Abstract
Introduction:
As electronic cigarette (e-cigarette) use has become more prevalent among adolescents, there is a growing body of evidence linking e-cigarette use to the initiation of other substances. Whether there is a threshold level of e-cigarette use that is predictive of other substance use is unknown. The current study examines patterns of e-cigarette use over time and determines whether different patterns of early adolescent e-cigarette use are concurrently and prospectively associated with alcohol and marijuana use in late adolescence.
Method:
Eight hundred and one adolescents (13-15 years old at baseline recruitment) completed five on-line surveys over a two-year period. Latent class growth analysis was used to model different developmental courses of e-cigarette, alcohol (drinking to intoxication), and marijuana use. Logistic regression was used to test the association between e-cigarette use trajectory patterns and alcohol and marijuana use trajectories.
Results:
Three developmental courses of e-cigarette use were identified: 1) high and increasing, 2) low and increasing, and 3) never. Compared to adolescents who had never used e-cigarettes, those in the other two groups were more likely to have been intoxicated and to be in the moderate and increasing marijuana use group.
Conclusion:
Both high and low levels of e-cigarette use patterns are associated with increasing use of other substances (alcohol and marijuana use) over time. Findings highlight the need for early intervention and prevention of e-cigarette use among adolescents.
Keywords: Adolescent behavior, electronic cigarettes, alcohol, marijuana
1. Introduction
Adolescent electronic cigarette (e-cigarette) use is an emerging public health problem. E-cigarette use among high school students has risen sharply from 1.5% in 2011 to 20.8% in 2018, surpassing conventional cigarette use (Centers for Disease Control and Prevention [CDC], 2018a; Johnston et al., 2019; U.S. Department of Health and Human Services [DHHS], 2016). Whether e-cigarettes are beneficial or harmful is widely debated; however, for adolescents, e-cigarette use may have significant adverse health implications. E-cigarettes contain nicotine and other chemicals, which may be especially harmful to adolescent brain development and have long-lasting negative effects (Peterson & Hecht, 2017). In addition, unlike adult users who may use e-cigarettes for smoking cessation, adolescent users may be more likely to transition from e-cigarettes to regular cigarettes (Fulton, Gokal, Griffiths, & Wild, 2018).
There is growing concern that e-cigarettes may be a gateway to other substance use. Use of e-cigarettes has been positively associated with use of other substances, including marijuana and alcohol (Audrain-McGovern, Stone, Barrington-Trimis, Unger, & Leventhal, 2018; Lessard et al., 2014; Milicic & Leatherdale, 2017). However, most of the studies have been cross-sectional (e.g., Milicic & Leatherdale, 2017; Roberts et al., 2018), making it difficult to ascertain whether e-cigarette use is predictive of other substance use or whether it is one of several substances being used concurrently. The few studies that have used longitudinal designs indicate prospective associations between e-cigarette use and marijuana, alcohol, and cigarette use among older (Audrain-McGovern et al., 2018; Bold et al., 2018) and younger adolescents (Dai, Catley, Richter, Goggin, & Ellerbeck, 2018; Westling, Rusby, Crowley, & Light, 2017), with e-cigarette use predicting later substance use. What has yet to be established is whether there are different patterns of e-cigarette use that are more strongly associated with substance use over time than others.
Substance use frequently increases across adolescence (Chen & Jacobson, 2012; Johnston et al., 2019). Studies examining the trajectories of alcohol, marijuana, and conventional cigarette use generally show that escalating use of one substance is typically associated with increased use of other substances as well. Individuals with high/chronic and increasing patterns of any of these substances tend to have the highest rates of poly substance use and the worst outcomes. For example, compared with low and infrequent users, adolescents demonstrating a chronic (high and occasional) and increasing pattern of marijuana use had higher rates of nicotine dependence, alcohol abuse/dependence, emotional dysregulation, and having a marijuana using partner in early adulthood (Brook, Zhang, Leukefeld, & Brook, 2016). Passarotti, Crane, Hedeker, and Mermelstein (2015) found that adolescents with a chronic and high pattern of marijuana use had the highest rates of cigarette use; however, those with an initially lower but increasing pattern of marijuana use also showed similarly high levels of cigarette use six years later. In another study of alcohol and cigarette use from adolescence to early adulthood, relative to those who decreased their cigarette use over time, those who drank and smoked cigarettes consistently had higher rates of deviant behavior and substance use in early adulthood (Orlando, Tucker, Ellickson, & Klein, 2005). In one of the few studies to examine patterns of e-cigarette use and their relation to substance use over time, Westling et al. (2017) identified two classes of e-cigarette users in a large sample of 8th grade students: a low use group (comprised of non-users and infrequent users) and an escalating group. Compared with the low use group, the escalating group had higher rates of cigarette, alcohol, and marijuana use at the end of 8th and 9th grades. Taken together, these studies show that early high and chronic use of substances is positively associated with other substance use over time. However, there is a dearth of research examining the trajectories of e-cigarette use, particularly extending into late adolescence. Thus, it is unknown whether e-cigarette use continues to escalate over time and whether there are different patterns of e-cigarette use that are differentially associated with substance use.
1.1. Aims of this Study
Health consequences of polysubstance use may be more serious than use of any one substance alone (Ives & Ghelani, 2006). Thus, it is important to understand co-occurrence and developmental changes in these behaviors over time. Moreover, little is known regarding the potential of e-cigarettes to be a gateway substance leading to subsequent alcohol and marijuana use over time. The goal of this study was to: (1) determine the developmental trajectories of e-cigarette, alcohol intoxication, and marijuana use over time from early to late adolescence; (2) examine the association of e-cigarette use with alcohol and marijuana use trajectories; and (3) examine whether adolescent e-cigarette use trajectory patterns (group membership) predict alcohol and marijuana use trajectory patterns over time. We hypothesized that: (1) there are distinct trajectory patterns of e-cigarette, alcohol, and marijuana use, including high use, low use, and never user groups; (2) e-cigarette use is associated with alcohol and marijuana use concurrently and prospectively; and (3) adolescents who are in the high use e-cigarette trajectory group are more likely than low or never use groups to be in the high use trajectory groups for alcohol and marijuana use.
2. Methods
2.1. Participants
Participants were a community sample of adolescents (N = 801, 57% female) aged 13 to 15 years at baseline recruitment (M = 14.45, SD = 0.85). Consistent with the demographics of Erie County, NY, 81% of participants self-identified as European American, 12% as African American, 4% as multiracial, 1% as Asian American, and < 1% as Native American. About 7% of the participants indicated that they were Hispanic/Latino. Participants were in 8th (32%), 9th (35%), or 10th (23%) grades and attended a public (85%), private (12%), or charter school (3%). About 61% of mothers had a college degree, and the median family income was $80,000 or more.
2.2. Procedure
2.2.1. Recruitment.
A sample of adolescents and their mothers in Erie County, NY were recruited using address-based sampling for a larger study examining the role of peer victimization (i.e., bullying and sexual harassment) in the development of adolescent substance use. Each household received at least two mailings; up to four mailings were sent to addresses in ethnically diverse and low-income neighborhoods to increase response rate. Interested individuals were screened for eligibility over the phone. To be eligible, adolescents had to be between the ages of 13 and 15, attending a public, charter, or private school, living with a mother or legal female guardian who was also willing to participate in the study, and had to speak and read English at a 6th-grade level. Of the 916 adolescents who were enrolled in the study, 801 (87.4%) completed the baseline surveys.
The web-based surveys were administered using a secure server at baseline and every six months post-baseline over a two-year period for a total of five waves. Each survey took about 1 ½ hours to complete and participants received a $25 check for their participation in each wave. Parental consent and adolescent assent were obtained and all procedures were approved by the Institutional Review Board of the affiliated university.
2.3. Measures
2.3.1. Demographics.
Demographic data were collected at baseline (age, sex, race and ethnicity, mother’s education, and household income). Household income was taken from the mothers’ reports.
2.3.2. E-cigarette use.
At Waves 1-5, e-cigarette use was assessed using the question, “On how many occasions (if any) did you use an electronic cigarette or e-cigarette (vape pen) during the last 6 months?”. This item was rated on a 7-point scale (0 = Never; 1 = 1-2 times; 2 = 3-5 times; 3 = 6-9 times; 4 = 10-19 times; 5 = 20-39 times; 6 = 40 or more times).
2.3.3. Alcohol use and intoxication.
Some alcohol use in late adolescence is normative (Johnston et al., 2019), making any alcohol use a poor indicator of problematic behavior. Because alcohol intoxication is one of the important indicators of adolescent problematic drinking behavior and has been associated with adverse physical and psychological health outcomes (Lipperman-Kreda & Grube, 2019; Lamminippa, 1995; WHO, 2018), we examined whether e-cigarette use was associated with drinking to intoxication. For adults, the standard for heavy drinking is 4 drinks for women and 5 drinks for men (NIAAA, 2018). There is no standard heavy drinking index for adolescents. Given that adolescents are less developed, less experienced with alcohol, and may (or may not) weigh less than adults, it is feasible that some adolescents may become intoxicated with fewer than 4-5 drinks. Thus, we opted to assess self-reported intoxication (i.e., being drunk) as our index of risky alcohol use. Alcohol intoxication, a proxy for binge drinking, was measured using the single item, “On how many occasions (if any) have you been drunk during the last 6 months?”. This item was assessed at all five Waves, and rated on a 7-point scale (0 = Never; 1 = 1-2 times; 2 = 3-5 times; 3 = 6-9 times; 4 = 10-19 times; 5 = 20-39 times; 6 = 40 or more times). In addition, participants also reported on the number of drinks it typically took for them to feel drunk.
2.3.4. Marijuana use.
Adolescent marijuana use was assessed at all five Waves using the question, “On how many occasions have you used marijuana (e.g., pot, grass, hashish) that was not prescribed for you by a doctor during the last 6 months?” The item was also rated on a 7-point scale (0 = Never; 1 = 1-2 times; 2 = 3-5 times; 3 = 6-9 times; 4 = 10-19 times; 5 = 20-39 times; 6 = 40 or more times).
2.4. Analytic Strategy
Analyses were performed in Mplus (Version 8.2; Muthén & Muthén, 2017) using maximum likelihood estimation with robust standard errors. Our primary analyses included three steps. In Step 1, we conducted linear growth analysis (LGA) models to examine developmental trajectories of e-cigarette use, alcohol intoxication, and marijuana use over time separately. Trajectory patterns were estimated by adolescent age rather than study wave (Duncan, Duncan, Strycker, & Chaumeton, 2007; Mehta & West, 2000). Accordingly, analyses spanned ages 13 to 17.5 years. Two participants (2/801; 0.2%) became 16 shortly after recruitment but before completing the baseline survey (i.e., within days or a couple of weeks), and we coded them as 15.5 years old at baseline in our analysis. All three substance use variables were analyzed as count variables.
In Step 2, we examined the concurrent (i.e., within-age) and prospective (i.e., cross-lagged) associations between e-cigarette use modeled as time-varying covariates and both alcohol intoxication and marijuana use over time in LGA models. Specifically, we tested the within-age associations (e.g., e-cigarette at aged 13 → intoxication at aged 13, … e-cigarette at aged 17.5 → intoxication at aged 17.5) and cross-lagged associations (e.g., e-cigarette at aged 13 → intoxication at aged 13.5, … e-cigarette at aged 17 → intoxication at aged 17.5). In Step 3, we examined associations between trajectory patterns of e-cigarette use with trajectories of alcohol and marijuana use over time. A series of latent class growth analysis (LCGA) models were used to identify the developmental courses of e-cigarette use, intoxication, and marijuana use. Evaluation of the best fitting models was based on conventional standards, including the Bayesian information criterion (BIC), the Lo–Mendell–Rubin likelihood ratio test, the bootstrapped likelihood ratio test, and entropy (Lo, Mendell, & Rubin, 2001; Nylund, Asparouhov, & Muthén, 2007). Next, logistic regression modeling from e-cigarette use to intoxication and from e-cigarette use to marijuana use were conducted. Sex (0= male; 1 = female), mother’s education level, and mother-reported household income were control variables in our models.
3. Results
Of 801 adolescents who completed baseline (i.e., Wave 1), 752 (93.9%) completed Wave 2, 728 (90.9%) completed Wave 3, 676 (84.4%) completed Wave 4, and 715 (89.3%) completed Wave 5 data. Missing analyses showed that, at baseline, adolescents who had missing data at some time point(s) were more likely than those who had complete data to be males, χ2(1, 801) = 8.367, p = .004; to use marijuana more frequently, t(799) = 2.882, p = .004; to have a mother with lower education, t(793) = −2.132, p = .033; and to have lower household income, t(781) = −3.016, p = .003. No differences were found in age, e-cigarette use, or intoxication. Adolescents also reported how old they were the first time they tried an e-cigarette (M = 14.14, SD = 1.37), the first time they got drunk (M = 14.47, SD = 1.25), and the first time they used marijuana (M = 14.20; SD = 1.50). There were no sex differences in age at first time they used these substances. On average, adolescents reported typically consuming 3.85 drinks (SD = 1.69) for them to feel drunk. Males (M = 4.53, SD = 1.80) typically consumed more drinks than females (M = 3.48, SD = 1.51) to feel drunk, t(247) = 4.924, p < .001. These quantities are slightly lower than the standard definitions of heavy drinking for male and female adults. Descriptives of primary variables from ages 13 to 17.5 years are displayed in Table 1, and frequencies for e-cigarette use, alcohol intoxication, and marijuana use over time are displayed in Table 2.
Table 1.
Mean Level of E-Cigarette Use, Drinking To Intoxication, and Marijuana Use across Measurement Occasions from Ages 13 to 17.5 Years
| Male |
Female |
Total |
|||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Variable | Min | Max | Mean | SD | N | Min | Max | Mean | SD | N | Min | Max | Mean | SD | N |
| Mother education (W1)a | 2.00 | 7.00 | 5.78 | 1.01 | 339 | 2.00 | 7.00 | 5.80 | 1.03 | 456 | 2.00 | 7.00 | 5.79 | 1.02 | 795 |
| Household income (W1)b | 1.00 | 4.00 | 3.32 | 0.82 | 336 | 1.00 | 4.00 | 3.30 | 0.86 | 447 | 1.00 | 4.00 | 3.31 | 0.84 | 783 |
| Electronic cigarette usec | |||||||||||||||
| 13 | 0.00 | 3.00 | 0.11 | 0.46 | 71 | 0.00 | 2.00 | 0.04 | 0.26 | 71 | 0.00 | 3.00 | 0.08 | 0.38 | 142 |
| 13.5 | 0.00 | 6.00 | 0.21 | 0.82 | 117 | 0.00 | 6.00 | 0.17 | 0.74 | 132 | 0.00 | 6.00 | 0.18 | 0.78 | 249 |
| 14 | 0.00 | 6.00 | 0.39 | 1.35 | 159 | 0.00 | 4.00 | 0.26 | 0.68 | 211 | 0.00 | 6.00 | 0.31 | 1.02 | 370 |
| 14.5 | 0.00 | 6.00 | 0.52 | 1.38 | 217 | 0.00 | 6.00 | 0.45 | 1.21 | 293 | 0.00 | 6.00 | 0.48 | 1.29 | 510 |
| 15 | 0.00 | 6.00 | 0.76 | 1.68 | 260 | 0.00 | 6.00 | 0.59 | 1.39 | 360 | 0.00 | 6.00 | 0.66 | 1.52 | 620 |
| 15.5 | 0.00 | 6.00 | 0.62 | 1.55 | 234 | 0.00 | 6.00 | 0.75 | 1.61 | 354 | 0.00 | 6.00 | 0.70 | 1.59 | 588 |
| 16 | 0.00 | 6.00 | 0.71 | 1.53 | 170 | 0.00 | 6.00 | 0.83 | 1.78 | 282 | 0.00 | 6.00 | 0.79 | 1.69 | 452 |
| 16.5 | 0.00 | 6.00 | 1.02 | 1.85 | 129 | 0.00 | 6.00 | 0.79 | 1.74 | 195 | 0.00 | 6.00 | 0.88 | 1.79 | 324 |
| 17 | 0.00 | 6.00 | 1.07 | 2.07 | 71 | 0.00 | 6.00 | 0.68 | 1.61 | 120 | 0.00 | 6.00 | 0.83 | 1.80 | 191 |
| 17.5 | 0.00 | 6.00 | 1.59 | 2.42 | 32 | 0.00 | 6.00 | 0.41 | 1.27 | 59 | 0.00 | 6.00 | 0.82 | 1.84 | 91 |
| Drinking to intoxicationd | |||||||||||||||
| 13 | 0.00 | 1.00 | 0.03 | 0.16 | 37 | 0.00 | 1.00 | 0.05 | 0.23 | 38 | 0.00 | 1.00 | 0.04 | 0.20 | 75 |
| 13.5 | 0.00 | 1.00 | 0.03 | 0.18 | 63 | 0.00 | 1.00 | 0.02 | 0.15 | 86 | 0.00 | 1.00 | 0.03 | 0.16 | 149 |
| 14 | 0.00 | 5.00 | 0.17 | 0.75 | 101 | 0.00 | 2.00 | 0.11 | 0.35 | 142 | 0.00 | 5.00 | 0.13 | 0.55 | 243 |
| 14.5 | 0.00 | 5.00 | 0.18 | 0.62 | 147 | 0.00 | 5.00 | 0.27 | 0.83 | 215 | 0.00 | 5.00 | 0.24 | 0.75 | 362 |
| 15 | 0.00 | 6.00 | 0.44 | 1.11 | 174 | 0.00 | 5.00 | 0.39 | 0.94 | 277 | 0.00 | 6.00 | 0.41 | 1.01 | 451 |
| 15.5 | 0.00 | 6.00 | 0.43 | 1.07 | 174 | 0.00 | 5.00 | 0.68 | 1.21 | 289 | 0.00 | 6.00 | 0.58 | 1.16 | 463 |
| 16 | 0.00 | 6.00 | 0.56 | 1.18 | 133 | 0.00 | 5.00 | 0.68 | 1.19 | 233 | 0.00 | 6.00 | 0.63 | 1.19 | 366 |
| 16.5 | 0.00 | 4.00 | 0.72 | 1.23 | 98 | 0.00 | 6.00 | 0.89 | 1.44 | 158 | 0.00 | 6.00 | 0.82 | 1.36 | 256 |
| 17 | 0.00 | 5.00 | 0.70 | 1.30 | 54 | 0.00 | 6.00 | 0.77 | 1.25 | 97 | 0.00 | 6.00 | 0.75 | 1.26 | 151 |
| 17.5 | 0.00 | 6.00 | 0.76 | 1.41 | 29 | 0.00 | 6.00 | 0.82 | 1.38 | 49 | 0.00 | 6.00 | 0.79 | 1.38 | 78 |
| Marijuana usee | |||||||||||||||
| 13 | 0.00 | 2.00 | 0.04 | 0.26 | 71 | 0.00 | 0.00 | 0.00 | 0.00 | 72 | 0.00 | 2.00 | 0.02 | 0.19 | 143 |
| 13.5 | 0.00 | 4.00 | 0.03 | 0.37 | 117 | 0.00 | 4.00 | 0.04 | 0.36 | 132 | 0.00 | 4.00 | 0.04 | 0.36 | 249 |
| 14 | 0.00 | 6.00 | 0.14 | 0.83 | 160 | 0.00 | 4.00 | 0.12 | 0.55 | 212 | 0.00 | 6.00 | 0.13 | 0.68 | 372 |
| 14.5 | 0.00 | 6.00 | 0.30 | 1.08 | 218 | 0.00 | 6.00 | 0.19 | 0.83 | 294 | 0.00 | 6.00 | 0.24 | 0.94 | 512 |
| 15 | 0.00 | 6.00 | 0.32 | 1.09 | 261 | 0.00 | 6.00 | 0.33 | 1.03 | 360 | 0.00 | 6.00 | 0.33 | 1.06 | 621 |
| 15.5 | 0.00 | 6.00 | 0.42 | 1.23 | 233 | 0.00 | 6.00 | 0.42 | 1.15 | 354 | 0.00 | 6.00 | 0.42 | 1.18 | 587 |
| 16 | 0.00 | 6.00 | 0.51 | 1.34 | 171 | 0.00 | 6.00 | 0.46 | 1.25 | 282 | 0.00 | 6.00 | 0.48 | 1.28 | 453 |
| 16.5 | 0.00 | 6.00 | 0.67 | 1.60 | 129 | 0.00 | 6.00 | 0.61 | 1.44 | 194 | 0.00 | 6.00 | 0.63 | 1.51 | 323 |
| 17 | 0.00 | 6.00 | 0.56 | 1.57 | 71 | 0.00 | 6.00 | 0.68 | 1.65 | 121 | 0.00 | 6.00 | 0.64 | 1.62 | 192 |
| 17.5 | 0.00 | 6.00 | 1.03 | 1.99 | 32 | 0.00 | 6.00 | 0.73 | 1.66 | 59 | 0.00 | 6.00 | 0.84 | 1.78 | 91 |
Bivariate correlations among e-cigarette use, drinking to intoxication, and marijuana use were 0.21 < rs < 0.78, all ps < .01.
Mothers reported their highest level of education completed at baseline across seven categories (1 = Less than 8th grade; 2 = Junior high school; 3 = Partial high school; 4 = High school; 5 = Partial college or associate’s degree; 6 = College graduate; 7 = Graduate degree).
Mothers reported their total household income over the last year (1 = Less than $10,000; 2 = Between $10,000 - $39,999; 3 = Between $40,000 – $79,999; 4 = $80,000 and over).
E-cigarette use was rated on a 7-point scale (0 = Never; 1 = 1-2 times; 2 = 3-5 times; 3 = 6-9 times; 4 = 10-19 times; 5 = 20-39 times; 6 = 40 or more times).
Drinking to intoxication was rated on a 7-point scale (0 = Never; 1 = 1-2 times; 2 = 3-5 times; 3 = 6-9 times; 4 = 10-19 times; 5 = 20-39 times; 6 = 40 or more times).
Marijuana use was rated on a 7-point scale (0 = Never; 1 = 1-2 times; 2 = 3-5 times; 3 = 6-9 times; 4 = 10-19 times; 5 = 20-39 times; 6 = 40 or more times).
Table 2.
Frequencies for E-cigarette Use, Drinking to Intoxication, and Marijuana Use from Ages 13 to 17.5 Years
| Variable | Age | Never | 1-2 times | 3-5 times | 6-9 times | 10-19 times | 20-39 times | 40 or more times | Total N |
|---|---|---|---|---|---|---|---|---|---|
| E-cigarette use | 13 | 135 (95.1%) | 4 (2.8%) | 2 (1.4%) | 1 (0.7%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 142 (100.0%) |
| 13.5 | 229 (92.0%) | 10 (4.0%) | 2 (0.8%) | 4 (1.6%) | 2 (0.8%) | 0 (0.0 %) | 2 (0.8%) | 249 (100.0%) | |
| 14 | 319 (86.2%) | 27 (7.3%) | 9 (2.4%) | 4 (1.1%) | 3 (0.8%) | 1 (0.3%) | 7 (1.9%) | 370 (100.0%) | |
| 14.5 | 412 (80.8%) | 48 (9.4%) | 13 (2.5%) | 11 (2.2%) | 6 (1.2%) | 5 (1.0%) | 15 (2.9%) | 510 (100.0%) | |
| 15 | 478 (77.1%) | 54 (8.7%) | 19 (3.1%) | 20 (3.2%) | 13 (2.1%) | 10 (1.6%) | 26 (4.2%) | 620 (100.0%) | |
| 15.5 | 451 (76.7%) | 47 (8.0%) | 24 (4.1%) | 11 (1.9%) | 19 (3.2%) | 7 (1.2%) | 29 (4.9%) | 588 (100.0%) | |
| 16 | 340 (75.2%) | 35 (7.7%) | 18 (4.0%) | 14 (3.1%) | 10 (2.2%) | 8 (1.8%) | 27 (6.0%) | 452 (100.0%) | |
| 16.5 | 235 (72.5%) | 28 (8.6%) | 14 (4.3%) | 11 (3.4%) | 7 (2.2%) | 5 (1.5%) | 24 (7.4%) | 324 (100.0%) | |
| 17 | 144 (75.4%) | 14 (7.3%) | 10 (5.2%) | 3 (1.6%) | 1 (0.5%) | 3 (1.6%) | 16 (8.4%) | 191 (100.0%) | |
| 17.5 | 72 (79.1%) | 3 (3.3%) | 3 (3.3%) | 2 (2.2%) | 2 (2.2%) | 2 (2.2%) | 7 (7.7%) | 91 (100.0%) | |
| Variable | Age | Never | 1-2 times | 3-5 times | 6-9 times | 10-19 times | 20-39 times | 40 or more times | Total N |
| Drinking to intoxication | 13 | 72 (96.0%) | 3 (4.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 75 (100.0%) |
| 13.5 | 145 (97.3%) | 4 (2.7%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 149 (100.0%) | |
| 14 | 222 (91.4%) | 16 (6.6%) | 3 (1.2%) | 0 (0.0%) | 0 (0.0%) | 2 (0.8%) | 0 (0.0%) | 243 (100.0%) | |
| 14.5 | 316 (87.3%) | 25 (6.9%) | 10 (2.8%) | 5 (1.4%) | 4 (1.1%) | 2 (0.6%) | 0 (0.0%) | 362 (100.0%) | |
| 15 | 360 (79.8%) | 48 (10.6%) | 16 (3.5%) | 10 (2.2%) | 13 (2.9%) | 2 (0.4%) | 2 (0.4%) | 451 (100.0%) | |
| 15.5 | 333 (71.9%) | 65 (14.0%) | 21 (4.5%) | 23 (5.0%) | 12 (2.6%) | 8 (1.7%) | 1 (0.2%) | 463 (100.0%) | |
| 16 | 255 (69.7%) | 52 (14.2%) | 22 (6.0%) | 17 (4.6%) | 16 (4.4%) | 3 (0.8%) | 1 (0.3%) | 366 (100.0%) | |
| 16.5 | 165 (64.5%) | 33 (12.9%) | 24 (9.4%) | 13 (5.1%) | 15 (5.9%) | 5 (2.0%) | 1 (0.4%) | 256 (100.0%) | |
| 17 | 99 (65.6%) | 20 (13.2%) | 14 (9.3%) | 10 (6.6%) | 6 (4.0%) | 1 (0.7%) | 1 (0.7%) | 151 (100.0%) | |
| 17.5 | 50 (64.1%) | 11 (14.1%) | 9 (11.5%) | 4 (5.1%) | 1 (1.3%) | 1 (1.3%) | 2 (2.6%) | 78 (100.0%) | |
| Variable | Age | Never | 1-2 times | 3-5 times | 6-9 times | 10-19 times | 20-39 times | 40 or more times | Total N |
| Marijuana use | 13 | 141 (98.6%) | 1 (0.7%) | 1 (0.7%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 143 (100.0%) |
| 13.5 | 246 (98.8%) | 1 (0.4%) | 0 (0.0%) | 2 (0.8%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 249 (100.0%) | |
| 14 | 352 (94.6%) | 10 (2.7%) | 3 (0.8%) | 2 (0.5%) | 2 (0.5%) | 0 (0.0%) | 3 (0.8%) | 372 (100.0%) | |
| 14.5 | 470 (91.8%) | 12 (2.3%) | 9 (1.8%) | 5 (1.0%) | 7 (1.4%) | 4 (0.8%) | 5 (1.0%) | 512 (100.0%) | |
| 15 | 545 (87.8%) | 27 (4.3%) | 17 (2.7%) | 9 (1.4%) | 9 (1.4%) | 6 (1.0%) | 8 (1.3%) | 621 (100.0%) | |
| 15.5 | 494 (84.2%) | 35 (6.0%) | 19 (3.2%) | 11 (1.9%) | 13 (2.2%) | 3 (0.5%) | 12 (2.0%) | 587 (100.0%) | |
| 16 | 373 (82.3%) | 29 (6.4%) | 17 (3.8%) | 9 (2.0%) | 7 (1.5%) | 8 (1.8%) | 10 (2.2%) | 453 (100.0%) | |
| 16.5 | 255 (78.9%) | 21 (6.5%) | 14 (4.3%) | 7 (2.2%) | 8 (2.5 %) | 5 (1.5%) | 13 (4.0%) | 323 (100.0%) | |
| 17 | 159 (82.8%) | 7 (3.6%) | 3 (1.6%) | 8 (4.2%) | 1 (0.5%) | 3 (1.6%) | 11 (5.7%) | 192 (100.0%) | |
| 17.5 | 68 (74.7%) | 8 (8.8%) | 2 (2.2%) | 2 (2.2%) | 3 (3.3%) | 2 (2.2%) | 6 (6.6%) | 91 (100.0%) | |
3.1. Developmental Trajectories of E-Cigarette Use, Intoxication, and Marijuana Use
3.1.1. Trajectory patterns of e-cigarette use, intoxication, and marijuana use.
LGA zero-inflated Poisson modeling indicated a significant increase in e-cigarette use over time, b = 0.483 (0.094), p < .001 (top panel of Figure 1), and significant individual variances over the 24 months, b = 0.445 (0.042), p < .001. LGA Poisson modeling indicated a significant increase in intoxication over time, b = 0.701 (0.074), p < .001 (top panel of Figure 2), and significant individual variability, b = 0.118 (0.033), p < .001. Similarly, LGA Poisson modeling indicated a significant increase in marijuana use over time, b = 0.739 (0.126), p < .001 (top panel of Figure 3), and individual variability, b = 0.220 (0.025), p < .001.
Figure 1.
Developmental trajectories of e-cigarette use over time, LGA (top); Developmental trajectories of e-cigarette use over time, LCGA (bottom). Trajectory patterns were estimated by adolescent age.
Figure 2.
Developmental trajectories of drinking to intoxication over time, LGA (top); Developmental trajectories of drinking to intoxication over time, LCGA (bottom). Trajectory patterns were estimated by adolescent age.
Figure 3.
Developmental trajectories of marijuana use over time, LGA (top); Developmental trajectories of marijuana use over time, LCGA (bottom). Trajectory patterns were estimated by adolescent age.
3.1.2. Trajectory group membership of e-cigarette use, intoxication, and marijuana use.
LCGA modeling identified different developmental courses that adolescents followed for e-cigarette use, intoxication, and marijuana use. The majority of adolescents followed a never course of e-cigarette use (n = 533; 66.6% of the sample), one fifth followed a low and increasing frequency of e-cigarette use (n = 161; 20.1%), and the remainder followed a high and increasing course (n = 106; 13.3%) (bottom panel of Figure 1). As shown in Table 3, although entropy was highest for the two-group solution, a three-group solution was selected because the BIC value was lowest for the three-group solution, and all other indicators were good. We also believed that the three-group solution would add substantially to the conceptual understanding of group patterns. Specifically, the average latent class assignment probabilities for adolescents were 0.947 for the high and increasing group (i.e., 13.3% of the sample).
Table 3.
Model Fit Indices for Latent Class Growth Analysis (LCGA)
| No. of Group | BIC | LMR-LRT | BLRT | Entropy |
|---|---|---|---|---|
| E-cigarette use | ||||
| 1 class | 6379.449 | NA | NA | NA |
| 2 class | 5536.475 | < .0001 | < .0001 | 0.838 |
| 3 class | 5350.335 | < .0001 | < .0001 | 0.777 |
| 4 class | 5370.389 | .5003 | .5003 | 0.823 |
| Drinking to intoxication | ||||
| 1 class | 5149.033 | NA | NA | NA |
| 2 class | 3706.805 | < .0001 | < .0001 | 0.847 |
| 3 class | 3582.640 | .0001 | .0001 | 0.724 |
| 4 class | 3571.234 | .0549 | .0496 | 0.691 |
| Marijuana use | ||||
| 1 class | 6560.704 | NA | NA | NA |
| 2 class | 4011.708 | < .0001 | < .0001 | 0.934 |
| 3 class | 3739.278 | < .0001 | < .0001 | 0.841 |
| 4 class | 3708.482 | .4343 | .4236 | 0.828 |
Note: BIC, Bayesian information criterion; LMR, Lo–Mendell–Rubin likelihood ratio test; BLRT, bootstrapped likelihood ratio test.
For intoxication and marijuana use, we chose a two-group solution due to their highest entropy and all other indicators were good (Table 3). The BIC value was lowest for the four-group solution for both intoxication and marijuana use, but other indicators such as LMR-LRT and/or BLRT were poor. The majority of adolescents followed a never or low course of intoxication (n = 498; 76.9% of the sample), with the remainder following a moderate increasing frequency of intoxication (n = 150; 23.1%) (bottom panel of Figure 2). With respect to marijuana use, the two developmental courses were: moderate increasing (n = 128; 16.0%) and never or low (n = 673; 84.0%) (bottom panel of Figure 3).
3. 2. Concurrent and Prospective Associations of E-Cigarette Use and Intoxication and Marijuana Use Trajectories.
Conditional LGA modeling examined the concurrent and prospective associations between e-cigarette use and intoxication and marijuana use over time. We treated e-cigarette use as time-varying covariates and controlled for Wave 1 sex, mother’s education, and household income. As shown in Table 4, we observed within-age associations between e-cigarette use and intoxication from ages 13.5 to 16 years (but not at 13 or 16.5 – 17.5). We also observed some evidence of the prospective association between e-cigarette use and subsequent alcohol use. For example, adolescent e-cigarette use at 14 was positively associated with intoxication six months later, and this was also the case for e-cigarette use at ages 15 and 16. Baseline sex predicted the initial status (the intercept) and rate of change (the slope) of adolescent intoxication. On average, whereas males had lower scores on intoxication at initial status than females, their rate of change (i.e., increasing) over the 24 months was somewhat faster. As displayed in Table 5, e-cigarette use was concurrently associated with marijuana use (i.e., within-age), and some prospective association between e-cigarette use and subsequent marijuana use also emerged.
Table 4.
Conditional Linear Growth Analysis (LGA), Drinking to Intoxication
| Drinking to intoxication |
|||
|---|---|---|---|
| Estimate (S.E.) | 95% CI | ||
| Means | |||
| Intercept | −4.771 (1.221)*** | [−7.164, −2.379] | |
| Slope | 0.302 (0.198) | [−0.087, 0.691] | |
| Variances | |||
| Intercept | 4.313 (0.908)*** | [2.534, 6.092] | |
| Slope | 0.028 (0.023) | [−0.016, 0.072] | |
| Covariates predicting the intercept | |||
| Sex | 0.923 (0.330)** | [0.277, 1.570] | |
| Mother education | −0.023 (0.181) | [−0.378, 0.333] | |
| Household income | −0.036 (0.222) | [−0.472, 0.400] | |
| Covariates predicting the slope | |||
| Sex | −0.110 (0.049)* | [−0.206, −0.015] | |
| Mother education | 0.018 (0.028) | [−0.036, 0.072] | |
| Household income | 0.020 (0.035) | [−0.048, 0.087] | |
| Time-varying covariates (concurrent) | |||
| Intoxication 13 <– e-cig 13 | 0.682 (0.350) | [−0.004, 1.369] | |
| Intoxication 13.5 <– e-cig 13.5 | 0.532 (0.213)* | [0.115, 0.950] | |
| Intoxication 14 <– e-cig 14 | 0.514 (0.091)*** | [0.336, 0.692] | |
| Intoxication 14.5 <– e-cig 14.5 | 0.283 (0.073)*** | [0.140, 0.426] | |
| Intoxication 15 <– e-cig 15 | 0.446 (0.058)*** | [0.333, 0.559] | |
| Intoxication 15.5 <– e-cig 15.5 | 0.269 (0.057)*** | [0.158, 0.380] | |
| Intoxication 16 <– e-cig 16 | 0.243 (0.054)*** | 0.137, 0.350] | |
| Intoxication 16.5 <– e-cig 16.5 | 0.079 (0.068) | [−0.055, 0.214] | |
| Intoxication 17 <– e-cig 17 | 0.051 (0.052) | [−0.051, 0.152] | |
| Intoxication 17.5 <– e-cig 17.5 | 0.046 (0.072) | [−0.095, 0.186] | |
| Time-varying covariates (lagged) | |||
| Intoxication 13 <– | -- | -- | |
| Intoxication 13.5 <– e-cig 13 | −0.270 (0.447) | [−1.146, 0.606] | |
| Intoxication 14 <– e-cig 13.5 | 0.104 (0.128) | [−0.147, 0.355] | |
| Intoxication 14.5 <– e-cig 14 | 0.335 (0.161)* | [0.019, 0.650] | |
| Intoxication 15 <– e-cig 14.5 | −0.014 (0.066) | [−0.143, 0.116] | |
| Intoxication 15.5 <– e-cig 15 | 0.171 (0.048)*** | [0.077, 0.264] | |
| Intoxication 16 <– e-cig 15.5 | 0.069 (0.054) | [−0.037, 0.175] | |
| Intoxication 16.5 <– e-cig 16 | 0.173 (0.067)** | [0.043, 0.304] | |
| Intoxication 17 <– e-cig 16.5 | 0.078 (0.075) | [−0.069, 0.225] | |
| Intoxication 17.5 <– e-cig 17 | 0.014 (0.092) | [−0.168, 0.195] | |
Note.
p < .001;
p < .01;
p < .05.
Table 5.
Conditional Linear Growth Analysis (LGA), Marijuana Use
| Marijuana use |
|||
|---|---|---|---|
| Estimate (S.E.) | 95% CI | ||
| Means | |||
| Intercept | −1.425 (1.409) | [−4.187, 1.337] | |
| Slope | −0.185 (0.490) | [−1.146, 0.776] | |
| Variances | |||
| Intercept | 8.222 (1.264)*** | [5.746, 10.699] | |
| Slope | 0.657 (0.106)*** | [0.449, 0.866] | |
| Covariates predicting the intercept | |||
| Sex | 0.146 (0.440) | [−0.716, 1.008] | |
| Mother education | −0.392 (0.253) | [−0.888, 0.104] | |
| Household income | −0.404 (0.312) | [−1.016, 0.209] | |
| Covariates predicting the slope | |||
| Sex | 0.022 (0.154) | [−0.279, 0.324] | |
| Mother education | 0.139 (0.093) | [−0.044, 0.321] | |
| Household income | −0.015 (0.120) | [−0.249, 0.220] | |
| Time-varying covariates (concurrent) | |||
| Marijuana use 13 <– e-cig 13 | −1.033 (0.131)*** | [−1.290, −0.776] | |
| Marijuana use 13.5 <– e-cig 13.5 | 1.016 (0.397)* | [0.238, 1.794] | |
| Marijuana use 14 <– e-cig 14 | 0.701 (0.124)*** | [0.459, 0.943] | |
| Marijuana use 14.5 <– e-cig 14.5 | 0.507 (0.090)*** | [0.330, 0.684] | |
| Marijuana use 15 <– e-cig 15 | 0.440 (0.078)*** | [0.287, 0.594] | |
| Marijuana use 15.5 <– e-cig 15.5 | 0.374 (0.041)*** | [0.293, 0.454] | |
| Marijuana use 16 <– e-cig 16 | 0.424 (0.047)*** | [0.332, 0.516] | |
| Marijuana use 16.5 <– e-cig 16.5 | 0.064 (0.079) | [−0.091, 0.219] | |
| Marijuana use 17 <– e-cig 17 | 0.012 (0.063) | [−0.111, 0.135] | |
| Marijuana use 17.5 <– e-cig 17.5 | 0.192 (0.086)* | [0.023, 0.361] | |
| Time-varying covariates (lagged) | |||
| Marijuana use 13 <– | -- | -- | |
| Marijuana use 13.5 <– e-cig 13 | −0.437 (0.100)*** | [−0.633, −0.241] | |
| Marijuana use 14 <– e-cig 13.5 | −0.059 (0.462) | [−0.964, 0.846] | |
| Marijuana use 14.5 <– e-cig 14 | 0.141 (0.084) | [−0.024, 0.306] | |
| Marijuana use 15 <– e-cig 14.5 | 0.137 (0.093) | [−0.046, 0.319] | |
| Marijuana use 15.5 <– e-cig 15 | 0.210 (0.050)*** | [0.112, 0.308] | |
| Marijuana use 16 <– e-cig 15.5 | 0.053 (0.044) | [−0.033, 0.139] | |
| Marijuana use 16.5 <– e-cig 16 | 0.357 (0.081)*** | [0.199, 0.515] | |
| Marijuana use 17 <– e-cig 16.5 | 0.316 (0.083)*** | [0.153, 0.479] | |
| Marijuana use 17.5 <– e-cig 17 | −0.030 (0.121) | [−0.267, 0.207] | |
Note.
p < .001;
p < .01;
p < .05.
To ensure that the pattern of results was robust, we tested the effects of concurrent and prospective associations using two-level growth modeling with random intercept. On average, both concurrent e-cigarette use (b = 0.247 [0.042], p < .001) and lagged e-cigarette use (b = 0.084 [0.032], p = .009) were associated with adolescent intoxication. E-cigarette use was also associated with marijuana use both concurrently (b = 0.377 [0.051], p < .001) and prospectively (b = 0.226 [0.041], p < .001).
3.3. E-Cigarette Use Trajectory Group Membership Predicting Trajectory Group Membership for Intoxication and Marijuana Use
We used logistic regression to test the relationship between e-cigarette use trajectory patterns (group membership) and intoxication trajectory patterns (Table 6). Compared to adolescents who had never used e-cigarettes, those who belonged to the high and increasing developmental course or to the low and increasing group were more likely to have been intoxicated during the last 6 months.
Table 6.
Drinking to Intoxication as a Function of E-Cigarette Use
| Variable | Drinking to intoxication1 |
|||
|---|---|---|---|---|
| Estimate (S.E.) | 95% CI | Odds Ratio (S.E.) | 95% CI for OR | |
| e-cig: low and increasinga | 1.502 (0.250)*** | [1.012, 1.992] | 4.491 (1.123)** | [2.750, 7.332] |
| e-cig: high and increasingb | 2.891 (0.275)*** | [2.352, 3.429] | 18.002 (4.946)** | [10.507, 30.846] |
| Sexc | 0.375 (0.219) | [−0.054, 0.804] | 1.455 (0.319) | [0.948, 2.235] |
| Mother’s education | 0.151 (0.122) | [−0.088, 0.390] | 1.163 (0.142) | [0.916, 1.478] |
| Household income | −0.013 (0.150) | [−0.307, 0.281] | 0.987 (0.148) | [0.736, 1.324] |
p < .001,
p < .01,
p < .05
The reference class is: never.
The reference class is: never.
The reference class is: never.
The reference group is male. Sex (0 = male; 1 = female)
Logistic regression was also conducted to test the association between e-cigarette use trajectory patterns (group membership) and marijuana use. As shown in Table 7, adolescents who belonged to the high and increasing developmental course or to the low and increasing e-cigarette group were more likely to report increasing marijuana use during the last 6 months compared to those who never used e-cigarettes. The mother reported household income was negatively related to adolescent marijuana use. Adolescents with higher household income were less likely to be in the moderately increasing marijuana group than in the never or low user group.
Table 7.
Marijuana Use as a Function of E-Cigarette Use
| Variable | Marijuana use1 |
|||
|---|---|---|---|---|
| Estimate (S.E.) | 95% CI | Odds Ratio (S.E.) | 95% CI for OR | |
| e-cig: low and increasinga | 1.841 (0.271)*** | [1.310, 2.373] | 6.305 (1.709)** | [3.706, 10.726] |
| e-cig: high and increasingb | 3.389 (0.295)*** | [2.810, 3.968] | 29.629 (8.754)** | [16.604, 52.870] |
| Sexc | 0.098 (0.226) | [−0.345, 0.542] | 1.104 (0.250) | [0.708, 1.719] |
| Mother’s education | 0.054 (0.137) | [−0.215, 0.323] | 1.055 (0.145) | [0.806, 1.381] |
| Household income | −0.580 (0.144)*** | [−0.862, −0.297] | 0.560 (0.081)*** | [0.422, 0.743] |
p < .001,
p < .01,
p < .05
The reference class is: never.
The reference class is: never.
The reference class is: never.
The reference group is male. Sex (0 = male; 1 = female)
4. Discussion
These findings replicate prior research showing concurrent and prospective positive associations between use of e-cigarettes and alcohol (Hershberger, Karyadi, VanderVeen, & Cyders, 2016; Lessard et al., 2014; Roberts et al., 2018) and e-cigarettes and marijuana (Audrain-McGovern et al., 2018; Dai et al., 2018; Passarotti et al., 2015) . Our findings extend the literature by identifying unique trajectory patterns of e-cigarette, alcohol intoxication, and marijuana use, and examining the associations of e-cigarette use trajectories with alcohol and marijuana use trajectories from early to late adolescence using five waves of data over two years. The results indicate that increasing use of e-cigarettes, regardless of whether use is initially relatively low or high, is associated with increasing use of other substances over time. The finding that those who are initially low in e-cigarette use tend to increase use over time and exhibit patterns of alcohol and marijuana use that are comparable to those in the high e-cigarette trajectory group is counter to expectations. This suggests that even low e-cigarette use in early adolescence may be a risk factor, or at least a risk marker, for other substance use. The results highlight the need for early interventions that encompass prevention of e-cigarette use as well as other substances.
The concurrent and prospective associations and similarity in patterns of e-cigarette, alcohol, and marijuana use trajectories indicate that there are groups at high-risk for poly substance use (e-cigarettes, alcohol, and marijuana use). Although adolescents in this sample initiated e-cigarettes slightly earlier than other substances, it is still unclear whether e-cigarette use may be an entry point to other substance use or whether it may be a marker for poly-substance use. Similar patterns and concurrent associations between e-cigarette and alcohol intoxication trajectories indicate that e-cigarette users may share similar profiles with alcohol users, given that 59% of young people had used alcohol by 12th grade, 30% of 12th graders had reported binge drinking during the past 30 days (CDC, 2019; Johnston et al., 2019), and e-cigarettes appeal more to adolescents who are not traditionally “at risk” than combustible cigarettes (Kristjansson, Mann, Smith, & Sigfusdottir, 2018). Moreover, the concurrent associations between e-cigarette and marijuana use trajectories may reflect the increasing pattern of e-cigarette device use for marijuana delivery by adolescents (Trivers, Phillips, Gentzke, Tynan, & Neff, 2018). Prospective associations may be related to the fact that e-cigarettes are perceived by adolescents as less harmful than other substances (Park, Kwon, Gaughan, Livingston & Chang, 2019; Roditis & Halpern-Felsher, 2015). Another possibility is that e-cigarette users may be more likely to affiliate with delinquent peer groups in which these substances are normalized than non-users (Kristjansson et al., 2017), through which they become exposed to other substances. More research is needed to identify the contexts in which adolescents use e-cigarettes.
We examined influences of sex and socioeconomic status (SES) on the alcohol and marijuana use developmental trajectories in addition to e-cigarette use. Females were more likely to drink to intoxication compared to males at baseline, but males’ rate of change (increase) was somewhat faster than that of females in developing patterns over time. Previous research reported inconsistent results regarding the initiation and trajectories of substance use over time depending on sex (Bolland et al., 2016; Chen & Jacobson, 2012; Niño, Cai, Mota-Back, & Comeau, 2017). Higher household income reduced the likelihood of being in the moderate and increasing group for marijuana use. Mother’s education was not a significant factor for development of alcohol and marijuana use, which is consistent with the previous study finding (Dai et al., 2018).
4.1. Limitations
There are some limitations to this study. Although attempts were made to recruit from ethnically and economically diverse populations, the participants were largely affluent and 80% non-Hispanic White students. Importantly, the findings of this study should be interpreted carefully, considering that the prevalence rates of substance use vary among different racial and SES groups. White groups showed higher rates of e-cigarette use (White: 26.8% vs. Black: 20.8%, Hispanic: 14.8%), and alcohol use (White: 32.4% vs. Black: 20.8%, Hispanic: 31.3%), and lower marijuana use (White: 17.7% vs. Black: 25.3%, Hispanic: 23.4%) compared to low-SES and non-White populations (Gentzke et al., 2019; CDC, 2018b). In addition, these results were consistent with national data that reported on the different trajectory patterns of alcohol use among distinctive race groups. White adolescents were more likely to be heavy episodic drinkers than non-White groups after controlling for gender and SES, although no significant difference in marijuana use trajectory was reported depending on race (Park, McCoy, Erausquin, & Bartlett, 2018). Given such differences in substance use related to sociodemographic compositions, the findings may not be generalizable. Future studies should target more diverse samples to examine racial and ethnic differences in the development of adolescent substance use. We used self-reported measures to assess e-cigarette, marijuana, and alcohol use, although self-report using valid instruments is the most commonly used method for assessing these risky problems. Finally, we did not assess whether participants were using e-cigarettes to consume flavoring only, nicotine, or cannabis products, so the purposes of e-cigarette use were not distinguished, and only the associations of e-cigarette use with other substances regardless of contained substances in e-cigarettes were examined. Future studies need to examine the association between e-cigarette and other substances depending on various kinds of substances consumed in e-cigarettes. In fact, there is wide variability in terms of nicotine content, and adolescents are often unaware of exactly what they are vaping. The nature of this sample did not allow us to include conventional cigarette use in our growth models. Very few adolescents reported having tried smoking cigarettes during the last 6 months (32/800, 4.0% at W1; 19/663, 2.9% at W2; 15/628, 2.4% at W3; 16/618, 2.6% at W4; 11/612, 1.8% at W5).
4.2. Implications and Conclusions
This study used five waves of data from early to late adolescence and employed class and trajectory methodologies to understand developmental patterns of e-cigarette, alcohol, and marijuana use over time. Our large sample and longitudinal analysis enabled us to explore whether adolescent e-cigarette use predicts alcohol and marijuana use during adolescence. To our knowledge, this is the first study to reveal how different trajectory patterns of e-cigarette use are associated with marijuana and alcohol use trajectory patterns. We modeled marijuana use and alcohol use behaviors independently, which allowed us to compare the different patterns and factors predicting those behaviors respectively. Finally, the age span of participants over the course of the study was from 13 to 17.5 years old, which provides insights on this phenomenon during the critical period of adolescent substance initiation and use.
The findings of this study have important implications for public health researchers and practitioners. Considering that e-cigarettes are highly accepted among adolescents and e-cigarette use at increasing level increases the chances of substance use, there is a critical need to develop and deliver e-cigarette prevention interventions early in adolescence to prevent the initiation of e-cigarette use. For those who have initiated, interventions targeting multiple substances including e-cigarettes, alcohol, and marijuana are needed to address the longitudinal development of these behaviors. Additional research is needed to understand the factors that predict the initiation and escalation of e-cigarette use (i.e., peer affiliation, parental substance use, substance used with e-cigarettes) and the mechanisms through which e-cigarette use contributes to other substance use. Future studies should investigate profiles of the high-risk developmental groups in order to identify potential contextual factors that contribute to poly use, including e-cigarettes as well as their long-term physical and mental health consequences.
Acknowledgements
The authors wish to thank Jennifer Haas, Cynthia Warthling, Ashley Rupp, and Carrie Pengelly for their assistance with data collection. We would also like to thank Denise Feda and James Zemer for their assistance with data management.
Funding Source
This research was funded by R01 AA021169 awarded to Jennifer A. Livingston by the National Institute on Alcohol Abuse and Alcoholism.
References
- Audrain-McGovern J, Stone MD, Barrington-Trimis J, Unger JB, & Leventhal AM (2018). Adolescent e-cigarette, hookah, and conventional cigarette use and subsequent marijuana use. Pediatrics, 142(3), e20173616 10.1542/peds.2017-3616 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bold KW, Kong G, Camenga DR, Simon P, Cavallo DA, Morean ME, & Krishnan-Sarin S (2018). Trajectories of e-cigarette and conventional cigarette use among youth. Pediatrics, 141(1), e20171832. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bolland KA, Bolland JM, Tomek S, Devereaux RS, Mrug S, & Wimberly JC (2016). Trajectories of adolescent alcohol use by gender and early initiation status. Youth & Society, 48(1), 3–32. 10.1177/0044118X13475639 [DOI] [Google Scholar]
- Brook JS, Zhang C, Leukefeld CG, & Brook DW (2016). Marijuana use from adolescence to adulthood: Developmental trajectories and their outcomes. Social Psychiatry and Psychiatric Epidemiology, 51(10), 1405–1415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Centers for Disease Control and Prevention. (2018a). Notes from the field: Use of electronic cigarettes and any tobacco product among middle and high school students – United States, 2011-2018. Morbidity and Mortality Weekly Report (MMWR), 68(6), 157–164. Retrieved from https://www.cdc.gov/mmwr/volumes/68/wr/mm6806e1.htm?s_cid=mm6806e1_w Accessed March 15, 2019. [Google Scholar]
- Centers for Disease Control and Prevention. (2019). Underage drinking. Retrieved from https://www.cdc.gov/alcohol/fact-sheets/underage-drinking.htm
- Center for Disease Control and Prevention. (2018b). Youth Risk Behavior Surveillance – United State 2017. 67(8), 1–479. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chen P & Jacobson KC (2012). Developmental trajectories of substance use from early adolescence to young adulthood: Gender and racial/ethnic differences. Journal of Adolescent Health, 50(2), 154–163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dai H, Catley D, Richter KP, Goggin K, & Ellerbeck EF (2018). Electronic cigarettes and future marijuana use: A longitudinal study. Pediatrics, 141(5), e20173787 DOI: 10.1542/peds.2017-3787 [DOI] [PubMed] [Google Scholar]
- Duncan SC, Duncan TE, Strycker LA, & Chaumeton NR (2007). A cohort-sequential latent growth model of physical activity from ages 12 to 17 years. Annals of Behavioral Medicine, 33(1), 80–89. 10.1207/s15324796abm3301_9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fulton E, Gokal K, Griffiths S, & Wild S (2018). More than half of adolescent e-cigarette users had never smoked a cigarette: Findings from a study of school children in the UK. Public Health, 161, 33–35. 10.1016/j.puhe.2018.04.014 [DOI] [PubMed] [Google Scholar]
- Gentzke AS, Creamer M, Cullen KA, Ambrose BK, Willis G, Jamal A, & King BA (2019). Vital Signs: Tobacco Product Use Among Middle and High School Students — United States, 2011-2018. MMWR Morb Mortal Wkly Rep 68, 157–164. DOI: 10.15585/mmwr.mm6806e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hershberger AR, Karyadi KA, VanderVeen JD, & Cyders MA (2016). Combined expectancies of alcohol and e-cigarette use relate to higher alcohol use. Addictive Behaviors, 52, 13–21. doi: 10.1016/j.addbeh.2015.08.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ives R, & Ghelani P (2006). Polydrug use (the use of drugs in combination): A brief review. Drugs: Education, Prevention and Policy, 13(3), 225–232. doi: 10.1080/09687630600655596 [DOI] [Google Scholar]
- Johnston LD, Miech RA, O’Malley PM, Bachman JG, Schulenberg JE, & Patrick ME (2019). Monitoring the Future national survey results on drug use, 1975-2018: Overview, key findings on adolescent drug use. Ann Arbor: Institute for Social Research, The University of Michigan. Retrieved from http://monitoringthefuture.org/pubs/monographs/mtf-overview2018.pdf [Google Scholar]
- Kristjansson AL, Mann MJ, Smith ML, & Sigfusdottir ID (2018). Social profile of middle school-aged adolescents who use electronic cigarettes: Implication for primary prevention. Prevention Science, 19, 6, 805–812. [DOI] [PubMed] [Google Scholar]
- Lipperman-Kreda S, & Grube JW (2019). Associations of early age of first intoxication with past year drinking contexts and problems. Substance Use & Misuse, 54(7), 1146–1153. 10.1080/10826084.2018.1563187 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lamminpää A (1995). Alcohol intoxication in childhood and adolescence. Alcohol and Alcoholism, 30(1), 5–12. 10.1093/oxfordjournals.alcalc.a045700 [DOI] [PubMed] [Google Scholar]
- Lessard J, Henrie J, Livingston JA, Leonard KE, Colder CR & Eiden RD (2014). Correlates of ever having used electronic cigarettes among older adolescent children of alcoholic fathers. Nicotine and Tobacco Research, 16, 1656–1660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lo Y, Mendell NR, & Rubin DB (2001). Testing the number of components in a normal mixture. Biometrika, 88(3), 767–778. 10.1093/biomet/88.3.767 [DOI] [Google Scholar]
- Mehta PD, & West SG (2000). Putting the individual back into individual growth curves. Psychological Methods, 5(1), 23–43. 10.1037/1082-989X.5.1.23 [DOI] [PubMed] [Google Scholar]
- Milicic S, & Leatherdale ST (2017). The associations between e-cigarettes and binge drinking, marijuana use, and energy drinks mixed with alcohol. Journal of Adolescent Health, 60(3), 320–327. [DOI] [PubMed] [Google Scholar]
- Muthén LK and Muthén BO (1998-2017). Mplus user’s guide (8th Ed.). Los Angeles, CA: Muthén & Muthén. [Google Scholar]
- National Institute of Alcohol Abuse and Alcoholism (2018). Alcohol facts and statistics. Retrieved from https://www.niaaa.nih.gov/sites/default/files/publications/AlcoholFactsAndStats.pdf
- Niño MD, Cai T, Mota-Back X, & Comeau J (2017). Gender differences in trajectories of alcohol use from ages 13 to 33 across Latina/o ethnic groups. Drug and Alcohol Dependence, 180, 113–120. 10.1016/j.drugalcdep.2017.08.008 [DOI] [PubMed] [Google Scholar]
- Nylund KL, Asparouhov T, & Muthén BO (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling, 14(4), 535–569. 10.1080/10705510701575396 [DOI] [Google Scholar]
- Orlando M, Tucker JS, Ellickson PL, & Klein DJ (2005). Concurrent use of alcohol and cigarettes from adolescence to young adulthood: An examination of developmental trajectories and outcomes. Substance Use & Misuse, 40(8), 1051–1069. [DOI] [PubMed] [Google Scholar]
- Park E, McCoy TP, Erausquin JT, & Bartlett R (2018). Trajectories of risk behaviors across adolescent and young adulthood: the role of race and ethnicity. Addictive Behaviors, 76, 1–7. [DOI] [PubMed] [Google Scholar]
- Park E, Kwon M, Gaughan MR, Livingston JA, & Chang Y-P (2019). Listening to adolescents: their perceptions and information sources about e-cigarettes. Journal of Pediatric Nursing. [DOI] [PubMed] [Google Scholar]
- Passarotti AM, Crane NA, Hedeker D, & Mermelstein RJ (2015). Longitudinal trajectories of marijuana use from adolescence to young adulthood. Addictive Behaviors, 45, 301–308. 10.1016/j.addbeh.2015.02.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Peterson LA, & Hecht SS (2017). Tobacco, e-cigarettes, and child health. Current Opinion in Pediatrics, 29(2), 225–230. doi: 10.1515/ijamh-2016-0033 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Roberts W, Moore KE, Peltier MR, Verplaetse TL, Oberleitner L, Hacker R, & McKee SA (2018). Electronic cigarette use and risk of harmful alcohol consumption in the U.S. population. Alcoholism, Clinical and Experimental Research, 42(12), 2385–2393. 10.1111/acer.13889 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Roditis ML, & Halpern-Felsher B (2015). Adolescents’ perceptions of risks and benefits of conventional cigarettes, e-cigarettes, and marijuana: A qualitative analysis. Journal of Adolescent Health, 57(2), 179–185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Trivers KF, Phillips E, Gentzke AS, Tynan MA, & Neff LJ (2018). Prevalence of cannabis use in electronic cigarettes among US youth. JAMA Pediatrics, 172(11), 1097–1099. doi: 10.1001/jamapediatrics.2018.1920 [DOI] [PMC free article] [PubMed] [Google Scholar]
- US Department of Health and Human Services. (2016). E-Cigarette use among youth and young adults: A report of the Surgeon General. Atlanta, GA: Retrieved from https://www.cdc.gov/tobacco/data_statistics/sgr/e-cigarettes/pdfs/2016_sgr_entire_report_508.pdf/ Accessed 18 January 2019. [Google Scholar]
- Westling E, Rusby JC, Crowley R, & Light JM (2017). Electronic cigarette use by youth: Prevalence, correlates, and use trajectories from middle to high school. Journal of Adolescent Health, 60(6), 660–666. [DOI] [PMC free article] [PubMed] [Google Scholar]
- World Health Organization. (2018). Adolescent alcohol-related behaviors: Trends and inequalities in the WHO European Region, 2002-2014. UK: WHO Regional Office for Europe. [Google Scholar]