Does Cannabis and Electronic Nicotine Delivery System (ENDS) Use Interfere with Cigarette Smoking Cessation in Young Adulthood?

Abstract

Aims:

Reducing combustible cigarette smoking is a major public health concern. Research identifies both electronic nicotine delivery systems (ENDS) and cannabis use as factors that are related to combustible cigarette cessation. The current study examined whether using substances in similar ways (e.g., inhaling them) triggers the urge to smoke, thereby undermining attempts to quit.

Design:

Analyses examined how using ENDS and smoking or vaping cannabis, both separately and together, were associated with reductions in daily combustible cigarette smoking among young adult smokers.

Participants:

Three waves of data from the Community Youth Development Study, including 909 participants who reported daily cigarette smoking at age 23, 26, and/or 28, were included in analyses.

Setting:

Participants were dispersed throughout 48 US states.

Measurements:

Latent growth modeling was used to model daily cigarette smoking over time. Models using past-month ENDS use, past-month smoking/vaping cannabis, and past-month co-use of ENDS and cannabis (using ENDS and smoking/vaping cannabis within the past month) as time-varying covariates were tested.

Findings:

Over time, there was a tendency towards cessation of daily combustible cigarette use among this smoking sample.

Smoking/vaping cannabis was associated with a decreased rate of daily combustible cigarette cessation among the sample, whereas ENDS use was associated with an increased rate of daily smoking cessation. The predicted additive effect of using ENDS and smoking/vaping cannabis was not significant.

Conclusions:

Among young adult daily cigarette smokers, smoking cannabis, on its own, poses a risk to quitting combustible cigarettes, while using ENDS may promote cigarette cessation, possibly through substitution.

1. Introduction

Smoking cigarettes is a leading cause of preventable disease and death.^1–3 Young adulthood (YA; e.g., age 18–30) represents an important developmental window for smoking cessation, as rates of combustible cigarette, electronic nicotine delivery systems (ENDS; also referred to as e-cigarettes), and cannabis use are high during adolescence and YA.^4–7 Quitting smoking earlier confers more health benefits than quitting later.^8–11 Despite population-level reductions in cigarette smoking,^{3; 5} rates of YA ENDS and cannabis use have been rising.^12–14 However, how ENDS and cannabis use impact cigarette smoking trajectories during YA is not well understood.¹⁵

1.1. Common route of administration theory

The common route of administration theory suggests that using substances in similar ways (e.g., by inhaling them) increases the risk of trying other substances that are delivered through similar methods, as well as risk of using multiple substances.¹⁶ Several factors can impact this, including the body habituating to one route (e.g., inhalation), making intitating other substance use through this route more normative. A similar process is hypothesized based on developing more favorable attitudes towards substances that are used in similar ways.¹⁷ Additionally, the similar route of administration may act as a shared cue via physiological, psychological, and behavioral responses that reinforces cravings for use of other substances administered in similar ways. Finally, such cues may make it more likely to use other substances, and make quitting more difficult.^{18; 19} Given that combustible cigarettes, ENDS, and cannabis are all used in similar ways (e.g., by inhaling them^{20; 21}) understanding how using ENDS (e.g., vaping e-cigarettes) and smoking/vaping cannabis, both separately and together, impacts cigarette smoking can offer important intervention targets.

1.2. ENDS use and combustible cigarette smoking

ENDS use has been steadily increasing among adolescents and young adults (YAs),^{7; 13; 14; 22} with bidirectional effects of cigarettes and ENDS initiation reported. Literature demonstrates elevated risk of cigarette smoking initiation for nonsmokers who began using ENDS, particularly among adolescents and YAs.^23–28 Increases in ENDS use among former and current cigarette smokers is also high, with one study showing that 25.2% of former cigarette smokers who quit smoking within the past year reported the highest current ENDS use.⁷ Among current cigarette smokers, 49.4% reported ever, and 9.7% reported currently using ENDS.

Many cigarette smokers adopt ENDS because they believe they are less dangerous than combustible cigarettes, or as an aid to quit smoking.^{29; 30} However, the evidence that using ENDS helps with quitting smoking is mixed. While some studies suggest that ENDS use is associated with higher rates of quitting^{31; 32} or maintaining reductions in smoking cigarettes,^{33; 34} others suggest that ENDS use may inhibit efforts to quit.^35–37 Additionally, some studies have found that using both ENDS and combustible cigarettes leads to increases in overall nicotine use,³⁸ potentially due to increased opportunities for nicotine use in places where combustible use is prohibited. Differences in findings across studies may be due to differences across study participant demographics (e.g., European adults, U.S. college students), differences in how ENDS, combustible cigarette, and cannabis use was assessed (e.g., daily smoker versus past 30 days), and differences in study design. In their review of studies examining the relationship between ENDS use and cigarette quitting among YAs, Glasser and colleagues¹⁵ concluded that none of the 26 studies reviewed contained the necessary components to specifically isolate the relationship between ENDS use and subsequent smoking due to omissions of pre-existing conditions and lack of inclusion of third variables.

1.3. Cannabis and combustible cigarette use

Research suggests that rates of cannabis use in the United States have also been increasing.^{12; 39} Cannabis use has been firmly linked to a higher likelihood of combustible cigarette use.^40–42 Additionally, increases in concurrent cannabis and combustible cigarette use^{43; 44} highlight the need to better understand how concurrent cannabis and cigarette use impacts trends in cigarette smoking. When examined in isolation, using cannabis is consistently associated with less successful quitting of combustible cigarettes.^{40; 45} Individuals who smoke both cigarettes and cannabis are more likely to report lower interest and success in quitting tobacco^{42; 44; 46} and are at greater risk of developing nicotine dependence and smoking more cigarettes.⁴¹

1.4. Current study

Studies have demonstrated relationships between cigarette, cannabis, and ENDS use such that increased use of any of these substances may increase risk of concurrent or subsequent cannabis, cigarette, and ENDS use among YAs.^47–50 The common route of administration may be one explanation for the overlap between the use of the three substances.¹⁶ The current study examines how using ENDS and smoking/vaping cannabis (hereafter referred to cannabis use), both separately and together, is associated with combustible cigarette smoking trajectories among YA (age 23 to 28) daily smokers. We hypothesized that combustible cigarette smoking would decline over time⁵ and that time-varying ENDS use, cannabis use, and their interaction would be associated with increases in combustible cigarette smoking (i.e., less cessation). Additionally, we hypothesized that using ENDS or cannabis would be associated with less cigarette cessation from age 23 to 28, as ENDS, cannabis, and combustible cigarettes are each inhaled. We expected that both using ENDS and smoking/vaping cannabis in the same month (from here on: co-use) would have a synergistic effect such that using ENDS or cannabis may be associated with less cigarette cessation over time, while co-use of ENDS and cannabis may be associated with even less likelihood of cigarette cessation across YA.¹⁶

2. Methods

2.1. Participants

The current study used three waves of data from the Community Youth Development Study (CYDS).⁵¹ CYDS was a community-randomized trial established in 2003 to assess the efficacy of the Communities That Care (CTC) prevention system. CTC aims to reduce adolescent substance use and other problem behaviors, and promote positive youth development by assisting community coalitions in implementing preventive interventions targeting locally specific risk and protective factors using data from youth in the community.⁵² CYDS followed 4407 youth from 24 small towns in seven states (Colorado, Illinois, Kansas, Maine, Oregon, Utah, and Washington State) beginning when they were in fifth grade. Participants were surveyed annually from age 10 to 16, and again at ages 18, 19, 21, 23, 26, and 28 (in 2021). Analyses for the current study were based on surveys conducted at ages 23, 26, and 28. Retention of the active living sample at ages 23, 26, and 28 was 88%, 87%, and 82%, respectively. By age 28, participants were dispersed throughout 48 US states. Details about CYDS are described elsewhere.^{51; 52} Although this study did not specifically examine intervention effects in the current study, intervention condition was included as a covariate in all analyses.

The current investigation was limited to participants who reported daily cigarette smoking within the past month at age 23, 26, or 28 (N = 909; 21% of the full sample). This smoking subsample was similar to the full sample in regards to race, ethnicity, and intervention status; was composed of 51.8% male, 48.2% female, and was racially and ethnically diverse (1.1% Asian, 3.9% Black, 26.7% Hispanic, 6.4% Multiracial, 6.4% Native American, 13.9% Other Race/Ethnicity, 0.5% Pacific Islander, 67.8% White). At age 28, 26.9% had a bachelor’s degree or equivalent. All procedures were approved by the University of Washington Human Subjects Review Committee.

2.2. Measures

2.2.1. Combustible cigarette use

Combustible cigarette use was reported at each wave. Participants were asked “In the past month, how frequently did you smoke cigarettes?” Response options were 1 = ‘Not at all’; 2 = ‘Less than 1 per day’; 3 = ‘1 to 5 per day’; 4 = ‘½ pack per day’; 5 = ‘1 pack per day’; 6 = ‘1½ packs per day’; 7 = ‘2+ packs per day.’ A dichotomous measure of smoking was computed that defined daily smokers as those who reported smoking at least one cigarette per day within the past month at either age 23, 26, or 28 (coded as 1, 0 otherwise).

2.2.2. Cannabis use

Cannabis use was assessed at all included time points. A measure of smoking/vaping cannabis was derived from two items. First, participants were asked to report how many occasions within the past month they used marijuana (with “use” not further defined as part of this item). Response options ranged from 1 = ‘0 occasions’ and 2 = ‘1–2 occasions’ to 7 = ‘40 or more occasions.’ At each time point, a dichotomous measure of using cannabis regularly was computed such that those who used cannabis on at least three occasions within the past month were coded as 1 (0 otherwise).

Participants were also asked: “In the past 30 days, how did you usually use marijuana?” Because it was hypothesized that inhaling cannabis may be a particular risk factor for cigarette use, we created a dichotomous variable where any endorsement of smoking or vaping cannabis in the past month was coded as 1 and any other form of cannabis use (i.e., “ate it,” “drank it,” “used it by dabbing,” “used it in some other way”) was coded as 0.

To create a “regular smoking or vaping marijuana in the past month” variable, we combined the “regular cannabis use” and the “smoking or vaping cannabis” variables (described above) so that those who used cannabis on at least three occasions in the past month, and reported smoking or vaping as their method of past-month cannabis use were coded as 1; nonregular use (< 3 occasions in the past month) or use that did not include smoking or vaping were coded as 0.

2.2.3. ENDS use

ENDS use was assessed at all time points in analyses with the item: “In the past 30 days how often have you used electronic cigarettes?” Response options mirrored those for cigarette smoking and ranged from 1 = ‘0 occasions’ to 7 = ‘40 or more occasions.’ We computed a dichotomous measure of ENDS use in which using ENDS on three or more occasions within the past month was coded as 1 to capture regular past-month use, and using ENDS less than three times within the past month was coded as 0, to capture nonregular past-month use.

2.2.4. Background variables

Background variables included biological sex, race, ethnicity, education, and experimental condition. Background covariates were from self-reports at the initial data collection when participants were in the fifth to seventh grade, except for level of education which used participants’ highest level of education reported by age 28. All demographic variables were binary: sex (male or female), race (White or other), ethnicity (Hispanic or not Hispanic), education (bachelor’s degree/equivalent or less than bachelor’s degree/equivalent), and intervention condition (CTC or control).

2.3. Analyses

Latent growth modeling using logistic regression was conducted in Mplus v8 ⁵³ to examine change in daily past-month cigarette use from age 23 to 28 among daily cigarette smokers. We included vaping ENDS (referred to as ENDS use), smoking or vaping cannabis (referred to as cannabis use), and their interaction, measured at each time point, as time-varying predictors of change in daily cigarette smoking. The effect of time-varying coefficients were assumed to be consistent across time and, therefore, for parsimony, were held equal across time points examined. Sensitivity analyses allowing coefficients to vary showed the same pattern of results. Background variables were included as time-invariant controls. The weighted least square mean and variance adjusted estimator was used in analyses and yielded probit estimates.

We examined model fit primarily using conventional fit criteria, including the comparative fit index (CFI),⁵⁴ the Tucker-Lewis Index (TLI),⁵⁵ and root mean square error of approximation (RMSEA),⁵⁶ Good fit was indicated by CFI and TLI values ≥ .90, and RMSEA values < .08.⁵⁷ Missing data were addressed with full information maximum likelihood.⁵³ The initial intercept and slope of the cigarette smoking trajectory for all participants was fitted first (Model 1), followed by models that included ENDS use alone (Model 2), smoking/vaping cannabis use alone (Model 3), both ENDS use and smoking/vaping cannabis (Model 4), and the full model with main effects and the interaction (Model 5). Because Models 1–3 did not include all variables, and Model 4 exhibited slightly stronger model fit than Model 5, we interpret findings primarily from Model 4. Notably, while the study design is clustered by community of origin, the intraclass correlations were negligible across ages 23 to 28 (.000 - .011), indicating that clustering was not necessary in analyses.⁵⁸

3. Results

Descriptive statistics demonstrated an overall reduction in cigarette smoking from 82% to 56% of this YA smoking sample from age 23 to age 28 (Table 1). Similarly, there was a reduction in cannabis use from 42% to 31% from age 23 to age 28. In contrast, ENDS use increased from 23% at age 23 to 31% at age 28. Lastly, there was a slight reduction in how many people used both ENDS and cannabis, decreasing from 9% to 8% from age 23 to 28.

Table 1.

Substance use prevalence for ages 23, 26, and 28.

	Age 23	Age 26	Age 28
Past-month daily cigarette smoker, n (%)	665 (81.5%)	552 (69.3%)*	402 (55.6%)*
Past-month regular ENDS users, n (%)	188 (23.0%)	270 (34.3%)*	225 (31.3%)
Past-month regularly smoke/vape cannabis, n (%)	348 (43.5%)	305 (39.2%)*	259 (36.3%)*
Past-month regular ENDS user and vape/smoke cannabis, n (%)	75 (9.2%)	117 (14.7%)*	91 (12.6%)

Note: Overall N = 909; all percentiles presented represent the valid percent, taking into account missing data. ENDS = electronic nicotine delivery systems.

^*represents significant change from prior age measured (p < .05)

3.1. Base latent growth model

All models demonstrated strong goodness of fit. Model 1 (without inclusion of time-varying covariates) demonstrated a significant decrease in daily combustible cigarette smoking among smokers from age 23 to 28 (see Table 2). No included background variable was consistently associated with cigarette smoking at age 23 or change in use from age 23 to age 28. In Model 4, identifying as White was associated with a steeper decline in cigarette smoking compared to non-White daily smokers.

Table 2.

Associations between ENDS use, smoking/vaping cannabis, and daily cigarette smoking from age 23 to 28.

	Model 1. Base model without time-varying covariates	Model 2. Model with ENDS	Model 3. Model with smoking/vaping cannabis	Model 4. Model with ENDS and smoking/vaping cannabis	Model 5. Model with ENDS, smoking/vaping cannabis, and their interaction
Estimate (standard error)
Slope mean	−0.624 (.261)*	−0.830 (.183)***	−0.469 (.292)	−0.675 (.240)**	−0.763 (.235)*
Slope
ENDS	--	−.083 (.057)	--	−.151 (.054)**	−.097 (.063)
Cannabis	--	--	.324 (.072)***	.255 (.062)***	.273 (.066)***
ENDS x cannabis interaction	--	--	--	--	−.117 (.091)
Slope
Male sex	−.006 (.112)	.007 (.089)	.000 (.135)	.039 (.114)	.061 (.112)
White race	−.288 (.149)	−.197 (.106)	−.323 (.165)	−.279 (.139)*	−.251 (.134)
Hispanic ethnicity	.243 (.165)	.250 (.131)	.251 (.199)	.318 (.172)	.377 (.173)*
Bachelor’s degree	−.133 (.263)	.108 (.149)	−.312 (.219)	−.171 (.172)	−.134 (.165)
Intervention status	−.034 (.112)	.002 (.089)	−.082 (.137)	−.024 (.116)	−.014 (.114)
Intercept
Male sex	.015 (.089)	−.005 (.085)	.007 (.091)	−.031 (.089)	−.052 (.089)
White race	.198 (.104)	−.167 (.116)	.187 (.106)	.178 (.103)	.155 (.102)
Hispanic ethnicity	−.196 (.125)	−.225 (.121)	−.170 (.130)	−.243 (.127)	−.304 (.131)*
Bachelor’s degree	.167 (.099)	−.222 (.103)*	−.130 (.112)	−.146 (.108)	−.164 (.107)
Intervention status	.027 (.088)	−.001 (.085)	.025 (.091)	−.015 (.089)	−.026 (.089)
Goodness-of-fit
CFI	1.00	.996	1.00	.995	.992
TLI	1.00	.981	1.00	.975	.963
RMSEA	0.00	.012	0.00	.016	.020

^*= p < .05;

^**= p < .01;

^***= p < .001

N = 909

Note: A negative coefficient indicates an increase in the downward trend of cigarette smoking. A positive coefficient indicates a reduction in the downward trend of cigarette smoking such that it offsets the downward trend. ENDS = electronic nicotine delivery systems; CFI = comparative fit index; TLI = Tucker-Lewis Index; RMSEA = root mean square error of approximation.

3.2. Latent growth model with time-varying substance use covariates

Model 2, which included ENDS use as a time-varying predictor, also demonstrated a significant downward slope in daily combustible cigarette smoking from age 23 to 28. However, ENDS use was not significantly associated with change in daily smoking in this model. When adjusting for cannabis use (Model 3), the downward trajectory in daily combustible cigarette smoking was no longer signficant. Cannabis use was associated with an increased probability of daily cigarette smoking and thus a significantly reduced rate of daily combustible cigarette smoking cessation over time. Sensitivity analyses comparing findings when “past-month regular smoking/vaping cannabis” was replaced by “past-month regular just smoking cannabis” and “past-month regular any cannabis use” demonstrated the same patterns of findings across models.

In Model 4, which included both ENDS and cannabis use main effects, the downward trajectory in daily combustible cigarette smoking was again significant. ENDS use was associated with a greater decline in cigarette use, while cannabis use was again associated with an increased probability of smoking cigarettes, and thus a reduced rate of cigarette cessation over time. Results from Model 5, which included both ENDS and cannabis use, as well as their interaction, continued to support a significant downward trajectory in daily cigarette smoking. However, when considering the synergistic effects of ENDS and cannabis co-use, the interaction between ENDS use and cannabis use was not significant. Thus, results demonstrate that the effect of ENDS use on change in daily cigarette smoking does not depend on cannabis use or vice versa. Instead, both ENDS use and cannabis use appear to have independent, and opposite, effects on daily smoking (Model 4; Table 2).

4. Discussion

The current study demonstrates that, similar to national trends,^{3; 5} daily cigarette smoking decreased from age 23 to 28 in this sample. Consistent with hypotheses, cannabis use was associated with less of a decline in daily cigarette smoking over time. In contrast to hypotheses, ENDS use was positively associated with a steeper decline in daily cigarette smoking over time for individuals who used ENDS, whereas cannabis use (regardless of route of administration) was associated with less cessation. Additionally, results demonstrated that ENDS and cannabis use had independent, but not interactive effects on their relationship with cigarette smoking across YA.

Findings support the growing literature documenting that cannabis may be a deterrent for quitting smoking.^{40; 44–46} However, findings failed to support the hypothesis that ENDS use would inhibit cigarette smoking cessation, either independently, or when used regularly during the same month as cannabis. Thus, among YA daily smokers, cannabis use on its own may pose a risk to combustible cigarette cessation, while using ENDS may promote cigarette cessation, possibly through substitution.

Our cannabis results are in line with research suggesting that smoking cigarettes and cannabis may increase risk of greater use of each substance, with recent research highlighting that up to 90% of cannabis users also smoke tobacco.⁵⁹ Results demonstrate similar findings regardless of if one vapes, smokes, or administers cannabis in any of its forms, with any form of cannabis use consistently associated with an attenuated decline in cigarette smoking over time. Thus, cigarettes and cannabis may share unique properties that make them more likely to be used together.¹⁷ Physiological or neurological pathways may explain these links, or differences in expectancies or reasons for use that might differ for cannabis and ENDS could explain the observed risks associated with cannabis use, but not ENDS, in our sample.

For ENDS, given the ongoing narratives and marketing that promotes ENDS as a healthier alternative to combustible cigarettes,^{36; 37; 60} further research is needed to understand how ENDS use might be related to cigarette use, especially in the context of cannabis use. Results suggest that using ENDS is related to more successful attempts to quit, reduce, or maintain reductions in smoking cigarettes (e.g., ^{32; 33; 34}). More nuanced investigation to better understand how, and for whom, ENDS might aid in efforts to quit smoking is needed. It could be that the impact of ENDS may differ for nondaily smokers versus daily smokers. Daily smokers might be more likely to use ENDS to quit smoking,^{29; 30} while using ENDS may increase risk for light smokers or nonsmokers to initiate or increase cigarette smoking.^25–27 Additionally, our results suggest that any benefits of using ENDS for smoking cessation may be offset by the effects of cannabis when a person also smokes/vapes cannabis. Future research is needed to further examine the potential moderating role of cannabis in the relationship between ENDS use and cigarette cessation.

4.1. Strengths and limitations

Strengths of this study include use of a longitudinal design and high rates of retention. Limitations include that participants recruited as children originated primarily from small towns in the Midwest and Western U.S. Therefore, results may not generalize to more urban communities or other regions. Additionally, this study was not designed to determine causal relationships and, while there are many factors that might relate to why someone uses multiple drugs, these analyses look at individual-level change and shed light on the importance of addressing their comorbidity.⁴⁰ Examination of levels of dependence on these different substances is also an important area for future research. We also examined a specific age range of YA daily smokers and results may not generalize to other ages or to nondaily smokers or nonsmokers. Further, age 28 data collection occurred in 2021 during the COVID-19 pandemic. It will be important to expand this longitudinal work post-pandemic to examine if patterns hold or differ for this data collection point. Data was also collected via self-report, and ENDS use was assessed broadly and did not allow modeling of the potential of different types of ENDS products used, potency, or the potential role of flavoring.⁶¹ Similarly, we were unable to assess potency of cannabis in this study. Notably, the ENDS market has evolved rapidly and ongoing research should continue to assess ENDS use and trends as ENDS products and marketing evolve.

4.2. Conclusions

This is the first study to our knowledge to examine the relationships between cannabis, ENDS, and combustible cigarette smoking across YA, despite the wide documentation of their co-use.^{21; 41} Given the high rates and patterns of co-use of combustible cigarette, ENDS, and cannabis during adolescence and YA,^4–7 better understanding how ENDS and cannabis use relate to cigarette smoking trajectories during YA is an important public health goal.

In line with other research, our results suggest that addressing cannabis use, regardless of method of use, may be an important target for intervention to reduce cigarette smoking.^{44; 62} Unfortunately, our understanding of how to best treat co-use of cannabis and cigarettes is lacking.⁵⁹ A growing literature highlights the potential positive impact of treating tobacco and cannabis use simultaneously to improve cigarette and cannabis cessation outcomes.^{17; 62} However, current programs often focus on one substance at a time while largely ignoring comorbid substance use. Additionally, although it is possible that using ENDS may promote cigarette cessation, possibly through substitution, it may be that smokers need to quit cannabis before they can successfully use ENDS as a quitting aid.⁶³ Future research should assess intervention strategies that are effective for reducing use of other substances—such as contingency management⁶⁴—for reducing co-use of ENDS, cannabis, and combustible cigarettes. Additionally, more longitudinal studies are needed to examine temporal ordering of initiation and cessation of co-use of different substances to determine best practices for preventing and treating co-use problems sequentially or simultaneously.

Highlights for Does Cannabis and Electronic Nicotine Delivery System (ENDS) Use Interfere With Cigarette Smoking Cessation in Young Adulthood?

• Cannabis use is a risk to quitting cigarette smoking among young adult smokers

• Vaping electronic cigarettes may promote cigarette cessation among smokers

• Cannabis and electronic cigarettes impact cigarette smoking independently

Data Availability Statement:

The data underlying this article will be shared on reasonable request to the second author.