Subjective Experiences at E-Cigarette Initiation: Implications for E-Cigarette and Dual/Poly Tobacco Use among Youth

Abstract

Background:

Subjective experiences (SEs) at initiation of cigarettes, cigars, and smokeless tobacco have been established as predictors of continued use. To date, less is known about the relationships between SEs at e-cigarette initiation and subsequent e-cigarette use behaviors.

Methods:

This study used data from Waves 1–6 of the Texas Adolescent Tobacco and Marketing Surveillance (TATAMS) system; a population-based rapid response study of adolescents in major metropolitan areas of Texas. Participants were adolescents who self-reported ever using e-cigarettes across all 6 waves (n=1,104; N=460,069). Factor analyses examined structure of SEs at e-cigarette initiation. Weighted, multilevel, multivariate regression models examined role of SEs at e-cigarette initiation on subsequent past 30-day e-cigarette use behaviors.

Results:

Factor analyses identified a positive (i.e., euphoria, relaxation) and a negative (i.e., dizziness, cough, nausea) domain of SEs. Positive SEs at e-cigarette initiation predicted 1.20 (95% CI: 1.02 – 1.42) greater odds of subsequent past 30-day e-cigarette use. Similarly, positive SEs at e-cigarette initiation predicted greater relative risk of dual/poly e-cigarette use, relative to non-use (RRR: 1.61; 95% CI: 1.24 – 2.10) and exclusive e-cigarette use (RRR: 1.68; 95% CI: 1.26 – 2.24).

Conclusion:

This is the first study to observe longitudinal relationships between SEs at e-cigarette initiation and subsequent e-cigarette use behaviors. Findings highlight the importance of preventing initial e-cigarette use among adolescents.

INTRODUCTION

E-cigarettes are the most commonly used tobacco product among adolescents in the United States, with prevalence increasing nearly 48% from 2014 (13.4%) to 2020 (19.8%) (Wang et al., 2020). Adolescent e-cigarette use presents numerous public health concerns including nicotine dependence (Case et al., 2018) and initiation of combustible tobacco use (National Academies of Sciences & Medicine, 2018). Furthermore, a large proportion of adolescent e-cigarette users also report use of combustible cigarettes and cigars (i.e., dual/poly e-cigarette use) (Mantey, Omega-Njemnobi, & Montgomery, 2019; Wang et al., 2020). However, little is known about what factors influence the transition from e-cigarette initiation or experimentation (i.e., ever use) to current (e.g., past 30-day) e-cigarette use and dual/poly e-cigarette use among youth.

Subjective experiences at tobacco initiation is a predictor of progression from experimentation to established use, across tobacco products (DiFranza et al., 2004; Klein, Sterk, & Elifson, 2013; Mantey et al., 2017; Sanouri Ursprung, Savageau, & DiFranza, 2011). Greater positive subjective experiences (e.g., euphoria, relaxation) at initiation of combustible cigarette and smokeless tobacco use is a risk factor for sustained (e.g., past 30-day) use as well as dual/poly tobacco use (DiFranza et al., 2004; Klein et al., 2013; Mantey et al., 2017; Sanouri Ursprung et al., 2011; Zabor et al., 2012). The relationship between negative subjective experiences (e.g., nausea) at initiation and progression to sustained tobacco use is more complex. For example, negative subjective experiences at initiation reduce the probability of sustained use for cigar products (Mantey et al., 2017) and smokeless tobacco (Zabor et al., 2012) but increase the risk for progression to sustained use (Klein et al., 2013; Sanouri Ursprung et al., 2011) of combustible cigarettes and onset of nicotine dependence (DiFranza et al., 2004; Sanouri Ursprung et al., 2011).

Research examining subjective experiences at e-cigarette initiation remains limited. To date, studies have found e-cigarettes appear to elicit fewer subjective experiences at initiation than other tobacco products (Chen et al., 2017; Do et al., 2018; Mantey et al., 2017). One study found adolescents were 3.5 times more likely to report coughing and nearly 6 times more likely to report nausea at initiation of combustible cigarettes, relative to initiation of e-cigarettes (Chen et al., 2017). Further, the subjective experiences that constitute ‘positive’ and ‘negative’ domains differ for e-cigarettes, relative to other products (DiFranza et al., 2004; Klein et al., 2013; Mantey et al., 2017; Zabor et al., 2012); however, studies of e-cigarettes remain too limited to draw stark contrasts. To date, no study has established a statistical association between subjective experiences at e-cigarette initiation and subsequent e-cigarette use (Chen et al., 2017; Mantey et al., 2017).

This study has two aims. First, we examine the factor structure of subjective experiences at e-cigarette initiation to empirically differentiate between positive and negative experiences using data collected from a large population-based cohort of adolescent from Fall 2014 to Spring 2017. Second, we apply the factor structure to examine the longitudinal relationship between subjective experiences at e-cigarette initiation and subsequent (1) past 30-day e-cigarette use; and (2) past 30-day dual/poly e-cigarette use.

METHODS

Study Design

This study is a prospective analysis of longitudinal, cohort data collected from the Texas Adolescent Tobacco and Marketing Surveillance (TATAMS) system, a population-based rapid response study of adolescents (n=3,907; weighted sample [N]=460,069) in major metropolitan areas of Texas. Study participants were 6^th, 8^th, or 10^th grade students at enrollment (in 2014–15), attending school in the four largest metropolitan areas of Texas: Austin, Dallas/Fort Worth, Houston, and San Antonio. This study analyzes 6 waves of data collected biannually via a web-based survey from Fall of 2014 to Spring of 2017. Surveys were designed to describe tobacco use trajectories among youth and identify the impact of product characteristics (e.g., flavors) and other potential risk factors on these trajectories (Pérez et al., 2017).

Study Sampling

TATAMS used a complex, multi-stage sampling technique to recruit a population-based cohort of adolescents. The sampling frame of TATAMS was developed to include all private, public, and charter schools within the study’s geographic region to produce a sample that was representative of student enrollment in these areas. Sample weights were developed using the enrollment data from Texas Education Agency to calculate probability proportional to the enrollment by grade, sex, and race/ethnicity of participants at baseline (Pérez et al., 2017). The TATAMS sampling frame accounted for 97% of 6^th, 8^th, and 10^th grade students in the study region, and more than 40% of 6^th, 8^th, and 10^th grade students in Texas; further explanation of sampling procedures are published elsewhere (Pérez et al., 2017).

Participation was voluntary. Parental consent and student assent were obtained for each participant (HSC-SPH-13–0377). Students received a $10 Amazon gift card at completion of Wave 1 and Wave 2 surveys and a $25 Amazon gift card at the completion of Wave 3–6 surveys.

Study Sample

For this analysis, we examined n=1,003 participants who self-reported ever using an e-cigarette across Waves 1–6 of the study and had complete data on all variables included in this study (see below). Ever e-cigarette use was assessed by asking participants: “Have you EVER used an electronic cigarette, vape pen, or e-hookah, even one or two puffs?” Those who reported “yes” were considered ever e-cigarette users and eligible for the study.

For this longitudinal study, ‘baseline’ observation for each participant was the first wave at which ever e-cigarette use was reported. Participants were required to have initiated e-cigarette use and completed at least one survey wave. The n=1,003 participants provided 3,523 observations (i.e., completed surveys). Of the n=1,003 participants, 6.6% (n=233) provided only one observation while 47.9% (n=528) provided 4 or more observations (i.e., participated in 4 or more waves of the survey). Of these 1,003 participants, 103 initiated e-cigarette use at Wave 6. Attrition analyses (see below) examine the extent to which missing data impacted this study.

Measures

Subjective Experiences.

The exposure variable for this study was self-reported subjective experiences at initiation of e-cigarette use. Participants who reported ever using an e-cigarette were asked “Think back to the FIRST time you used an electronic cigarette, vape pen, or e-hookah. Did you experience any of the following?” Participants were presented with the following five possible subjective experiences at initiation of e-cigarette use: (1) Pleasurable Rush or Buzz; (2) Relaxed or Good; (3) Dizziness; (4) Coughing; (5) Sick to Your Stomach. Items were adapted from the Early Smoking Experiences Questionnaire (alpha: 0.67; 0.75) (Rodriguez & Audrain-McGovern, 2004).

Guided by previous research conducted by our team (Mantey et al., 2017) and factor analyses (described below), two measures reflecting cumulative positive and cumulative negative subjective experiences at initiation were created. For positive subjective experiences, “felt pleasurable rush or buzz” and “relaxed or good” were combined to produce a cumulative measure with possible score of 0–2. For negative subjective experiences, “dizziness”, “coughing”, and “sick to your stomach” were combined to produce a cumulative measure with possible score of 0–3. As this variable could only be reported once (at the Wave of first reported e-cigarette use), the corresponding values for both positive and negative subjective experiences at e-cigarette initiation were considered “time-invariant” exposures and held constant across all waves for each participant.

E-Cigarette Use.

The first outcome variable of this study is past 30-day e-cigarette use, among ever e-cigarette users. Participants were asked “During the past 30-days, on how many days did you use an electronic cigarette, vape pen, e-hookah, MOD or tank system? Remember, marijuana DOES NOT count. Please enter the number of days (from 0 to 30 days)” Participants that reported 0 days were considered non-users (referent) while those who reported 1 or more days were considered e-cigarette users. Past 30-day e-cigarette use was assessed at all waves (i.e., time-variant).

Other Tobacco Product Use.

This study also examined past 30-day use of tobacco products other than e-cigarettes, at each wave (i.e., time-variant). Participants were asked to self-report use of the following tobacco products within the past 30-days: cigarettes; hookah; smokeless tobacco; cigar products (i.e., large cigars; little filtered cigars; or cigarillos). Participants that reported using one or more of these products in the past 30-days were considered other tobacco users.

Dual/Poly Use.

Informed by measures noted above, a categorical variable that combined past 30-day use of e-cigarettes and past 30-day use of other tobacco products was created. Participants were categorized into the following mutually exclusive groups: no e-cigarette use (referent), exclusive e-cigarette use; and e-cigarette and other tobacco product use.

Covariates.

This study controlled for the following socio-demographic variables, measured at Wave 1 (i.e., time-constant): biological sex, race/ethnicity, and grade. Biological sex is a dichotomous variable; males served as the referent group. Race/ethnicity was categorized as follows: Hispanic/Latino (referent); NH-White; NH-Black; and NH-“other” or multiracial. For this study, “other” includes NH-Asian, NH-Native American/Alaskan Native, NH-Native Hawaiian/Other Pacific Islander, and any other race/ethnicity; multiracial reflects the combination of two or more racial groups but not Hispanic/Latino ethnicity. Grade was coded as a categorical variable reflecting a participant’s grade at enrollment into the TATAMS study (i.e., cohort); categorical responses were 6^th grade (referent), 8^th grade, and 10^th grade. This study also controlled flavor used at initiation with possible responses being: fruit or candy (referent); mint or menthol; “other” (tobacco, spice, coffee, alcohol); and “I do not remember.”

Attrition Analyses

Attrition analyses were conducted using bivariate analyses (i.e., chi-square test) and post hoc tests (i.e., phi coefficient, Cramer’s V) to examine possible selection bias due to missingness. These analyses examined if (and to what degree) participants with complete data differed significantly from those who were removed due to incomplete data.

Bivariate analyses comparing complete observations (n=3,523) to those excluded due to missing data (n=384) revealed no statistical differences across any study variable; it should be noted that this rate of missingness is this study (Pérez et al., 2017) is consistent with similar observational studies (McDonald, Haardoerfer, Windle, Goodman, & Berg, 2017). Additionally, prior examination of attrition for the TATAMS dataset found no relationship between e-cigarette or other tobacco use at Wave 1 and missingness at Wave 2 (Pérez et al., 2017).

Statistical Analysis

An exploratory factor analysis (EFA) was conducted to examine the factor structure of subjective experiences at e-cigarette initiation (Study Aim 1). The EFA identified factor loadings of the five subjective experiences at e-cigarette initiation items. Barlett’s test for sphericity and Kaiser-Meyer Olkin were calculate and evaluate the intercorrelation of latent factors. Principal component analysis with varimax rotation was used (DeVellis, 2016). Items with factor loadings greater than 0.30 were retained for the corresponding factor. For the EFA, only one observation for each participant was examined (n=1,104) as this analysis was focused on e-cigarette initiation, measured by first report of ever e-cigarette use.

A random-effect, multiple logistic regression model was used to assess the relationship between positive and negative subjective experiences at e-cigarette initiation (informed by the factor analysis describe above), separately and subsequent past 30-day e-cigarette use among adolescents who reported ever e-cigarette use. Finally, a random-effect, multinomial logistic regression model assessed the relationships between positive and negative subjective experiences at e-cigarette initiation, separately, and specific subsequent past 30-day e-cigarette use behaviors (i.e. no e-cigarette use, exclusive use of e-cigarettes, dual/poly use).

For this longitudinal analysis of panel data the outcome variables (i.e., past 30-day use of e-cigarettes and other tobacco products) were time-varying meaning use behaviors could change from one survey to the next for each participant. As surveys were nested within participants, responses were assumed to not be independent. To account for the correlation of repeated measures for a single participant (i.e., within-individual variation) over time, survey wave was included in all statistical analyses, serving as the random effect in all statistical models. Subjective experience at first report of ever use of e-cigarettes was the exposure variable; this could have occurred at the same wave as past 30-day use or a prior one. Thus, relationships in this study were a mix of concurrent (e.g., subjective experiences at first report of ever use of e-cigarettes at wave 1 and past 30-day e-cigarette use at wave 1) as well as longitudinal (e.g., subjective experiences at first report of ever use of e-cigarettes at wave 2 and e-cigarette use at wave 2–6).

All analyses also controlled for time-constant covariates (e.g., sex, race/ethnicity, grade at enrollment). Sampling weights were applied to these analyses to generalize findings back to the study population from which they were drawn. All analyses were conducted using STATA 14.2 (College Station, TX).

RESULTS

Descriptive Statistics

Participant ages ranges from a young as 10 to 18 years old over the course of the study. Mean age of e-cigarette initiation among the n=1,003 participants was 15.4 (SD: 1.5). Relaxation (46.9%) was the most commonly reported positive subjective experience at e-cigarette initiation. Coughing (18.1%) was the most commonly reported negative subjective experience at e-cigarette initiation. Over the course of all study waves (Wave 1–6), 12.1% of ever e-cigarette users reported exclusive past 30-day e-cigarette use and 8.5% reported past 30-day e-cigarette dual/poly use. Further descriptive statistics are available in Table 1.

Table 1.

Descriptive Statistics of Full Sample and by Past 30-Day E-Cigarette Use Behavior (n = 3,523, N = 567,879).

	Full Sample	E-Cigarette Use Behaviors
		Non-Usera	Exclusive E-Cigarette Useb	E-Cigarette Dual/Poly Userc
Percent of Sample Positive Subjective Experiences d	100%	79.4%	12.1%	8.5%
Relaxation	22.0%	46.1%	43.6%	59.7%
Buzz	46.9%	20.6%	18.4%	39.3%
Total:	0.66 (0.8)	0.63 (0.7)	0.60 (0.8)	0.96 (0.9)
Negative Subjective Experiences e
Nausea	4.9%	5.1%	2.0.7%	5.7%
Dizziness	4.9%	4.6%	4.3%	8.7%
Cough	18.1%	17.7%	19.3%	20.1%
Total:	0.27 (0.6)	0.27 (0.5)	0.26 (0.7)	0.34 (0.6)
Sex
Males	49.6%	48.0%	54.5%	56.9%
Females	50.4%	52.0%	45.4%	43.1%
Grade (at baseline)
6th Grade	10.0%	10.8%	8.0%	5.9%
8th Grade	32.1%	32.0%	36.9%	26.6%
10th Grade	57.9%	57.3%	55.2%	67.5%
Race/Ethnicity f
Hispanic/Latino	47.2%	47.8%	39.4%	52.5%
Non-Hispanic, White	14.0%	12.2%	21.4%	20.6%
Non-Hispanic, Black	15.1%	16.4%	11.2%	8.3%
Non-Hispanic, Other or multiracial	23.7%	23.6%	28.0%	18.7%
Flavor at Initiation g
Fruit or Candy	60.5%	59.4%	66.2%	62.4%
Mint or Menthol	6.2%	6.7%	3.3%	6.2%
Other	15.1%	16.4%	19.2%	15.6%
“I do not remember”	17.8%	18.9%	14.2%	13.0%

^aSelf-reported not using e-cigarettes in the past 30-days.

^bSelf-reported using e-cigarette, but not other tobacco products, in the past 30- days.

^cSelf-reported using e-cigarettes and one ore more tobacco products in the past 30-days.

^dPositive subjective experiences included feeling of relaxation and feeling a “buzz” (coded as 0–2).

^eNegative subjective experiences included nausea, dizziness, or coughing (coded as 0–3).

^fRace/ethnicity was categorized as follows: Hispanic/Latino (referent group); non-Hispanic White; non-Hispanic Black; non-Hispanic “other”; and non-Hispanic multiracial. For the purposes of this study, “other” includes non- Hispanic Asian, non-Hispanic Native American/Alaskan Native, non-Hispanic Native Hawaiian or Other Pacific Islander, and any other race/ethnicity. Similarly, non-Hispanic multiracial reflects the combination of two or more racial groups but not Hispanic/Latino ethnicity.

^gSelf-reported initiating e-cigarette use via a fruit or candy flavor.

A total of n=1,003 participants reported ever using an e-cigarette. Breakdown of first reported e-cigarette use by survey wave was: n=127 (11.6% of sample) at Wave 1/Fall 2014, n=7 (0.5%) at Wave 2 (Spring 2015), n=15 (2.1%) at Wave 3 (Fall 2015), n=633 (65.9%) at Wave 4 (Spring 2016), n=102 (9.1%) at Wave 5 (Fall 2016), and n=119 (10.7%) at Wave 6 (Spring 2017).

Factor Analysis

As seen in Table 2, two factors were retained for the subjective experiences at initiation scale for e-cigarettes. These figures revealed a Positive Subjective Experiences factor comprised of “felt a pleasurable rush or buzz” (α=0.64) and “felt relaxed or good” (α=0.65). Similarly, Negative Subjective Experiences factor comprised of “dizziness” α= (0.43) “coughing” (α=0.32), and “sick to your stomach” (α=0.45).

Table 2:

Exploratory Factor Analyses of Subjective Experiences of First E-Cigarette Use (n=1,104)

Symptom	Factor 1a	Factor 2b
Relaxed/Good	0.65
Rush/Buzz	0.64
Dizziness		.43
Coughing		.32
Sick/Nausea		.45
Alpha/correlation for scales¥	0.68	0.38

^¥alpha for scales with >2 items, pearson correlation for scales

^aReflects “positive subjective experiences”

^bReflects “negative subjective experiences”

Logistic Regression Analyses

As seen in Table 3, positive subjective experiences at e-cigarette initiation significantly predicted subsequent past 30-day e-cigarette use, adjusting for covariates. Specifically, the odds of progression to past 30-day e-cigarette use increased by 1.19 (95% CI: 1.01 – 1.40) with each additional positive subjective experience at e-cigarette initiation. There was no statistical relationship observed between negative subjective experiences at e-cigarette initiation and subsequent past 30-day e-cigarette use.

Table 3:

Association between Subjective Experiences at E-Cigarette Initiation and Past 30-day E-Cigarette Use; Full Sample and Stratified by Flavors at Initiation (n=3,523, N=567,879)

	Past 30-Day E-Cigarette Use
	Adj OR 95% Confidence Intervals
Positive Subjective Experiences a
Total	1.19* (1.01 – 1.40)
Negative Subjective Experiences b
Total	1.07 (0.85 – 1.34)
Sex
Males	1.00 (Referent)
Females	0.75* (0.58 – 0.96)
Grade (at baseline)
6th Grade	1.00 (Referent)
8th Grade	1.43 (0.94 – 2.17)
10th Grade	1.34 (0.92 – 1.94)
Race/Ethnicity c
Hispanic/Latino	1.00 (Referent)
Non-Hispanic, White	1.94*** (1.43 – 2.62)
Non-Hispanic, Black	0.67 (0.42 – 1.06)
Non-Hispanic, Other or Multiracial	1.12 (0.75 – 1.65)
Flavor at Initiation d
Fruit or Candy	1.00 (Referent)
Mint or Menthol	0.55 (0.29 – 1.02)
Other	0.97 (0.67 – 1.43)
“I do not remember”	0.67* (0.47 – 0.92)

^*p < .05

^**p < .01

^***p < .001

NOTE: All models also controlled for Survey Wave.

^aPositive subjective experiences included feeling of relaxation and feeling a “buzz” (coded as 0–2)

^bNegative subjective experiences included nausea, dizziness, or coughing (coded as 0–3)

^cRace/ethnicity was categorized as follows: Hispanic/Latino (referent group); non-Hispanic White; non-Hispanic Black; non-Hispanic “other”; and non-Hispanic multiracial. For the purposes of this study, “other” includes non-Hispanic Asian, non-Hispanic Native American/Alaskan Native, non-Hispanic Native Hawaiian or Other Pacific Islander, and any other race/ethnicity. Similarly, non-Hispanic multiracial reflects the combination of two or more racial groups but not Hispanic/Latino ethnicity.

^dSelf-reported initiating e-cigarette use via a fruit or candy flavor.

Multinomial Regression Analyses

As seen in Table 3, positive subjective experiences at e-cigarette initiation significantly predicted greater relative risk for subsequent past 30-day e-cigarette dual/poly use (RRR: 1.58; 95% CI: 1.21 – 2.10) but not exclusive e-cigarette use (RRR: 0.96; 95% CI: 0.81 – 1.13), relative to no e-cigarette use. In addition, positive subjective experiences at e-cigarette initiation significantly predicted greater relative risk for e-cigarette dual/poly use (RRR: 1.65; 95% CI: 1.24 – 2.20), relative to exclusive e-cigarette use, within the past 30-days. There was no statistical relationship observed between negative subjective experiences at e-cigarette initiation and subsequent past 30-day e-cigarette use or subsequent dual/poly use.

DISCUSSION

This study found that the profile of positive and negative subjective experiences at e-cigarette initiation mirror that of combustible cigarettes and cigars (Chen et al., 2017; Do et al., 2018; Mantey et al., 2017) and that positive subjective experiences at initiation longitudinally predict current (past 30-day) e-cigarette and current (past 30-day) dual/poly use among a cohort of adolescents. Each of these findings are novel to the field and have substantial implications for public health intervention, regulatory policy, and future research.

To our knowledge, this is the first study to link subjective experiences at e-cigarette initiation and subsequent e-cigarette use behaviors. However, findings do build on prior cross-sectional (Mantey et al., 2017) and longitudinal (Chen et al., 2017) research of e-cigarette initiation. Descriptively, our study replicates findings from prior research that found much higher prevalence of positive (compared to negative) subjective experiences at e-cigarette initiation (Chen et al., 2017; Do et al., 2018; Mantey et al., 2017). Additionally, the statistically significant findings from this study are an expansion of prior research that found a positive, but not statistically significant, relationship between positive subjective experiences at e-cigarette initiation and current e-cigarette use among adolescents (Chen et al., 2017; Mantey et al., 2017). This latter finding mirrors the established relationship between subjective experiences at initiation of combustible cigarettes and subsequent use (DiFranza et al., 2004; Klein et al., 2013; Sanouri Ursprung et al., 2011; Zabor et al., 2012).

This study also found that positive subjective experiences at e-cigarette initiation predicted dual/poly e-cigarette use, relative to both no e-cigarette and exclusive e-cigarette use. One previous study of adult tobacco users found that positive subjective experiences at e-cigarette initiation predicted greater risk of dual use of combustible cigarettes and smokeless tobacco (Zabor et al., 2012). Beyond the novelty of these findings, results have significant health implications given: (1) the high prevalence of dual/poly use among adolescent e-cigarette users (Mantey, Barroso, Kelder, & Kelder, 2019); (2) increased risk for nicotine dependence and long-term tobacco use associated with dual/poly tobacco use (Bandiera, Loukas, Li, Wilkinson, & Perry, 2017; Health & Services, 2012; Loukas et al., 2016); and (3) increased exposure to toxicants and carcinogens (Maciej Lukasz Goniewicz, Boykan, Messina, Eliscu, & Tolentino, 2019; Maciej L Goniewicz et al., 2018; Teo et al., 2006). Further research is needed to explore the reason(s) for e-cigarette initiation and subsequent use of multiple tobacco products among youth and adults.

Descriptive and analytic results indicate the importance of preventing and reducing e-cigarette initiation among youth. This study found that a sizable proportion of ever e-cigarette users transitioned to current e-cigarette use (12.1%) and dual/poly use (8.5%) and that positive subjective experiences during initiation increased risk for this continued use. The reinforcing effects of positive subjective experiences at e-cigarette initiation signal the need to reduce those who initiate as a method of reducing transition to current e-cigarette and dual/poly use. School-based programs are a proven method of delivering health promotion interventions and, in the context of tobacco prevention, several have been demonstrated to reduce initiation during adolescence (Kelder et al., 2020; Thomas, McLellan, & Perera, 2013, 2015). As such, concentrated efforts are needed to prevent e-cigarette initiation during adolescence.

Findings also underscore the need for greater regulation of e-cigarette product characteristics. For example, characterizing flavors (e.g., mint/menthol; fruit/candy) are specifically designed to improve subjective experiences for the user and are uniquely popular among youth (Dai, 2019; Harrell, Loukas, Jackson, Marti, & Perry, 2017; Harrell, Weaver, et al., 2017; Mantey, Omega-Njemnobi, et al., 2019). Studies have found that use of flavored products, particularly at initiation, increase the risk for continued e-cigarette use (Audrain-McGovern, Rodriguez, Pianin, & Alexander, 2019) and dual/poly use (Mantey, Omega-Njemnobi, et al., 2019). As such, restriction of characterizing flavors for e-cigarette products should be considered by regulatory and legislative entities.

Findings reflect the need to consider restricting nicotine concentration in e-cigarette products as a method of reducing the transition from initiation to sustained e-cigarette use and dual/poly use. Nicotine is the primary psychoactive chemical in e-cigarettes (National Academies of Sciences & Medicine, 2018) and elicits positive subjective experiences (e.g., relaxation; buzz/euphoria) in users (Benowitz, 2010, 2014). At this writing, there are no restrictions on nicotine concentrations for e-cigarette products in the US. As a result, popular brands of e-cigarettes (e.g., JUUL) are available in doses as high as 59 mg/mL (5.0% by weight); similar nicotine concertation as a pack of combustible cigarettes. Other regions (e.g., European Union) have adopted restrictions on nicotine concentration in e-cigarettes (Bertollini, Ribeiro, Mauer-Stender, & Galea, 2016). It should be noted that regions with restrictions on nicotine concentration in e-cigarettes have not reported a disruption in adults using e-cigarette for cigarette smoking cessation (Kapan et al., 2020) and thus similar restrictions in the US would not be expected to impede harm reduction efforts.

Findings have implications for public health research. Results from the factor analysis found that feelings of relaxation and euphoria loaded on one factor (i.e., “positive”) while coughing, dizziness, and nausea loaded on another (i.e., “negative”); this reflects similar findings from prior studies of combustible tobacco (Chen et al., 2017; Do et al., 2018; Mantey et al., 2017). Additionally, ever e-cigarette users report a low prevalence of negative subjective experiences at e-cigarette initiation, as is consistent with prior research on e-cigarettes (Chen et al., 2017; Do et al., 2018; Mantey et al., 2017) and in contrast with the profile of subjective experiences at combustible cigarette initiation, which elicits a high prevalence of negative subjective experiences (DiFranza et al., 2004; Klein et al., 2013; Zabor et al., 2012). Similarly, negative subjective experiences continue to not be significantly associated with e-cigarette use (Chen et al., 2017; Do et al., 2018; Mantey et al., 2017), which is another contrast from our understanding of combustible cigarettes and subjective experiences (Klein et al., 2013; Zabor et al., 2012).. Future studies may further examine similarities and differences in subjective experiences at initiation across tobacco products as well as e-cigarette device types.

This study has limitations. First, all data are self-reported and thus subject to response and recall bias. Second, this study could not control for prior exposure to nicotine via other tobacco products. As users build up a tolerance to nicotine with greater exposure (Benowitz, 2010; Benowitz, Hukkanen, & Jacob, 2009), it is plausible that prior tobacco use influenced subjective experiences during e-cigarette initiation. Third, subjective experiences were assessed via dichotomous response, thus we cannot speak to severity/intensity of subjective experiences. Fourth, study findings are specific to adolescents in urban Texas and thus may not be representative of other locations or populations. Fifth, this study was unable to control for nicotine concentration and device type at e-cigarette initiation. As newer e-cigarette devices are able to deliver nicotine in doses similar to combustible cigarettes (Jackler & Ramamurthi, 2019) and are linked to sustained use (Sumbe et al., 2021), future research should explore these factors in the context of subjective experiences.

Despite limitations, this study expands the understanding of subjective experiences at initiation and subsequent e-cigarette use among adolescents. Findings show adolescents who report greater positive subjective experiences at e-cigarette initiation are more likely to progress to current e-cigarette use and dual/poly use. Future research should explore factors that impact subjective experiences at e-cigarette initiation include genetic sensitivity to nicotine (Chukwueke et al., 2020; Pérez-Rubio et al., 2020), product type (Sumbe et al., 2021), user topography (Lopez et al., 2015), and characterizing flavors. Similarly, as youth often endorse positive outcome expectations related to subjective experiences, such as feeling relaxed (Barker et al., 2019; Pokhrel, Little, Fagan, Muranaka, & Herzog, 2014), future study should explore the relationship between outcome expectations prior to use and subjective experiences at initiation.

Table 4:

Multinomial Logistic Regression for Relationship between Subjective Experiences at E-Cigarette Initiation and E-Cigarette Use Category; (n=3,523, N=567,879)

	Model 1: Non-Usersa as Referent Outcome		Model 2: Exclusive E-Cigarette Useb as Referent Outcome
	Exclusive E-Cigarette Useb	E-Cigarette Dual/Poly Usec	E-Cigarette Dual/Poly Usec
	Relative Risk Ratio 95% Confidence Intervals	Relative Risk Ratio 95% Confidence Intervals	Relative Risk Ratio 95% Confidence Intervals
Positive Subjective Experiences d
Total	0.96 (0.81 – 1.13)	1.58** (1.21 – 2.05)	1.65** (1.24 – 2.20)
Negative Subjective Experiences e
Total	1.01 (0.81 – 1.28)	1.14 (0.79 – 1.64)	1.12 (0.76 – 1.67)
Sex
Males	1.00 (Referent)	1.00 (Referent)	1.00 (Referent)
Females	0.81 (0.58 – 1.14)	0.68 (0.45 – 1.01)	0.83 (0.48 – 1.43)
Grade (at baseline)
6th Grade	1.00 (Referent)	1.00 (Referent)	1.00 (Referent)
8th Grade	1.53 (0.93 – 2.54)	1.26 (0.46 – 3.43)	0.82 (0.24 – 2.81)
10th Grade	1.16 (0.76 – 1.76)	1.63 (0.64 – 4.16)	1.41 (0.47 – 4.21)
Race/Ethnicity f
Hispanic/Latino	1.00 (Referent)	1.00 (Referent)	1.00 (Referent)
NH, White	2.31*** (1.61 – 3.31)	1.48 (0.95 – 2.31)	0.64 (0.38 – 1.09)
NH, Black	0.87 (0.51 – 1.48)	0.46* (0.25 – 0.83)	0.53* (0.29 – 0.96)
NH, Other/Multiracial	1.47 (0.89 – 2.43)	0.73 (0.44 – 1.22)	0.50* (0.26 – 0.96)
Flavor at Initiation g
Fruit or Candy	1.00 (Referent)	1.00 (Referent)	1.00 (Referent)
Mint or Menthol	0.42* (0.21 – 0.84)	0.76 (0.32 – 1.78)	1.79 (0.71 – 4.54)
Other	0.91 (0.60 – 1.37)	1.08 (0.64 – 1.82)	1.19 (0.69 – 2.07)
“I do not remember”	0.63* (0.43 – 0.93)	0.70 (0.37 – 1.33)	1.11 (0.51 – 2.39)

^*p < .05

^**p < .01

^***p < .001

NOTE: All models also controlled for Survey Wave.

^aSelf-reported not using e-cigarettes in the past 30-days

^bSelf-reported using e-cigarette, but not other tobacco products, in the past 30-days

^cSelf-reported using e-cigarettes and one ore more tobacco products in the past 30-days

^dPositive subjective experiences included feeling of relaxation and feeling a “buzz” (coded as 0–2)

^eNegative subjective experiences included nausea, dizziness, or coughing (coded as 0–3)

^fRace/ethnicity was categorized as follows: Hispanic/Latino (referent group); non-Hispanic White; non-Hispanic Black; non-Hispanic “other” or multiracial. For the purposes of this study, “other” includes non-Hispanic Asian, non-Hispanic Native American/Alaskan Native, non-Hispanic Native Hawaiian or Other Pacific Islander, and any other race/ethnicity. Similarly, non-Hispanic multiracial reflects the combination of two or more racial groups but not Hispanic/Latino ethnicity. “NH” reflects “non-Hispanic”

^gSelf-reported using tobacco, coffee, spice, or alcohol flavors at e-cigarette initiation

Negative subjective experiences included nausea, dizziness, or coughing (coded as 0–3)

Highlights.

• Subjective experiences (SEs) for e-cigarettes mirror those of other tobacco products.

• Positive SEs at initiation predicted 1.20 greater odds of continued e-cigarette use

• Positive SEs at initiation predicted 1.68 greater odds of dual/poly tobacco use.