the 10 items on a checklist and for subjects that responded to both checklists, we used the larger of the two sum scores. Eight percent of subjects had missing data for 1–8 items.

Exposures

Respondents were instructed to think back to their first vaping experience and queried, “Which flavor did you try first?” Responses included “Flavorless,” “Tobacco flavored,” “Menthol/Mint,” “Fruit/Sweet (fruit, candy, dessert, buttery, etc.),” or “Other.” Flavorless, tobacco flavored, and other responses were combined into a single reference category because the policy significance of their comparison to fruit/sweet and menthol/mint, resulting in a trichotomous first flavor variable. The question, “Which device did you first use?” had responses: “Disposable device,” “Vape pen or pen-like rechargeable device (such as eGO or small startup kit),” “Mod or mech-mod rechargeable device,” “Box mod,” “Juul,” “Other [non-Juul] pod mod (e.g., Suorin),” “Other type of electronic device.” Responses were directly translated into a 7-level categorical variable, with vape pen as reference category because it had the highest prevalence.

Covariates

Because individuals experiencing sociodemographic or peer environments, who started vaping earlier (prior to the advent of certain devices), who use other tobacco products, or who have certain psychological characteristics might be more likely to be exposed to certain products,^20–24 we included the following a priori covariates: sociodemographics (age, gender, race/ethnicity, parent education, family finances, academic attainment, and living with biological parents), depressive symptoms measured by the 10-item version of the Center for Epidemiologic Studies Depression Scale,²⁵ a 4-item version of sensation seeking scale of the impulsive behavior scale,²⁶ number of five best friends that vape, time of vaping onset (calculated as the first survey wave participants reported ever vaping, range 1 [beginning of ninth grade] to 9 [post-high school]), and use of combustible tobacco products (see Table 1 for responses).

Table 1.

Descriptive Statistics of Sample Outcome and Covariates

Variable	Percent	Mean (SD)	Missing N
Outcomes
Past-30-day e-cigarette use	59%	6.7 (10.4) days	35
E-cigarette dependence	28%	0.8 (1.4) symptoms	11
Covariates
Personal and friend tobacco status
Combustible tobacco use			0
Never	37%
≥ 6 month ago	22%
1–6 month ago	14%
Past month	26%
Vaping onset time (survey wave)		4.7 (3.3)	0
No. 5 best friends that vape		3.0 (1.8)	41
Sociodemographic characteristics
Age, years		19.7 (0.5)	3
Female gender	44%		0
Race/ethnicity			2
Hispanic	54%
Asian/Pacific Islanders	20%
White	18%
Black	3%
Other	5%
Parental education			113
≤8th	3%
Some HS	9%
HS diploma	14%
Some college	20%
College graduate	34%
Advanced degree	19%
Family finances			55
Pretty well off	21%
Above Average	49%
Poor	17%
Varied	13%
Education attainment			55
<HS or GED	3%
HS diploma	38%
Some college	55%
2-year degree+	5%
Raised in two-parent household	64%		6
Psychological risk factors
Depression symptomsa		1.0 (0.6)	59
Sensation seekingb		1.7 (0.7)	59

N = 971.

GED = general education development, HS = high school.

^aBased on mean of Center for Epidemiologic studies depression scale 10-item version, with each item rated week on a 4-point Likert scale for past week symptoms, ranging from 0 (rarely or none of the time; 0–1 days) to 3 (most or all of the time; 5–7 days).

^bBased on mean of 4-item sensation seeking subscale of the UPPS-P Impulsive Behavior Scale. These items are rated on a 4-point Likert scale, ranging from 1 (disagree strongly) to 4 (agree strongly).

Data Analysis

After descriptive analyses, we fit two regression models, one to test associations of first flavor and device with past 30-day vaping outcomes and another for associations with e-cigarette dependence symptom outcomes. Both models co-adjusted for first device type (use of vape-pen as the reference) and flavor (use of tobacco/other flavors as the reference) as simultaneous regressors and all covariates mentioned above. A sensitivity analysis by excluding other flavor from the reference category was also conducted. We used hurdle regression models from the R program pscl package²⁷ for zero-inflated count outcomes. Each model splits outcomes into (1) hurdle—a binomial logistic regression of >1 versus 0 values; and (2) count—a truncated negative binomial regression of the nonzero values. Both parts are estimated simultaneously. Hurdle results are adjusted odds ratios (AORs) for any vaping days (>1 vs. 0) or any dependence symptoms (>1 vs. 0). Count results are adjusted rate ratios (ARR) for the mean number of past-month vaping days among current vapers (1–30) or dependence symptom count score among symptom-positive respondents (1–10 symptoms). We used multiple imputations to manage nonresponse using variables across all waves creating 100 imputed datasets using chained equations via the R-program MICE package²⁸ with 25 iterations with predictive mean matching for most quantitative variables and multinomial logistic regression for categorical variables. Standard methods were used to combine sets of results from each imputed dataset to a single set of final results.²⁹ 95% confidence intervals (CIs) surround estimates. Statistical significance was 0.05.

Results

Sample Description

Subjects were 19–20 years old (Mean[SD] = 19.7[0.5] years), 56% were male, and the racially diverse sample included Hispanics (54%), Asian/Pacific Islanders (20%), White (18%), and Non-Hispanic Blacks (3%) (Table 1). There were a range of socioeconomic backgrounds and academic attainment, with 38% having a high school diploma, and 55% with some college. The mean wave at which respondents reported initiation of e-cigarette vaping was 4.7 (SD = 3.3), occurring when most respondents were 15–16 years old. 37% had never used combustible tobacco, and 26% had used it in the past month. Descriptive statistics for depressive symptoms and other covariates are reported in Table 1.

The prevalence of past-30-day vaping was 59%, with a mean number of days of 6.6 (SD = 10.3). The prevalence of any dependence symptoms was 28% with a mean dependence symptom score of 1.2 (SD = 2.4). Most (71%) reported using a fruit/sweet flavor in their first device, 19% used menthol/mint, and 10% used tobacco/flavorless/other. The most common first device was a pen (37%), followed by Juul (22%), disposable (12%), mod (10%), box (10%), non-Juul pod (8%), and other (1%).

Association of First Flavor and Device With Past 30-Day E-Cigarette Use

Seventy-nine percent of those beginning with menthol/mint first flavor were past 30-day vapers at the time of survey during young adulthood, compared to 52% for those beginning with tobacco/flavorless/other and 54% who first used fruit/sweet (Figure 1A). Among past 30-day vapers, those reporting menthol/mint as first flavor (Figure 1B) reported almost double the mean days vaped (Mean: 6.9[95% CI = 4.2 to 11.3]) as compared to tobacco/other flavors (Mean: 3.3[95% CI = 1.7 to 6.4]), with first fruit/sweet falling in between (Mean: 4.5[95% CI = 2.9 to 7.1]). In the multivariable covariate-adjusted analysis (Table 2), menthol/mint versus tobacco/other flavor use was significantly associated with both a greater likelihood of any past 30-day vaping (AOR: 2.22[95% CI = 1.16 to 4.25]) and higher vaping frequency among past 30-day vapers (ARR = 2.12 [1.27, 3.56]). Fruit/sweet versus tobacco/other was not significantly associated with past-30-day vaping odds or frequency.

Table 2.

Association of First Vaping Flavor and Device With Past 30-Day Vaping^a,b

	Any current vaping		Number of days (1–30)
	Adjusted odds ratio (95% CI)	p	Adjusted rate ratio (95% CI)	p
Flavorc
Tobacco/other (n = 93)	Reference		Reference
Sweet/fruity (n = 693)	0.93 (0.54 to 1.59)	.73	1.40 (0.87 to 2.25)	.17
Menthol/mint (n = 185)	2.22 (1.16 to 4.25)	.02	2.12 (1.27 to 3.56)	.004
Device
Vape Pen (n = 361)	Reference		Reference
Juul (n = 210)	1.29 (0.84 to 1.98)	.25	1.18 (0.81 to 1.72)	.40
Disposable (n = 115)	1.00 (0.60 to 1.69)	.99	1.39 (0.88 to 2.20)	.16
Mod (n = 99)	2.40 (1.34 to 4.31)	.003	2.10 (1.37 to 3.23)	.001
Box (n = 93)	1.67 (0.92 to 3.02)	.09	1.76 (1.13 to 2.73)	.01
Non-Juul pod style (n = 81)	2.64 (1.41 to 4.92)	.002	1.96 (1.26 to 3.04)	.003
Other (n = 13)	2.15 (0.54 to 8.56)	.28	1.98 (0.70 to 5.54)	.20

CI = confidence interval.

^aTotal N = 971. Results based on 100 multiply imputed datasets.

^bAdjusted for age, gender, race/ethnicity, parent education, family income, bioparent(s) household, highest level of school attained, vaping onset, friend vaping, combustible tobacco use, depression symptoms, and sensation seeking.

^cTobacco/other = flavorless (n = 46) + tobacco (n = 23) + other (n = 24).

Figure 1.

Unadjusted past-30 day e-cigarette use outcome, stratified by first vaping device/flavor.(A) Any current vaping. (B) Vaping frequency among current vapers. Error bars are 95% confidence intervals.

Forty-eight percent of those beginning with a pen device were past 30-day vapers, compared to 73%–76% those beginning with mod, box, or non-Juul pod devices (Figure 1A). Among those that were currently using an e-cigarette at the time of survey (Figure 1B), first device mod, box, or non-Juul pod reported approximately double the number of mean past-month vaping days (7.6–8.6 days) than first device pen (3.4 days). In multivariable analyses (Table 2), compared to those reporting beginning with a pen device, using a mod or non-Juul pod-style for first device was associated with more than doubling of the current vaping odds (AORs: 2.40–2.64; ps <.05) and doubling (or near doubling) of the number of past-30-day vaping days among current vapers (ARRs: 1.96–2.10; ps <.05). Using a box as the first device was associated only with higher vaping frequency among current e-cigarette users (ARR: 1.76[95% CI = 1.13 to 2.73]) but not odds of any vaping. Use of a disposable or Juul device first was not significantly associated with either higher odds of past-30-day vaping or higher vaping frequency (Table 2).

In the sensitivity analysis by excluding other flavor from the reference group (Supplementary Table 1), the results are similar to the main analysis.

Association of First Flavor and Device With Vaping Dependence

Forty-six percent of those beginning with menthol/mint first flavor reported >1 dependence symptoms, compared to 22% beginning with tobacco/flavorless/other flavors and 24% for those starting with fruit/sweet flavors (Figure 2A). Among those with >1 dependence symptoms, mean symptom counts were not different for first flavor menthol/mint compared to tobacco/other (Figure 2B). In the multivariable analysis (Table 3), menthol/mint (vs. tobacco/other) first flavor was associated with a higher likelihood of experiencing >1 vs. 0 dependence symptoms (AOR: 1.95[95% CI = 1.003 to 3.79]) but not higher symptom count among those with >1 symptoms. Fruit/sweet flavor was not associated with dependence symptom status or count (Table 3).

Table 3.

Association of First Vaping Flavor and Device With E-Cigarette Dependence^a Symptoms^b,c

	Any dependence symptoms		Number of dependence symptoms (1–0)
	Adjusted odds ratio	p	Adjusted rate ratio	p
Flavord
Tobacco/other (n = 93)	Reference		Reference
Sweet/fruity (n = 693)	1.00 (0.55, 1.83)	1.00	0.84 (0.66, 1.06)	.14
Menthol/mint (n = 185)	1.95 (1.003, 3.79)	.049	0.86 (0.67, 1.10)	.22
Device
Vape pen (n = 361)	Reference		Reference
Juul (n = 210)	1.10 (0.68, 1.78)	.70	1.59 (1.29, 1.96)	.0001
Disposable (n = 115)	1.04 (0.59, 1.81)	.90	1.17 (0.90, 1.52)	.24
Mod (n = 99)	2.36 (1.36, 4.10)	.002	1.38 (1.10, 1.73)	.005
Box (n = 93)	3.42 (1.97, 5.93)	.000	1.45 (1.18, 1.79)	.0004
Non-Juul pod style (n = 81)	2.82 (1.57, 5.07)	.001	1.48 (1.17, 1.87)	.001
Other (n = 13)	4.01 (1.04, 15.5)	.045	1.34 (0.87, 2.06)	.18

^aBased on a self-reported checklist of 10 vaping product dependence symptoms reflecting loss of control over use, craving or urge, or withdrawal symptoms while abstinent and administered separately for devices of e-cigarette and Juul.

^bTotal N = 971. Results based on 100 multiply imputed datasets

^cModel adjusts for age, gender, race/ethnicity, parent education, family income, bioparent(s) household, highest level of school attained, vaping onset, friend vaping, combustible tobacco use, depression symptoms, and sensation seeking.

^dTobacco/other = flavorless (n = 46) + tobacco (n = 23) + other (n = 24).

Figure 2.

Unadjusted e-cigarette dependence outcome, stratified by first vaping device/flavor.(A) One or more symptoms. (B) Number of symptoms amongst those with at least one symptom. Error bars are 95% confidence intervals.

Whereas 18% of those beginning with a pen device reported any dependence symptoms, 41%–53% of those beginning with box, mod, or non-Juul pod devices experienced >1 symptoms (Figure 2A). In those with >1 dependence symptom, Juul, first device box, mod, and non-Juul pod also had higher mean number of dependence symptoms (mean: 4.4–5.2) than first device pen (mean: 3.0 symptoms) (Figure 2B). In multivariable analyses (Table 3), using a mod, box, non-Juul pod, or other for first device was associated more than a doubling of the odds of any dependence (AORs: 2.36–4.01; ps < .05) and mod, box, and non-Juul pod were also associated with higher rate ratios for number of dependence symptoms among those experiencing at least one symptom (ARR: 1.38–1.48; ps < .05). First device Juul (vs. pen) did not differentiate odds of experiencing >1 versus 0 dependence symptoms but was associated with experiencing 59% more dependence symptoms (ARR: 1.59[95% CI = 1.29 to 1.96]) among with >1 symptom.

In the sensitivity analysis that excluded other flavor from the reference group (Supplementary Table 2), the association between menthol/mint (vs. tobacco/flavorless) first flavor and experiencing >1 vs. 0 dependence symptoms was consistent with the main analysis with a slight attenuation (AOR: 1.67[95% CI = 0.81 to 3.44]).

Discussion

This study provides new evidence that the characteristics of the e-cigarette product first used may be associated with the persistence and intensity of vaping behaviors and dependence symptoms in young people. Our findings add to the existing literature of flavor use and vaping behaviors by reporting that young adults who initiated e-cigarette use with menthol/mint flavors, as compared to tobacco and other non-sweet/fruit flavors or flavorless products, had higher odds of continued and more frequent e-cigarette use, and greater nicotine dependence. It is possible that menthol/mint flavor increases the appeal of e-cigarettes to young adults by creating a cooling sensation and suppressing the harshness and bitterness of nicotine, which has been shown in previous experimental research.^13,30,31 The consequences of this cool sensory attribute might increase the likelihood that young people continue (vs. discontinue) use of e-cigarettes after initial trial, particularly products with higher nicotine concentration. Additionally, using mint/menthol-flavored products early in life may make it easier to inhale aerosol more deeply, which could facilitate nicotine delivery and acquisition and progression of e-cigarette dependence.

While fruit/sweet flavors were the most commonly used first flavor in this study, first use of this flavor was not associated with current e-cigarette use or dependence in this study. This result is inconsistent with a prior study conducted in 2015–2016 demonstrating that use of fruit/sweet flavors predicted vaping persistence and frequency in youths.¹⁴ It is likely that shifts in e-cigarette industry marketing strategies over time may explain why first use of menthol/mint, but not fruit/sweet, flavors were associated with vaping continuation and dependence in this study. Fruit/sweet flavors were seven times more commonly used than menthol/mint flavors among youth US e-cigarette users in 2015–2016³² and widely marketed. In 2018, Juul voluntarily stopped selling their fruit and dessert (but not menthol) flavors in US markets and other manufacturers subsequently adopted that marketing strategy. Respondents in this study that started with and came to prefer fruit/sweet flavors might have been inclined to reduce or stop vaping if their preferred flavors were no longer being sold at the time of this study’s 2018–2019 survey. By contrast, respondents who began with menthol/mint and continued using that flavor might have had continued access to their preferred flavor and no deterrent to keep them from vaping. Another possible explanation may be that those less nicotine dependent to begin with chose fruit or sweet flavors, though the current cross-sectional design precludes investigation of this hypothesis.

This study demonstrated that youth who first tried an e-cigarette from the later generation devices (e.g., mod/box) reported a higher level of vaping frequency and dependence symptoms than those who started on earlier second-generation pen-like devices. The later generation of vaping devices might typically deliver nicotine to the lung more efficiently,³³ thus increasing the nicotine uptake and dependence. Additionally, some later generation devices allow users to customize their product (e.g., voltage, airflow), potentially increasing nicotine delivery rates,^33,34 and in some cases improving the sensory experience of the aerosol.³⁵

A similar trend was observed with pod-style devices. Initial vaping with Juul (vs. pen-like devices) was not associated with a higher vaping prevalence and frequency but was associated with experiencing more severe vaping dependence among those with at least one vaping dependence symptom. Youth who first tried a non-Juul pod-style versus pen-style device reported higher level of vaping frequency and dependence symptoms. One possible reason for increased dependence is that pod-style products can have 2–10 times higher nicotine concentrations than e-cigarette solutions used in pen-like devices. They also use nicotine salt-formulations, and this combination may result in aerosol that is perceived as smooth and less harsh despite the high nicotine concentration.⁸ Another reason is the diminutive size and resemblance of USB drives of pod-style products may increase the ability to use the device in a discrete fashion, allowing users to conceal use even in situations where vaping is prohibited (e.g., at work, restaurants, movies, airplanes).²¹

It is important to note the recent surge of newer pod-style disposable e-cigarette products, such as Puff Bar, during the period³⁶ after the US FDA announced intentions to enforce noncompliance with marketing authorization requirements for cartridge-based e-cigarettes in flavors other than tobacco and menthol in February 2020.³⁷ This study did not find the association of initial disposable product use and subsequent vaping frequency and dependence. However, many cohort participants started vaping before the newer generation of disposable products entered the market, and it is likely that the majority of participants who first used a disposable e-cigarette likely used a cig-a-like product. Thus, these findings cannot address whether novel disposable pod-style products might promote vaping persistence, frequency, and dependence.

This study has limitations. First, the study sample was limited to those from Los Angeles, CA, USA, and thus the results may not be generalizable to other regions. Second, the sample was not large enough to be sufficiently powered to examine interactions between first device and first flavor used. Future studies should assess whether the combination of first device and first flavor used may have differential effects on future vaping frequency and dependence. Third, menthol/mint was administered as a single response option in the survey, and we were unable to separate mint and menthol flavor in the analyses. Given different regulatory actions on menthol and mint flavors in e-cigarettes,³⁷ further studies with more granular flavor categories is needed. Fourth, the observational single-time point design also precludes causal inferences. While unmeasured confounding cannot be ruled out, we statistically adjusted for 12 covariates that could differentiate the type of e-cigarette used first and propensity toward vaping progression and dependence. Finally, first vaping flavors and devices were determined retrospectively, and are subject to recall error. This cross-sectional study was unable to sufficiently control for nicotine dependence prior to vaping initiation. Therefore, it is possible that prior nicotine dependence influenced choice of first vaping flavor and device. Smoking status before e-cigarette use and transition between e-cigarette and combustible tobacco products over time were not assessed in this study. Continued vaping may play a role in preventing combustible tobacco smoking uptake, especially if flavors are satisfying and vaping devices are effective in delivering nicotine. Prospective studies are needed to examine the longitudinal relationships between first vaping characteristics, combustible tobacco use, and future vaping behaviors.

Notwithstanding these limitations, this study highlights the associations of exposure to mint/menthol flavor and mod/box/pod device type at first vaping with progression to frequent vaping and dependence symptoms in young people. Further studies are needed to assess the causal factors and conduct longitudinal surveillance in response to fast-changing landscapes in vaping products and marketing strategies. Regulations and tobacco control policies that address flavors and devices merit consideration for preventing young people who try e-cigarettes from vaping persistence and escalation as well as nicotine dependence.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Funding

Research reported in this publication was supported by grant numbers R01CA229617, U54CA180905 from the National Cancer Institute at the National Institutes of Health (NIH) and the Food and Drug Administration (FDA) Center for Tobacco Products (CTP), grant K01DA042950 from the National Institute for Drug Abuse at NIH, and grant 27-IR-0034 from the Tobacco Related Disease Research Program (TRDRP). The funder had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript.

Declaration of Interests

None declared.

Data Availability

Data available on request.