Early onset of e-cigarette use and subsequent use frequency among US high school students

Abstract

Objective

The aim of this study was to examine whether the age of e-cigarette use onset predicts subsequent use of e-cigarettes.

Methods

We used the National Youth Tobacco Survey (NYTS) from 2022. Our sample consisted of 4537 US high school students who had ever used e-cigarettes. Age of first e-cigarette use was assessed by a categorical variable ($\le$12 years, 13 years, 14 years, 15 years, 16 years, and $\ge$17 years). We also constructed a binary variable of early onset use ($<$14 years vs $\ge$14 years). E-cigarette use outcomes in the past 30 days included any use and frequent use (used on $\ge$20 days). Weighted multivariable logistic regressions were conducted for each outcome to assess the associations between early onset of e-cigarette use and subsequent use frequency, adjusting for a list of covariates.

Results

Among 4537 high school students who had ever used e-cigarettes, 49.5 % (95 % CI, 46.1 %–52.9 %) reported any use in the past 30 days and 22.8 % (95 % CI, 20.0 %–25.7 %) reported frequent e-cigarette use. Early-onset users, compared with those who tried e-cigarettes at age 14 or older, showed significantly higher risks of any use (aRR = 1.21, 95 % CI, 1.11–1.33) and frequent use (aRR = 1.88, 95 % CI, 1.60–2.20) in the past 30 days. We found younger age at first use to be associated with higher risk of current and frequent use.

Conclusions

Our findings highlight the importance for age-sensitive efforts, prioritizing younger adolescents, to prevent and delay e-cigarette use initiation.

1. Introduction

E-cigarette has become the most popular nicotine-containing product among US middle and high school students since 2014. In 2023, the prevalence of past 30-day e-cigarette use was 4.6 % (95 % CI, 3.6 %–5.8 %) for middle school students and 10.0 % (95 % CI, 8.8 %–11.4 %) for high school students. (Birdsey, 2023) Although these rates have declined substantially from the peak in 2019, (Cullen et al., 2019) a considerable proportion of e-cigarette users report frequent or daily use. For example, 40.0 % of high school e-cigarette users reported use on $\ge$20 days in the past 30 days and 29.9 % reported daily use in 2023. (Birdsey, 2023) This pattern is concerning because more frequent e-cigarette use is associated with higher levels of nicotine exposure, (Dai et al., 2024) potentially leading to elevated risks of nicotine dependence due to reinforcing properties of nicotine on developing brains. (Leslie, 2020) Other dangers of e-cigarette use include increased risks of respiratory and heart disease. (Peruzzi et al., 2020; Sun and Oates, 2024) In addition, dual cigarette and e-cigarette use is associated with higher odds of disease than exclusive smoking. (Pisinger and Rasmussen, 2022).

It is well-established for some substances that the age of initiation is a key factor in predicting subsequent use and long-term health risks. (Chen et al., 2009; Lopez-Quintero et al., 2011; American Lung Association, 2023) For example, 87 % of adult daily cigarette smokers tried smoking by age 18 and 95 % by age 21. (American Lung Association, 2023) Across various substances such as alcohol and cannabis, studies have found that onset of use in adolescence, compared with initiation in adulthood, is associated with higher likelihood of developing dependence or abuse. (Chen et al., 2009; Lopez-Quintero et al., 2011) In particular, initiation during early adolescence may exhibit even greater risks than onset in late adolescence or adulthood. (Hawke et al., 2020) Onset of substance use during adolescence is also strongly associated with lower academic achievement and occupational attainment. (Horwood et al., 2010).

However, the age of e-cigarette initiation is less studied. A few studies have investigated the role of e-cigarette initiation age on subsequent frequency of e-cigarette use, with inconclusive findings. (Bold et al., 2017; Conner et al., 2021; Harrell et al., 2021) While two studies found that younger age of e-cigarette initiation was associated with increased frequency of subsequent use, (Bold et al., 2017; Harrell et al., 2021) another one found the association to be non-significant after adjusting for relevant covariates. (Conner et al., 2021) In addition, all of these studies analyzed data collected between 2014 and 2019, before the recent decline in e-cigarette use. Two of these studies investigated adolescents in the US but included only selected adolescents in Connecticut (Bold et al., 2017) or Texas (Harrell et al., 2021). To the best of our knowledge, no studies have yet analyzed a national sample of US high school students to investigate the age of e-cigarette use onset and frequency of use. To fill these gaps in the literature, this study examined the association between early onset of e-cigarette use and subsequent frequency of use among a nationally representative sample of US high school students, using data from the National Youth Tobacco Survey (NYTS) in 2022.

2. Methods

2.1. Sample

NYTS is an annual survey that provides nationally representative data about middle and high school students' tobacco-related belief, attitudes, behaviors, and exposure to pro- and anti-tobacco influences. In 2022, NYTS was administered as a web-based survey, adopting a stratified, three-stage cluster sample design to select participants. The overall participation rate was 45.2 %. Using NYTS 2022 data, our sample consisted of 4537 high school students from grades 9–12 who had ever used e-cigarettes. The University of Alabama at Birmingham Institutional Review Board exempted this study because it used deidentified data.

2.2. Measures

2.2.1. Independent variable

We identified the age of e-cigarette use onset from the question, “How old were you when you first used an e-cigarette, even once or twice?” Response options were: “8 years or younger”, “9 years”, “10 years”, “11 years”, “12 years”, “13 years”, “14 years”, “15 years”, “16 years”, “17 years”, “18 years”, and “19 years or older”. Due to sample size, we categorized these answers into the following groups: $\le$12 years, 13 years, 14 years, 15 years, 16 years, and $\ge$17 years. In addition, following previous literature, (McCabe et al., 2018; Conner et al., 2021) we constructed a binary measure of early onset use, considering those who began using e-cigarettes before age 14 as early-onset users ($<$14 vs $\ge$14).

2.2.2. Outcome

Current e-cigarette use (0 vs 1) was categorized by any positive answer to the question, “During the past 30 days, on how many days did you use e-cigarettes?” Frequent use was defined as using e-cigarettes on $\ge$20 days in the past 30 days.

2.2.3. Covariates

Study covariates included self-reported demographic characteristics of age, sex (female vs male), race (white, black, and other), and ethnicity (Hispanic vs non-Hispanic). Other in race consisted of American Indian or Alaska Native, Asian, or Native Hawaiian or Other Pacific Islander. We also controlled for past 30-day other nicotine-containing product use (0 vs 1), covering cigarettes, cigars, chewing tobacco/snuff/dip, hookah, roll-your-own cigarettes, pipe, snus, dissolvable tobacco products, bidis, heated tobacco products, and nicotine pouches. Family nicotine use (0 vs 1) was given a value of 1 if anyone who lived with the participant now used any nicotine-containing products listed above or e-cigarettes. Mental health status was assessed by depression screening (0 vs 1) and anxiety screening (0 vs 1), adopting the Patient Health Questionaire-2 (PHQ-2) and the Generalized Anxiety Disorder-2 (GAD-2) respectively. Both PHQ-2 and GAD-2 contained two questions. A score $\ge$3 was used to identify a positive outcome of depression or anxiety screening. The Appendix Table 1 presents detailed questions.

2.3. Analysis

We first calculated the weighted prevalence of e-cigarette use and sample characteristics, by early onset use. Then we conducted multivariable logistic regressions using a complete case analysis to examine the association between early onset of e-cigarette use and subsequent use among high school students in the US, controlling for covariates mentioned above. The analyses were performed using Stata version 18, incorporating survey weights. We reported logistic regression results as adjusted relative risks (aRRs). Other studies have reported the association in adjusted odds ratios. (McCabe et al., 2018; Conner et al., 2021) However, odds ratios are commonly misinterpreted as relative risks and would produce biased estimates of relative risks due to the high prevalence of our outcomes. (Norton et al., 2018).

3. Results

Among 4537 high school students who had ever used e-cigarettes, 49.5 % (95 % CI, 46.1 %–52.9 %) reported current use and 22.8 % (95 % CI, 20.0 %–25.7 %) reported frequent use. Early-onset users had significantly higher prevalence of any use (58.2 % vs 45.1 %, p < .001) and frequent use (34.0 % vs 17.1 %, p < .001) in the past 30 days.

Table 1 describes the sample characteristics by early onset use. Around half (51.9 %, 95 CI, 49.3 %–54.5 %) of the overall sample were female, 70.9 % (95 % CI, 65.6 %–75.7 %) were white, and 25.2 % (95 % CI, 20.5 %–30.7 %) were Hispanic. 19.7 % (95 % CI, 16.5 %–23.3 %) used other nicotine-containing products in the past 30 days, 51.7 % (95 % CI, 48.5 %–54.8 %) lived with family who used nicotine, 35.2 % (95 % CI, 32.8 %–37.7 %) tested positive for depression screening, and 40.0 % (95 % CI, 37.6 %-42.4) for anxiety screening. Comparing early-onset users with those who were not, early-onset users were significantly younger (p < .001), more likely to report past 30-day use of other nicotine-containing products (p < .001), have family member who used nicotine (p < .001), and higher chance of testing positive for depression screening (p = .01) and anxiety screening (p = .02).

Table 1.

Descriptive characteristics of US high school students who had ever used e-cigarettes, from the National Youth Tobacco Survey in 2022.

Ever e-cigarette users in high school	Overall	Early-onset usersa	Non-early-onset users
	Weighted %, 95 % CI	Weighted %, 95 % CI	Weighted %, 95 % CI	P-valueb
Past 30 Days Use
Any use	49.5 (46.1–52.9)	58.2 (54.0–62.2)	45.1 (41.1–49.1)	<0.001
Frequent usec	22.8 (20.0–25.7)	34.0 (29.6–38.6)	17.1 (14.7–19.7)	<0.001
Characteristics
Age				<0.001
$\le$13 years	0.3 (0.1–0.7)	0.6 (0.2–1.8)	0.0 (0.0–0.0)
14 years	6.5 (5.1–8.3)	12.6 (10.1–15.7)	3.3 (2.4–4.6)
15 years	19.1 (17.0–21.4)	27.0 (23.0–31.4)	14.9 (12.6–17.5)
16 years	27.1 (24.2–30.2)	29.2 (25.0–33.8)	26.0 (23.2–29.1)
17 years	28.5 (25.9–31.2)	18.9 (15.5–22.7)	33.5 (30.6–36.4)
$\ge$18 years	18.6 (15.6–22.0)	11.7 (8.3–16.3)	22.3 (18.9–26.1)
Sex				0.87
Female	51.9 (49.3–54.5)	51.8 (48.5–55.2)	52.2 (49.2–55.1)
Male	48.1 (45.6–50.7)	48.2 (44.8–51.5)	47.9 (4498–51.0)
Race				0.30
White	70.9 (65.6–75.7)	70.1 (64.4–75.2)	71.5 (65.3–76.9)
Black	16.2 (12.4–20.7)	15.3 (11.8–19.6)	16.6 (12.3–22.0)
Otherd	12.9 (10.6–15.6)	14.6 (11.7–18.2)	12.0 (9.7–14.8)
Ethnicity				0.99
Hispanic	25.2 (20.5–30.7)	25.5 (20.3–31.6)	25.1 (20.2–30.7)
Non-Hispanic	74.8 (69.3–79.6)	74.5 (68.4–79.7)	75.0 (69.3–79.8)
Past 30-day other nicotine-containing product usee				<0.001
Yes	19.7 (16.5–23.3)	27.6 (22.7–33.1)	15.6 (12.5–19.3)
No	80.3 (76.7–83.5)	72.4 (67.0–77.3)	84.4 (80.7–87.5)
Family nicotine usef				<0.001
Yes	51.7 (48.5–54.8)	57.4 (53.7–61.0)	48.8 (44.9–52.7)
No	48.3 (45.2–51.5)	42.6 (39.0–46.3)	51.2 (47.3–55.1)
Depression Screeningg				0.01
Positive	35.2 (32.8–37.7)	39.3 (35.8–42.9)	33.2 (30.3–36.3)
Not positive	64.8 (62.4–67.2)	60.8 (57.1–64.3)	66.8 (63.7–70.0)
Anxiety Screeningh				0.02
Positive	40.0 (37.6–42.4)	44.8 (39.4–50.4)	37.7 (35.2–40.3)
Not positive	60.0 (57.6–62.4)	55.2 (49.6–60.6)	62.3 (59.7–64.8)

Notes

^aEarly onset use is defined by first e-cigarette use before age 14 ($<$14 vs $\ge$14).

^bP-value was calculated by comparing the distribution of the characteristic between early-onset users and non-early-onset users.

^cFrequent use is defined by reported e-cigarette use on $\ge$20 days in the past 30 days.

^dOther includes American Indian or Alaska Native, Asian, or Native Hawaiian or Other Pacific Islander.

^eOther nicotine-containing products include cigarettes, cigars, chewing tobacco/snuff/dip, hookah, roll-your-own cigarettes, pipe, snus, dissolvable tobacco products, bidis, heated tobacco products, and nicotine pouches.

^fFamily nicotine use is identified if anyone who lives with the participant now uses e-cigarettes, cigarettes, cigars/cigarillos/little cigars, chewing tobacco/snuff/dip, hookah or waterpipe, roll-your-own cigarettes, pipe, snus, dissolvable tobacco products, bidis, heated tobacco products, or nicotine pouches.

^gDepression screening outcome is identified using the Patient Health Questionaire-2 (PHQ-2) with a cutoff score of 3.

^hAnxiety screening outcome is identified using the Generalized Anxiety Disorder-2 (GAD-2) with cutoff score of 3.

The adjusted association between age of first e-cigarette use and current use was reported in Table 2. Early-onset users, compared with those who tried e-cigarettes at age 14 or older, showed significantly higher risks of any use (aRR = 1.21, 95 % CI, 1.11–1.33) and frequent use (aRR = 1.88, 95 % CI, 1.60–2.20) in the past 30 days. Using first use at age 17 or older as the reference group, younger age at first use was almost all significantly associated with higher risk of current use, with aRRs of 1.63 (95 % CI, 1.31–2.02) for those started $\le$12 years, 1.64 (95 % CI, 1.33–2.03) for first use at 13 years, 1.51 (95 % CI, 1.16–1.97) for 14 years, 1.30 (95 % CI, 1.02–1.67) for 15 years, and 1.18 (95 % CI, 0.92–1.50) for 16 years. Similarly, we found younger age at first use to be associated with higher risk of frequent use.

Table 2.

Association between early onset of e-cigarette use and subsequent e-cigarette use in the past 30 days, among US high school students who had ever used e-cigarettes (2022 National Youth Tobacco Survey).

	E-cigarette use in the past 30 days
	Any use aRR (95 % CI)	Frequent usea aRR (95 % CI)
Early onset ($<$14 vs $\ge$ 14 years)
Yes	1.21 (1.11–1.33)	1.88 (1.60–2.20)
No	Ref	Ref
Age at first use
$\le$12 years	1.63 (1.31–2.02)	4.80 (2.59–8.91)
13 years	1.64 (1.33–2.03)	1.30 (1.87–7.23)
14 years	1.51 (1.16–1.97)	1.18 (1.63–6.55)
15 years	1.30 (1.02–1.67)	2.06 (0.94–4.50)
16 years	1.18 (0.92–1.50)	1.11 (0.55–2.23)
$\ge$ 17 years	Ref	Ref

Notes

Multivariable logistic regressions adjusted for all study covariates: age ($\le$13, 14, 15, 16, 17, $\ge$18 years old), sex (female vs male), race (white, black, other), ethnicity (Hispanic vs non-Hispanic), past 30-day other nicotine-containing product use (including cigarettes, cigars, chewing tobacco/snuff/dip, hookah, roll-your-own cigarettes, pipe, snus, dissolvable tobacco products, bidis, heated tobacco products, and nicotine pouches), family nicotine use, depression, and anxiety.

^aFrequent use is defined by reported e-cigarette use on $\ge$20 days in the past 30 days.

4. Discussion

Using a nationally representative sample of US high school students in 2022, we found that early onset of e-cigarette use among ever e-cigarette users is significantly associated with higher risks of current and frequent e-cigarette use reported in the past 30 days. These findings are consistent with two previous studies. (Bold et al., 2017; Harrell et al., 2021) Bold et al. examined high school students in southeastern Connecticut in 2015 and reported that age of e-cigarette initiation was negatively associated with the number of days of e-cigarette use in the past 30 days. Using longitudinal data from adolescents in Texas between 2014 and 2019, Harrell et al. compared average e-cigarette use frequency between early initiators (13–14 years old), mid initiators (14–15 years old), and late initiators (15–16 years old). They found an overall trend that earlier age of initiation was associated with subsequent higher frequency of use. Our study extends previous research by examining a nationally representative sample of high school students with more recent data on e-cigarette use. Given the rapid development of vaping products and the impact of e-cigarette related policies, it is likely that adolescent utilization patterns are changing over time. Future studies are needed to further examine the association of early e-cigarette use onset and subsequent use frequency.

Our study results have significant implications for public health policies. The strong association between the early onset of e-cigarette use and subsequent use shows that targeting younger adolescents is critical for preventing e-cigarette use. One study, analyzing US adolescents between 2014 and 2021, found a decrease in the age at initiation of e-cigarette use. (Glantz et al., 2022) With adolescents initiating e-cigarette use at earlier ages, it is urgent to prioritize our prevention efforts to focus on younger adolescents.

In addition to prevention, efforts to delay e-cigarette initiation could also yield positive health benefits by reducing frequency of e-cigarette use and future nicotine dependence. Although no studies have yet estimated the impact of delaying the age of e-cigarette initiation, studies on other substances have revealed sizable effect. (Azagba et al., 2015; Jordan and Andersen, 2017; Rioux et al., 2018) For example, each year of delayed cannabis use onset was associated with a 31 % reduction in the odds of developing any drug abuse symptoms by age 28. (Rioux et al., 2018) In general, a one-year delay in the age of substance use initiation is associated with a decrease of 4–5 % in the likelihood of lifetime substance use disorder. (Jordan and Andersen, 2017) According to the 2023 report from Monitoring the Future survey, e-cigarette was the second most popular substance ever used by 8th graders. (Miech et al., 2024) Many adolescents likely tried e-cigarettes before other substances. As a result, delaying the age of e-cigarette use onset could postpone general substance use for some adolescents.

This study has a few limitations. First, self-reported cross-sectional data from NYTS are subject to biases such as the social desirability bias and recall bias. Future studies with longitudinal design could provide estimates of the association with limited recall bias. Second, because NYTS data were collected from students in schools, the findings might not be generalizable to those who are home-school, in detention centers, or have dropped out of school. Lastly, the COVID-19 pandemic from 2020 to 2023 may affect our findings from 2022 data. While some high school students reduced their consumption due to limited access to e-cigarettes, (Kreslake et al., 2021) others increased their e-cigarette consumption to cope with the pandemic. (Bennett et al., 2023) Future study is needed to identify potential changes in e-cigarette utilization patterns after the pandemic.

In summary, we found that high school students with early onset of e-cigarette use, compared with those who initiated later, were significantly more likely to report subsequent e-cigarette use. These results highlight the importance for age-sensitive efforts, prioritizing younger adolescents, to prevent and delay e-cigarette use initiation.

CRediT authorship contribution statement

Ruoyan Sun: Writing – original draft, Formal analysis, Conceptualization. Nengjun Yi: Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Appendix Table 1

Data availability

Data will be made available on request.