The Link Between Depressive Symptoms and Vaping Nicotine in U.S. Adolescents, 2017–2019

Abstract

Purpose:

While there is a well-established association between depression and cigarette use, the mental health sequelae of vaping nicotine remain unclear. This study examined whether adolescents with depressive symptoms had higher odds of vaping nicotine than others, and how this association differed when examining vaping with cigarette use, vaping without cigarette use, and cigarette use alone.

Methods:

Using 2017–2019 Monitoring the Future data, we examined U.S. adolescents in the eighth, 10th and 12th grades surveyed in schools across the contiguous states. Depressive symptoms were measured by using questions around negative affect and hopelessness. The outcome included vaping with cigarette use; vaping without cigarette use; cigarette use alone; and neither. Control covariates included sex, race, highest level of parental education, and average grades.

Results:

The sample included 32,636 adolescents. Depressive symptoms were positively associated with comorbid vaping and cigarette use across all grades (eighth graders: adjusted odds ratio [aOR] = 3.52 [95% confidence interval (CI): 1.94–6.39]; 10th graders: aOR = 2.26 [95% CI: 1.51–3.38]; 12th graders: aOR = 1.81 [95% CI: 1.05–3.12]); vaping without cigarette use among eighth graders (eighth graders: aOR = 2.01 [95% CI: 1.46–2.77]; 10th graders: aOR = 1.20 [95% CI: .97–1.49]; 12th graders: aOR = 1.20 [95% CI: .84–1.70]); and cigarette use alone among eighth and 10th graders (eighth graders: aOR = 2.91 [95% CI: 1.50–5.62]; 10th graders: aOR = 2.29 [95% CI: 1.35–3.88]; 12th graders: aOR = 1.73 [95% CI: .83–3.61]).

Conclusions:

Eighth grade adolescents with depressive symptoms have increased odds of vaping nicotine with and without cigarette use. As vaping prevalence increases, clinician assessment of adolescent vaping should concomitantly acknowledge potential mental health correlates. Vaping may be a marker for a broader constellation of adolescent health concerns, including mental health.

In recent years, vaping among U.S. adolescents has increased dramatically. In 2017, one in ten 12th grade adolescents reported past-month nicotine vaping; in 2019, this increased to one in four adolescents [1]. While vaping may engender fewer health harms than smoking combustible cigarettes, only 6% of adolescents who vape do so to reduce smoking [2]. Most adolescent vaping is among teens with no prior smoking experience [3], who are at subsequent greater risk of initiating cigarette use than those who do not vape [4]. Similar to cigarette use, vaping is associated with nicotine dependence [5], other substance use [6], health risk behaviors [7], and impaired respiratory function [8]. As a result, there is growing concern around the risk factors for and consequences of vaping in youth.

Among adolescents, depression is a known risk factor for cigarette use, including smoking initiation [9], number of cigarettes smoked [10], and nicotine dependence [11]. Cigarette use has also been identified as a risk factor for future depression in adolescents [12], and comorbid smoking and depression have been associated with adverse clinical and social outcomes, including lower rates of smoking cessation [13] and lower educational attainment after high school [14].

Several pathways may explain the relationship between smoking combustible cigarettes and depression, including the use of cigarettes to cope with psychosocial stress [15,16], self-medication via nicotine’s antidepressant effects [17,18], depression resulting from nicotine withdrawal [17,19], or altered serotonin function resulting from nicotine exposure during adolescent brain development [20]. If these are the pathways through which cigarette use and depression are related, then we may expect similar magnitudes of association when examining vaping nicotine and depression. However, the social context of vaping and cigarette use may be very differentd—vaping is a relatively new product, and might attract different users with fewer underlying risks for depression. Moreover, as vaping becomes increasingly commonplace, it may be initiated for more social, rather than coping or self-medication, purposes. Therefore, despite the existence of clear links between cigarette use and depression in adolescents, whether this association extends to vaping is unclear.

To our knowledge, only three studies to date have examined the link between vaping nicotine and depression. Two of these studies examined U.S. adults and found depression to be positively associated with initiation and current use of e-cigarettes [21,22]. A longitudinal study of 2,640 adolescents similarly found that e-cigarette use and depression were associated, with a bidirectional temporal relationship [23]. However, data were drawn from ninth graders in a single area of California in 2013, limiting generalizability. The association between adolescent vaping and depressive symptoms has not been investigated in nationally representative samples in recent years when adolescent vaping has become widespread and new vaping technology has increased the concentration of nicotine delivered [24]. There is also a dearth of evidence regarding the magnitude of this association relative to other substances, such as alcohol and marijuana, which have robust associations with adolescent depression [25].

We therefore examined the association between depressive symptoms and smoking and/or vaping in a nationally representative sample of adolescents from 2017 to 2019. Specifically, among eighth, 10th, and 12th grade students, we assessed whether those with depressive symptoms were more likely to vape nicotine, and whether this association differed when examining vaping among students who also smoked cigarettes, those who did not smoke cigarettes, and those who smoked cigarettes but did not vape. We then further compared the association of depression with vaping nicotine to those with binge drinking alcohol and using cannabis.

Methods

Setting, participants, and procedures

Deidentified data were drawn from the 2017, 2018, and 2019 Monitoring the Future Survey (MTF). MTF is an annual, nationally representative, cross-sectional survey of U.S. eighth, 10th, and 12th grade students [26]. The stages of sample selection include (1) selection of geographic areas; (2) high schools within geographic areas; and (3) students within high schools. Each school’s probability of selection is proportionate to its class size, and sample weights are calculated to compensate for different selection probabilities [26]. Any school that declines to participate is replaced with another randomly sampled school within the target geographic location and is of similar size and urbanicity as the originally selected school. For a detailed description of MTF, including the complex multistage sampling design, see the study by Bachman et al. [27].

Approximately 400 public and private schools in the 48 contiguous states are sampled each year [26]. After completing the consent process (active or passive, per school policy), participants in the 12th grade are randomly assigned to complete one of six self-administered surveys, while participants in the eighth and 10th grades are randomly assigned to complete one of four self-administered surveys. Each survey consists of “core” questions that all students are asked to complete and other questions that appear on only some of the other surveys. Question topics include demographic characteristics, substance use, mental health, marital and family plans, and more. Further details regarding the MTF design and procedures are available elsewhere [26]. Past-month cigarette use was asked about on the core questionnaire, while questions around past-month vaping nicotine and depressive symptoms were randomly assigned to different forms. A school was successfully recruited to participate in 90% of the target geographic regions selected in 2017, 2018, and 2019. Across the three grades sampled, student response rates ranged from 79% to 87% in 2017 and 81% to 89% in 2018 and 2019 [26]. Nearly all nonresponse was due to student absences. Prior analyses show that adjustment for absenteeism results in minimal effects on population estimates [26].

Measures

Outcomes.

Past-month vaping/smoking status was measured using answers to the following questions: “On how many days (if any) have you vaped nicotine during the last 30 days?” and “How frequently have you smoked cigarettes during the past 30 days?” For 12th graders in 2017 and 2018, the question regarding vaping read, “During the last 30 days, on how many occasions (if any) have you vaped an e-liquid with nicotine?” Those who indicated at least one occasion of vaping nicotine and no days of smoking cigarettes were coded as “vaping without cigarette use”. Those who indicated at least one occasion of vaping nicotine and at least one day of smoking cigarettes were coded as “vaping with cigarette use”. Those who indicated no occasions of vaping and at least one day of smoking cigarettes were coded as “cigarette use only”. Those who reported no cigarette smoking or vaping in the past 30 days were coded as “neither”. Therefore, the final outcome variable was in nominal categorical form with four levels: vaping with cigarette use, vaping without cigarette use, cigarette use only, and neither vaping nor cigarette use. “Neither” was used as the reference group for all analyses and referred to those adolescents who replied “None” and “Not at all” to questions regarding vaping nicotine and smoking cigarettes in the past 30 days.

Main exposures.

The presence of depressive symptoms was tapped with four questions, all using the same stem, “How much do you agree or disagree with each of the following statements?”. The statements were “Life often seems meaningless”, “The future often seems hopeless”, “It feels good to be alive”, and “I enjoy life as much as anyone”. Responses ranged from 1 (Strongly Disagree) to 5 (Strongly Agree). Positive statements were reverse coded. Item responses were averaged and then dichotomized using a mean score greater than 3 (indicating at least some agreement across statements) to indicate the presence of depressive symptoms. These items of depressive symptoms have been used numerous times in prior research [28–31], are similar to those on the Center for Epidemiologic Studies Depression Scale [32], and demonstrate excellent internal consistency across the three grades [28].

Control covariates.

Control covariates were demographic characteristics hypothesized to be independently associated with both depressive symptoms and smoking/vaping, including sex (male, female), race (non-Hispanic white, racial or ethnic minority), highest level of parental education (<high school, high school, at least some college), and average grades in school (≥B+, <B+ [the median grade of the combined sample]).

Statistical analysis.

To account for the multistage random sampling design, all analyses were weighted using published MTF sample weights [3]. Prevalences of all outcomes and demographic characteristics were calculated for each grade. Missing data were handled using multiple imputation with chained equations with 20 imputed data sets. All observations with missing outcome values for past-month smoking/vaping were excluded. In total, 60% of the sample had complete data on all variables, 31% had one imputed value, and 9% had ≥2 imputed values. Unadjusted and covariate-adjusted multinomial logistic regression models were used to assess the relationship between depressive symptoms and past-month vaping/smoking. A single model assessed the association between depressive symptoms and smoking behaviors in all grades. An interaction variable (depressive symptoms x grade) tested whether the association between depressive symptoms and smoking varied by grade. After confirming the significance of the interaction term (p < .01), this model was used to calculate separate estimates for eighth, 10th, and 12th graders. Unadjusted and adjusted odds ratios (ORs and aORs) reflected the odds of vaping nicotine but not smoking cigarettes, vaping nicotine and smoking cigarettes, or smoking cigarettes only for adolescents with versus without depressive symptoms. For comparisons with other substances, the adjusted model was rerun using past 2-week binge drinking and past-month cannabis use, respectively. Data management and preparation were conducted using SAS version 9.4 (Cary, NC). All imputations and analyses were conducted using the R statistical packages multiple imputation with chained equations and nnet. Because the data are publicly available and deidentified, the present study was exempt from review by the Columbia University Medical Center Institutional Review Board.

Sensitivity analyses.

Three sensitivity analyses tested the robustness of results. First, as other substance use, including marijuana and alcohol use, may lead to both vaping/smoking and depressive symptoms, we additionally controlled for past-month marijuana use (yes, no) and binge drinking in prior 2 weeks (yes, no). These covariates were not included in the main analyses because of uncertainty around the directionality of the relationship between other substance use and our main exposure and outcome. A second sensitivity analysis increased the threshold for mean depressive symptoms from >3.0 to >3.5 to reflect greater severity. Third, as the race variable in public use MTF datafiles is limited to preserve participant confidentiality (including only three categories of white, black, and Hispanic), our primary analysis combined black and Hispanic categories to create a dichotomous race variable (non-Hispanic white vs. racial or ethnic minority), which was imputed for participants missing information on race in the datafile. A final sensitivity analysis dropped race as a control covariate from the analysis.

Results

The final study sample consisted of 32,636 U.S. adolescents from 2017 to 2019, including 13,422 eighth grade students, 13,378 10th grade students, and 5,816 12th grade students. Sample characteristics of participants by grade are presented in Table 1.

Table 1.

Sample characteristics of eighth, 10th, and 12th grade adolescents, 2017–2019^a

Characteristics	Grade			Total
	Eighth	10th	12th
Sex
Male	6,459 (49.4)	6,599 (49.0)	2,672 (47.8)	15,730 (49.0)
Female	6,582 (50.6)	6,520 (51.0)	2,964 (52.2)	16,066 (51.0)
Race
Non-Hispanic white	5,911 (57.2)	6,698 (59.7)	3,090 (62.1)	15,699 (59.1)
Racial or ethnic minority	4,295 (42.8)	4,080 (40.3)	1,704 (37.9)	10,079 (40.9)
Parental education
<High school	1,515 (12.8)	1,274 (11.3)	596 (11.7)	3,385 (11.9)
High school	1,944 (16.3)	1,782 (15.2)	993 (18.6)	4,719 (16.2)
Some college	1,592 (13.7)	1,953 (16.3)	1,038 (19.3)	4,583 (15.7)
College diploma	6,847 (57.3)	7,515 (57.3)	2,957 (50.3)	17,319 (56.1)
Average grades
<B+	5,662 (42.8)	6,418 (49.0)	2,353 (40.9)	14,433 (45.1)
≥B+	7,780 (57.3)	6,960 (51.0)	3,463 (59.1)	18,203 (54.9)
Marijuana use in past 30 daysb	767 (5.6)	2,245 (16.9)	1,347 (22.7)	4,359 (13.2)
Binge drinking in past 2 weeksc	488 (3.9)	1,267 (9.4)	897 (15.0)	2,652 (8.0)
Depressive symptoms, mean (SE)d
“Life often seems meaningless”e	2.39 (.01)	2.47 (.01)	2.20 (.02)	2.39 (.01)
“The future often seems hopeless”f	2.15 (.01)	2.27 (.01)	2.20 (.02)	2.21 (.01)
“It feels good to be alive”g	4.09 (.01)	4.09 (.01)	4.09 (.02)	4.08 (.01)
“I enjoy life as much as anyone”h	3.78 (.01)	3.72 (.01)	3.72 (.02)	3.74 (.01)
Smoking status (past month)
Vaped nicotine and did not use cigarettes	713 (5.0)	1,808 (12.6)	1,073 (17.0)	3,594 (10.1)
Vaped nicotine and used cigarettes	145 (1.1)	321 (2.5)	281 (4.8)	747 (2.3)
Smoked cigarettes only	124 (.9)	174 (1.5)	124 (2.4)	422 (1.4)
Neither	12,460 (93.1)	11,075 (83.5)	4,338 (75.8)	27,873 (86.2)
No. of participants (N)	13,422	13,378	5,816	32,636

Values are given as freq (weighted %).

SE = standard error.

^aBased on Monitoring the Future eighth, 10th, and 12th grade public use datafiles, 2017–2019. Questions on depressive symptoms and nicotine vaping were administered to a randomly selected one-third of both eighth and 10th grade students and one-sixth of 12th grade students.

^bExcludes 1.7% of students for whom data on past-month marijuana use were not collected.

^cExcludes 4.5% of students for whom data on past 2-week binge drinking were not collected.

^dStem questions with scores ranging from 1 = strongly disagree to 5 = strongly agree. 88.5% of participants had 0–1 missing responses. 11.5% had ≥2 missing responses.

^eExcludes 10.6% of students for whom data on this question were not collected.

^fExcludes 11.8% of students for whom data on this question were not collected.

^gExcludes 12.4% of students for whom data on this question were not collected.

^hExcludes 11.2% students for whom data on this question were not collected.

Prevalences of vaping with and without cigarette use

As grade increased, the prevalences of vaping with cigarette use, vaping without cigarette use, and cigarette use alone increased (Table 1). Among all adolescents who vaped nicotine in the past month, 18.4% also smoked cigarettes. Among eighth graders, 10th graders, and 12th graders who vaped nicotine, 17.8%, 16.6%, and 21.2% also smoked cigarettes, respectively.

Prevalences of vaping/smoking by depressive symptom status

Among adolescents with nonimputed information on depressive symptoms (N = 28,546), 9.7% of those without depressive symptoms vaped nicotine but did not smoke cigarettes, 1.8% vaped nicotine in addition to smoking cigarettes, and 1.1% smoked cigarettes only. In contrast, among adolescents with depressive symptoms, 13.1% vaped nicotine but did not smoke cigarettes, 4.1% vaped nicotine and smoked cigarettes, and 2.4% smoked cigarettes only. Prevalences of vaping and/or cigarette use by depressive symptom status and grade are presented in Table 2.

Table 2.

Prevalence of vaping/smoking among adolescents with and without depressive symptoms (N = 28,546)^a

Prevalence (95% CI)	Past-month smoking
	Vaping without cigarette use	Vaping with cigarette use	Cigarette use only
Eighth Graders
No depressive symptoms	4.1 (3.7–4.5)	.7 (.5–.9)	.6 (.5–.8)
Depressive symptoms	8.7 (7.5–10.0)	2.5 (1.8–3.2)	1.7 (1.2–2.3)
Factor increase in prevalence	2.12	3.57	2.83
10th Graders
No depressive symptoms	12.1 (11.4–12.7)	1.8 (1.6–2.1)	1.1 (.9–1.3)
Depressive symptoms	14.9 (13.4–16.3)	4.5 (3.6–5.3)	2.5 (1.9–3.2)
Factor increase in prevalence	1.23	2.50	2.27
12th Graders
No depressive symptoms	16.5 (15.5–17.6)	4.1 (3.6–4.7)	2.2 (1.8–2.6)
Depressive symptoms	19.2 (16.5–21.9)	7.3 (5.5–9.1)	3.8 (2.5–5.2)
Factor increase in prevalence	1.16	1.78	1.73

CI = confidence interval.

^aSample includes all observations with nonimputed exposure information on depressive symptoms.

Are depressive symptoms associated with vaping nicotine in the absence of cigarette use?

Depressive symptoms were associated with vaping nicotine without cigarette use primarily among eighth grade adolescents. In adjusted analyses, compared with adolescents in the same grade without depressive symptoms, eighth graders with depressive symptoms had 2.01 times the odds of vaping without cigarette use (95% confidence interval [CI]: 1.46, 2.77); 10th graders with depressive symptoms had 1.20 times the odds of vaping without cigarette use (95% CI: .97, 1.49); and 12th graders with depressive symptoms had 1.20 times the odds of vaping without cigarette use (95% CI: .84, 1.70) (Table 3).

Table 3.

Odds ratios for the association between depressive symptoms and vaping/smoking

	Past-month smokinga
	Vaping without cigarette use	Vaping with cigarette use	Cigarette use only
Eighth Graders
Unadjusted OR (95% CI)	2.08 (1.52, 2.84)c	3.74 (2.07, 6.73)c	3.23 (1.68, 6.22)c
Adjusted ORb (95% CI)	2.01 (1.46, 2.77)c	3.52 (1.94, 6.39)c	2.91 (1.50, 5.62)d
10th Graders
Unadjusted OR (95% CI)	1.32 (1.07, 1.62)e	2.60 (1.76, 3.86)c	2.61 (1.55, 4.40)c
Adjusted ORb (95% CI)	1.20 (.97, 1.49)f	2.26 (1.51, 3.38)c	2.29 (1.35, 3.88)d
12th Graders
Unadjusted OR (95% CI)	1.27 (.90, 1.79)f	2.03 (1.19, 3.44)d	1.88 (.91, 3.90)f
Adjusted ORb (95% CI)	1.20 (.84, 1.70)f	1.81 (1.05, 3.12)e	1.73 (.83, 3.61)f

CI = confidence interval; HS = high school; OR = odds ratio.

^aReference = Neither vaping nicotine nor smoking cigarettes in the past month.

^bAdjusted for sex (male/female), race (non-Hispanic white/racial or ethnic minority), highest parental education (<HS/HS/at least some college), average grades in school (≥B+/<B+).

^cp < .001.

^dp < .01.

^ep < .05.

^fp > .05.

Are depressive symptoms associated with vaping nicotine in combination with cigarette use?

Depressive symptoms were associated with vaping nicotine in combination with cigarette use among eighth, 10th, and 12th grade adolescents. In adjusted analyses, compared with adolescents in the same grade without depressive symptoms, eighth graders with depressive symptoms had 3.52 times the odds of vaping in combination with cigarette use (95% CI: 1.94, 6.39); 10th graders with depressive symptoms had 2.26 times the odds of vaping in combination with cigarette use (95% CI: 1.51, 3.38); and 12th graders with depressive symptoms had 1.81 times the odds of vaping in combination with cigarette use (95% CI: 1.05, 3.12) (Table 3).

Are depressive symptoms associated cigarette use alone?

Depressive symptoms were associated with cigarette use alone among eighth and 10th grade adolescents. In adjusted analyses, compared with adolescents in the same grade without depressive symptoms, eighth graders with depressive symptoms had 2.91 times the odds of cigarette use alone (95% CI: 1.50, 5.62); 10th graders with depressive symptoms had 2.29 times the odds of cigarette use alone (95% CI: 1.35, 3.88); and 12th graders with depressive symptoms had 1.73 times the odds of cigarette use alone (95% CI: .83, 3.61) (Table 3).

How does the association of depressive symptoms with vaping nicotine compare with other substances?

Among eighth graders, depressive symptoms had a higher magnitude of association with vaping nicotine (aOR = 2.01, 95% CI: 1.46–2.77) than with either binge drinking (aOR = 1.67, 95% CI: 1.16–2.40) or marijuana use (aOR = 1.86, 95% CI: 1.37–2.54) (Supplemental Table 1). Among 10th and 12th graders, the magnitude of association of depressive symptoms with vaping (10th grade aOR = 1.20, 95% CI: .97–1.49; 12th grade aOR = 1.20, 95% CI: .84–1.70) was similar to that with binge alcohol use (10th grade aOR = 1.32, 95% CI: 1.02–1.72; 12th grade aOR = 1.20, 95% CI: .81–1.78) and lower than that with marijuana use (10th grade aOR = 1.65, 95% CI: 1.34–2.04; 12th grade aOR 1.37, 95% CI: .97, 1.93) (Supplemental Table 1).

Sensitivity analyses

Full sensitivity analyses are presented in Table 4. Increasing the threshold for depressive symptoms to reflect greater severity (mean score >3.5) and removing race from the control covariates did not meaningfully change the observed results. When marijuana use and binge drinking were additionally controlled, the association between depressive symptoms and vaping with cigarette use was notably reduced among 12th graders (aOR = 1.41, 95% CI: .76–2.62), suggesting that it may not represent a robust relationship.

Table 4.

Sensitivity analyses for the association between depressive symptoms and vaping/smoking

Adjusted OR (95% CI)	Past-month smokinga
	Vaping without cigarette use	Vaping with cigarette use	Cigarette use only
Raising the depressive symptom thresholdb
Eighth Graders	2.02 (1.40, 2.92)f	3.33 (1.71, 6.47)f	3.63 (1.81, 7.27)f
10th Graders	1.22 (.94, 1.60)i	2.04 (1.26, 3.32)g	3.01 (1.68, 5.41)f
12th Graders	1.13 (.71, 1.79)i	2.55 (1.38, 4.70)g	1.30 (.46, 3.64)i
Controlling for other substance usec
Eighth Graders	1.66 (1.19, 2.32)g	2.41 (1.26, 4.62)g	2.33 (1.18, 4.58)h
10th Graders	.95 (.74, 1.20)i	1.58 (1.01, 2.48)h	1.76 (1.03, 3.02)h
12th Graders	1.00 (.67, 1.49)i	1.41 (.76, 2.62)i	1.40 (.65, 3.01)i
Past-month marijuana used	6.71 (5.70, 7.90)f	13.40 (9.74, 18.4)f	7.56 (5.23, 10.92)f
Past 2-week binge drinkingd	4.74 (3.92, 5.74)f	12.67 (9.23, 17.39)f	5.53 (3.73, 8.20)f
Dropping race from the control covariatese
Eighth Graders	2.00 (1.46, 2.74)f	3.53 (1.95, 6.38)f	2.90 (1.50, 5.61)g
10th Graders	1.26 (1.02, 1.55)h	2.42 (1.63, 3.61)f	2.36 (1.39, 4.00)g
12th Graders	1.23 (.86, 1.74)i	1.91 (1.12, 3.26)h	1.76 (.84, 3.66)i

CI = confidence interval; HS = high school; OR = odds ratio.

^aReference = Neither vaping nicotine nor smoking cigarettes in the past month.

^bAdjusted for sex (male/female), race (non-Hispanic white/racial or ethnic minority), parental education (<HS/HS/at least some college), and average grades in school (≥B+/<B+).

^cAdjusted for sex (male/female), race (non-Hispanic white/racial or ethnic minority), highest parental education (<HS/HS/at least some college), average grades in school (≥B+/<B+), past-month marijuana use and past 2-week binge drinking.

^dOverall estimates across grades 8, 10, and 12 since prior modeling demonstrated a lack of interaction between other substance use and grade on smoking status.

^eAdjusted for sex (male/female), parental education (<HS/HS/at least some college), and average grades in school (≥B+/<B+).

^fp < .001.

^gp < .01.

^hp < .05.

ⁱp > .05.

Discussion

The present study examined the relationship between depressive symptoms and past-month nicotine vaping among eighth, 10th, and 12th grade adolescents in the U.S. from 2017 to 2019. Younger adolescents with depressive symptoms had increased odds of vaping nicotine both with and without cigarette use. To our knowledge, this is the first study to investigate adolescent mental health and vaping in a nationally representative sample, across different age groups, or in comparison with other commonly used substances, including alcohol and marijuana.

We outline three central findings. First, nearly one in five adolescents who vaped nicotine also smoked cigarettes in the past month. Using U.S. Census Bureau population estimates for adolescents aged 12–19 years [33,34], extrapolation of our study results suggests that approximately 4.22 million U.S. adolescents vape nicotine, of whom nearly 800,000 additionally smoked cigarettes. Second, depressive symptoms were associated with 2–3 times the odds of vaping in combination with cigarette use among eighth and 10th grade adolescents. Third, depressive symptoms were positively associated with vaping nicotine among eighth grade adolescents, for whom depressive symptoms approximately doubled the odds of vaping nicotine in the past month. In eighth grade adolescents, depressive symptoms have a greater association with vaping nicotine than with either binge drinking or using marijuana.

Overall, these findings suggest that vaping is more common among younger adolescents (~age 12–14 years) who have elevated depressive symptoms. Clinicians should query adolescents regarding vaping nicotine, and among those who report use, consider assessing a broader constellation of health issues, including mental health. While research on counselling for adolescent vaping is lacking [35], there is evidence that addressing adolescent depression can aid in reducing subsequent substance use. For example, one study of youths with major depressive disorder found that those who responded to depression treatment additionally showed a significant improvement in substance-related impairment [36]. Another study of adolescents with depressive symptoms found that a depression-prevention program significantly reduced substance use at 2-year follow-up [37]. Brief interventions for depression may similarly improve vaping nicotine and have been shown to be feasibly implemented in primary care settings [38] alongside smoking-cessation treatments [39].

The present study also highlights the need for research on the potential moderating role of depression in interventions for vaping in young adolescents. Among adolescent cigarette users, higher depression scores have been associated with lower rates of abstinence after treatment [40]. The same may be true for depressed adolescents who vape nicotine, requiring targeted vaping-cessation programs that incorporate therapies for depression. For example, a meta-analysis of 16 randomized controlled trials found that the addition of behavioral mood management to smoking-cessation interventions improved abstinence in depressed adults, while the addition of antidepressants did not [39]. The relative efficacy of different combinations and types of interventions for comorbid adolescent nicotine vaping and depression is an important area for future study.

Study limitations are noted. First, the present study is cross-sectional and therefore cannot determine the directionality of the relationship between nicotine use and depression. Longitudinal research should examine hypotheses generated by this work, including depression as a risk factor for vaping initiation in younger adolescents, as well as for progression from vaping nicotine to cigarette use (or vice-versa) across adolescent age groups. Second, we examined vaping and cigarette use ≥1 time in the past month and lacked information about the quantity of nicotine ingested. As adolescents with depression are at higher risk of nicotine addiction [11], future research should examine the association of depressive symptoms with daily or near-daily smoking/vaping or with consuming greater quantities of nicotine. Third, the measure of depressive symptoms was not based on a clinical assessment, and information on mental health service utilization was not available. Future research should aim to address questions around vaping and clinical depression using validated clinical measures. Still, the 4-item measure implemented in the present study has been used numerous times in prior research and demonstrates excellent internal consistency across the three grades sampled. Fourth, owing to confidentiality measures implemented in MTF public use data files, the measure used for race in the present study was limited, including only three mutually exclusive categories: white, black, and Hispanic. To address this issue, we conducted a sensitivity analysis where race was dropped from the analysis, which caused no change to the results. Future research should include more refined race/ethnic categories. Fifth, Monitoring The Future does not ask questions around early life trauma or parental demographic or clinical characteristics other than level of education that might predict both depressive symptoms and smoking behaviors. Future research should aim to use data sets which allow incorporating these factors. Finally, Monitoring the Future excludes adolescents who are not in school. As school dropout rates increase with age, the 12th grade adolescents in this study may be less representative of the U.S. adolescent population than those sampled in the eighth and 10th grades. Still, the effect on 17- to 19-year-olds should only be modest, as 92% of 12th grade youth in 2018 had not dropped out of high school [41], and, in addition, the results remain generalizable to all 12th graders in school.

Conclusion

As adolescent vaping increases across the U.S. and evidence aiding clinicians in addressing these behaviors is lacking, understanding the correlates and potential risk factors for vaping nicotine is increasingly urgent. Overall, the present study suggests that a substantial proportion of adolescents who vape nicotine also smoke cigarettes and that both single and dual use of these products are associated with depressive symptoms among younger adolescents. In clinical settings, questions around vaping behaviors should be accompanied by inquiries around smoking cigarettes and potential depression. Adolescents who vape or smoke should be screened for depressive symptoms, with a particular emphasis on younger adolescents (~12–14 years old), who may be at the highest risk of comorbid nicotine use and depression.

Supplementary Material

Supplementary data related to this article can be found at https://doi.org/10.1016/j.jadohealth.2021.07.003.

IMPLICATIONS AND CONTRIBUTION.

Nationally representative data show that vaping is more common among younger adolescents (~aged 12–14 years) with depressive symptoms than those without. In clinical settings, adolescents who vape and smoke should be screened for depressive symptoms. Younger adolescents may be particularly vulnerable.