Electronic Vapor Product Use and Sexual Risk Behaviors in US Adolescents

In this study, we used data from the 2017 National YRBS. The National YRBS is a cross-sectional study conducted every other year by the US Centers for Disease Control and Prevention to assess US high school students’ engagement in health-related risk behaviors. The 2017 National YRBS employed a 3-stage cluster sample design to generate a nationally representative sample of ninth- through 12th-grade students in public and private schools.⁸  The survey questionnaire was self-administered, and student participation was voluntary and anonymous. Data were weighted by sex, racial/ethnic background, and grade and adjusted for school and student nonresponse and for the oversampling of Black and Hispanic students to yield a nationally representative sample.⁸  Further information about the YRBS methodology has been published.⁹  Because we used publicly available, deidentified data, this study was exempt from institutional review board approval.

Current EVP and cigarette use were determined by the following questions: “During the past 30 days, on how many days did you use an electronic vapor product?” and “During the past 30 days, on how many days did you smoke cigarettes?” Examples of EVP brands and products were provided to respondents. Response options for the EVP and cigarette use frequency questions were the following: 0, 1 to 2, 3 to 5, 6 to 9, 10 to 19, 20 to 29, or all 30 days.

Respondents were categorized into 1 of 4 mutually exclusive groups on the basis of their past 30-day EVP and cigarette usage: did not use either EVPs or cigarettes (nonuse), used EVPs on ≥1 day but did not use cigarettes (EVP use only), used cigarettes on ≥1 day but did not use EVPs (cigarette use only), or used cigarettes on ≥1 day and EVPs on ≥1 day (dual use). Respondents were also categorized by their previous 30-day EVP use alone into 1 of 4 mutually exclusive groups: did not use EVPs (nonuse), used EVPs 1 to 9 days (occasional use), used EVPs 10 to 29 days (frequent use), used EVPs all 30 days (daily use).

Ten SRBs included in the 2017 YRBS were identified as dependent variables, and responses were dichotomized (Table 1).

Four demographic variables were included as confounders in the regression models: sex (male or female), grade (9, 10, 11, or 12), racial/ethnic background, and sexual orientation. These variables were selected on the basis of previous literature about sociodemographic factors and tobacco use.^7,8  Respondents were categorized into 1 of 6 mutually exclusive race and/or ethnicity categories: white, non-Hispanic; Black, non-Hispanic; Hispanic or Latino; multiple races, Hispanic; multiple races, non-Hispanic; or other. “Other” included ethnic groups for which the number of respondents was too small to support meaningful analysis: Asian, Native American and Native Alaskan, and Native Hawaiian and Pacific Islander. Respondents self-identified sexual orientation as one of the following: heterosexual or straight, gay or lesbian, bisexual, or not sure.

Respondents’ current usage of smokeless tobacco (including chewing tobacco, snuff, dip, snus, and dissolvable tobacco) and cigars (including cigarillos and little cigars) was also included as categorical confounding variables in all models. Response options for the smokeless tobacco and cigar use frequency questions were the following: 0, 1 to 2, 3 to 5, 6 to 9, 10 to 19, 20 to 29, or all 30 days.

The percentage of respondents in each category of previous 30-day EVP and/or cigarette use was calculated overall and by sex, racial/ethnic background, grade, and sexual orientation. Second-order Rao-Scott adjusted χ² tests were performed to assess differences by demographic category. A modified Poisson regression model,¹⁰  adjusted for the confounders listed above, produced adjusted prevalence ratios (aPRs) and 95% confidence intervals (CIs) for the associations between past 30-day EVP and/or cigarette usage and the 10 SRBs. Nonusers of both EVPs and cigarettes served as the reference group. A second modified Poisson regression, adjusted for the confounders plus categorical past 30-day cigarette usage, generated aPRs and 95% CIs for the associations between past 30-day EVP usage frequency and the 10 SRBs. Nonusers of EVPs constituted the reference group for the second regression.

Linear contrasts were conducted to compare aPRs in each regression model across categories of EVP and/or cigarette use and EVP use frequency. The following comparisons were made in the first regression: EVP-only use versus cigarette-only use, dual use versus cigarette-only use, and dual use versus EVP-only use. In the second regression, comparisons were made for the following EVP use frequency categories: frequent use versus occasional use, daily use versus occasional use, and daily use versus frequent use. P values from significance testing of linear contrasts were adjusted by using the Bonferroni correction for multiple comparisons.

The svy suite of commands in Stata 14.2 (Stata Corp, College Station, TX) was used to perform all statistical analyses. This software package accounted for the YRBS complex survey design. For all tests, a P value of <.05 was considered statistically significant; all statistical tests were 2-tailed.

A total of 14 765 students were included in the 2017 National YRBS sample. One hundred forty-four schools participated in the survey, with a school response rate of 75%. The student response rate was 81%, and the overall response rate was 60%. Demographic characteristics of the National YRBS sample have been described elsewhere.⁸ 

A total of 12 667 students responded to both the current cigarette use and the current EVP use questions. In 2017, 84.2% of US high school students used neither cigarettes nor EVPs in the past 30 days, 6.5% used EVPs only, 2.8% used cigarettes only, and 6.5% used both products (Table 2). Prevalence of EVP-only use and dual use was higher among male students (8.2% and 7.6%, respectively) than female students (4.8% and 5.4%, respectively). Prevalence of cigarette use, EVP-only use, and dual use was highest among non-Hispanic white students, and prevalence of all 3 activities increased by grade.

EVP-only use was most prevalent among heterosexual high schoolers at 7.0%, with prevalence of EVP-only use among gay or lesbian and bisexual students at 5.4% and 5.6%, respectively. Cigarette-only use was most prevalent among gay or lesbian and bisexual students at 5.5% for both groups, compared with 2.6% prevalence for heterosexual students. Prevalence of dual use was higher among gay or lesbian students (9.0%) and highest among bisexual students (12.5%), compared with heterosexual students (6.0%).

Absolute prevalence of SRB engagement generally increased from nonusers to EVP-only users to cigarette-only users, with absolute prevalence highest among dual users. Dual users were more likely than nonusers to engage in all 10 SRBs examined. EVP-only and cigarette-only users were more likely than nonusers to engage in 9 of the 10 SRBs (Table 3).

EVP and/or cigarette users in all 3 categories were twice as likely as nonusers to have ever engaged in sexual intercourse and to have engaged in intercourse in the previous 3 months (aPR range: 2.07–2.59), were at least twice as likely to have experienced sexual debut at <13 years (aPR range: 2.68–3.76) and at <16 years (aPR range: 2.48–3.09), and were at least 2.5 times more likely to have had ≥4 lifetime sexual partners (aPR range: 2.93–5.31) and to have consumed alcohol or used drugs before the most recent intercourse (aPR range: 2.64–3.62). Whereas cigarette-only users and dual users were slightly more likely than nonusers to have not used a condom during the most recent sexual intercourse (aPR range: 1.30–1.44), EVP-only users were similarly as likely as nonusers to have not used a condom (aPR: 0.96; 95% CI: 0.81–1.14).

Users of EVPs and/or cigarettes were also more likely than nonusers to have experienced sexual violence, including forced sexual intercourse, in one’s lifetime (aPR range: 1.92–2.75) and sexual violence in the previous 12 months (aPR range: 1.92–2.52). EVP-only users and dual users were more likely than nonusers to have experienced sexual dating violence in the previous 12 months (aPR range: 1.69–1.79), whereas there was no difference for this outcome between cigarette-only users and nonusers (aPR: 1.14; 95% CI: 0.72–1.80).

Modest differences in the likelihood of SRB engagement were found among the EVP and cigarette use categories. Dual users were more likely than EVP-only users to have had ≥4 lifetime sexual partners, used alcohol or drugs before the most recent intercourse, and not used a condom during the most recent intercourse. Dual users were also more likely than cigarette-only users to have engaged in intercourse in the past 12 months. Cigarette-only users were more likely than EVP-only users to have had ≥4 lifetime sexual partners and to have not used a condom during the most recent intercourse. For the remaining 6 behaviors, there was no difference in prevalence ratios among EVP-only users, cigarette-only users, and dual users.

Absolute prevalence of SRB engagement was generally lowest among EVP nonusers and highest among daily EVP users, with occasional and frequent EVP users at an intermediate absolute prevalence. Occasional EVP users were more likely than nonusers to engage in 9 of 10 SRBs (Table 4).

EVP users at any past 30-day frequency were more likely than nonusers to have ever engaged in sexual intercourse and to have engaged in intercourse in the previous 3 months (aPR range: 1.63–2.35), were more likely to have experienced sexual debut at <16 years (aPR range: 1.86–2.22), were more likely to have had ≥4 lifetime sexual partners (aPR range: 2.28–2.68), and were more likely to have used alcohol or drugs before the most recent intercourse (aPR range: 1.89–2.51). Whereas occasional and daily EVP users were more likely than nonusers to have experienced sexual debut at <13 years (aPR range: 2.67–2.79), frequent users were not more likely than nonusers to have done so (aPR: 1.65; 95% CI: 0.89–3.06). There was no difference between EVP users at any frequency and EVP nonusers in likelihood of condom use during the most recent intercourse (aPR range: 0.89–1.05).

EVP users at any previous 30-day frequency were more likely than nonusers to have experienced sexual violence in the previous 12 months (aPR range: 1.88–2.40). Of the other 2 sexual violence behaviors, only occasional EVP users were more likely than EVP nonusers to have ever been forced into sexual intercourse (aPR: 1.75), and only occasional and daily EVP users were more likely than nonusers to have experienced sexual dating violence in the past 12 months (aPR range: 1.78–2.23).

Only 1 linear contrast of EVP usage frequency categories yielded a significant result: the aPR for the association between vaping and sexual intercourse was smaller for daily EVP users than for frequent EVP users (but not occasional users). For the 9 other SRBs, no difference in aPRs was found among occasional, frequent, and daily EVP users.

Our results reveal that US high school students who currently use EVPs and/or cigarettes are more likely than nonusers to engage in a wide range of SRBs. Previous studies have documented associations between traditional cigarette use and SRBs as well as EVP use and various risk behaviors, such as substance use, violent behaviors, and poor exercise and dietary habits.^5–7  This analysis identifies 7 new risky behaviors associated with nicotine-containing EVP use: early sexual debut (at age <13 and age <16), use of alcohol and/or drugs before the most recent intercourse, lifetime sexual intercourse engagement, and experiencing forced intercourse, sexual violence, or dating violence. Only 1 of 10 SRBs examined (lack of condom use at the last intercourse) was not associated with EVP-only use, a result that is consistent with a previous study.⁵  However, cigarette-only and dual use were associated with lack of condom use at the most recent intercourse.

Our results corroborate past evidence of a gradient of risk behavior engagement for adolescent nicotine users^5,11,12 : nonusers had the lowest absolute prevalence of SRB engagement, followed by EVP-only users, cigarette-only users, and, finally, dual users. However, previous research has found that youth EVP-only users are more likely to initiate cigarette smoking or dual use, thus moving up the risk gradient and increasing the likelihood of developing other substance use behaviors.¹³ 

Our findings also indicate that youth who used EVPs at any past 30-day frequency were more likely than nonusers to engage in 6 of the 10 SRBs examined, and students who used EVPs only occasionally were more likely than nonusers to engage in 9 of 10 SRBs. This result is consistent with the previous finding that youth likelihood of engagement in several substance use risk behaviors does not vary among occasional, frequent, or daily EVP users and supports medical and public health experts’ efforts to reduce even experimental EVP use among adolescents.^6,14,15  Furthermore, this study lends evidence to the “risk behavior syndrome” theory that youth EVP use clusters with other risky activities, such as drug use and SRBs.¹²  Proposed drivers of the risk behavior syndrome include low perceived social support from family and peers as well as underlying mental health conditions.¹⁶ 

Prevalence of dual EVP and cigarette use was particularly high among US high school students who self-identified as lesbian, gay, or bisexual. This result adds to previous findings of heightened nicotine (including EVP) use among lesbian, gay, bisexual, transgender, queer (LGBTQ+) adolescents.^8,17  Other researchers have hypothesized that sexual-minority youth may be more likely to use nicotine products because of stigma and discrimination from family and society as well as a lack of targeted prevention programs to counteract tobacco marketing to the LGBTQ+ community.^17,18  Because sexual-minority youth are also more likely to engage in SRBs,^19,20  analysis of the associations between EVP use and SRBs specifically among LGBTQ+ youth is an important area of future research.

The strengths of this study include the large number of novel SRB outcomes studied as well as the large nationally representative sample. However, the study is limited by its cross-sectional design, which prohibits causal inference or determination of the temporality of risk behavior engagement. For example, lifetime SRBs may have occurred before EVP and/or cigarette initiation. In previous research, investigators have been unable to infer causality in the SRB-substance use relationship as well.²¹  It is possible that unmeasured variables (such as knowledge about substance use disorders and SRBs, peer influence, heightened activation of the social reward network, mental health conditions, and social vulnerabilities, such as bullying) explain both youth EVP use and SRB engagement.^7,22–24  Additional research is needed to identify adolescent motivations for EVP use and engagement in SRBs.

Other limitations include possible underreporting of youth risk behavior engagement by respondents and the fact that the YRBS is not representative of youth who do not attend school. However, studies have revealed that YRBS respondents tend to report their behaviors reliably, and in 2013, only 5% of all US 16- and 17-year-olds were not enrolled in high school and did not have a high school diploma or equivalent.^25,26  Finally, it is possible that the YRBS questionnaire language did not fully and accurately capture SRB engagement. For example, because “sexual intercourse” was not defined in the survey, students may not have interpreted this behavior to include both same-sex and heterosexual intercourse, or they may have considered other sexual activities when answering questions about intercourse.

The American Academy of Pediatrics has proposed several policy changes to reduce youth access to and use of EVPs, including increasing the minimum age to purchase tobacco to 21 and banning youth-friendly EVP flavorings; some of these recommendations have been implemented.^15,27,28  However, EVPs remain the most popular tobacco product among US youth, with prevalence of use increasing year on year.^1,29  Enhanced regulatory enforcement efforts are needed to prevent retailers (including e-retailers) from selling EVPs to adolescents and to dissuade EVP manufacturers from employing youth-accessible marketing strategies. Moreover, continued public education initiatives through media and schools are essential to ensuring that youth and families are aware of the health consequences of EVP use.^15,30 

At the patient-provider level, adolescent health practitioners can contribute to reducing youth vaping by screening patients for EVP use and exposure, offering preventive counseling, and providing treatment and/or resources for youth who are tobacco dependent.^31,32  Moreover, because American Academy of Pediatrics guidelines suggest assessing adolescents’ nicotine use as part of regular sexual and reproductive health care,³³  this assessment presents an excellent opportunity for health practitioners to offer information and counseling to youth and their families about the associations between SRBs and EVP use at any frequency.

US adolescents who use nicotine-containing EVPs, with or without concurrent cigarette use, are more likely than nonusers to engage in numerous SRBs. Likelihood of SRB engagement generally does not differ among occasional, frequent, or daily EVP users, and youth who use EVPs only occasionally are more likely than nonusers to engage in 9 of the 10 SRBs examined. Health providers should be aware of the behavioral risks associated with EVPs, the most commonly used tobacco product among youth, to effectively screen and counsel adolescents about nicotine use and sexual health.

We thank Eli Rapoport for his review of the manuscript.

aPR

adjusted prevalence ratio

CI

confidence interval

EVP

electronic vapor product

LGBTQ+

lesbian, gay, bisexual, transgender, queer

SRB

sexual risk behavior

YRBS

Youth Risk Behavior Survey