Effects of Very Low Nicotine Content Cigarettes and Nicotine Vaping Device Characteristics on Choices to Smoke, Vape, or Abstain in Early Young Adults

Abstract

Introduction

A national nicotine reduction policy could reduce the public health toll of smoking. However, reducing nicotine in cigarettes may lead to changes in the use of other tobacco products such as nicotine vaping devices, particularly among young people. Product use outcomes may depend on the characteristics of available nicotine vaping devices. We aimed to determine the impact of cigarette nicotine content, vaping device nicotine concentration, and vaping device flavors on choices to smoke, vape, or abstain.

Aims and Methods

Early young adults (ages 18–20 inclusive, N = 80) who reported smoking daily and vaping nicotine at least twice in their lifetime participated in a laboratory study. Participants received either very low nicotine content (VLNC; 0.4 mg nicotine/g of tobacco) or normal nicotine content (NNC; 15.8 mg/g) cigarettes. First, participants chose between their assigned cigarette or abstaining. Subsequently, participants chose between 2 cigarette puffs, 2 vape puffs, or abstaining. Vaping device nicotine concentration (3 mg vs. 18 mg/mL) and flavor (tobacco vs. non-tobacco) were manipulated within subjects.

Results

When only cigarettes were available, there were no differences between the VLNC and NNC groups on cigarette choices. When the nicotine vaping device was concurrently available, the VLNC group made fewer choices to smoke than the NNC group. Non-tobacco flavors and lower vaping device nicotine concentration were associated with fewer choices to smoke.

Conclusions

Nicotine vaping device availability reduced choices to smoke VLNC cigarettes, and vaping devices with lower nicotine and non-tobacco flavors led to the fewest choices to smoke. Regulators should consider that the availability and characteristics of alternative tobacco products can moderate the product standard’s impact.

Implications

The U.S. Food and Drug Administration may enact a reduced nicotine product standard that would affect all commercially available cigarettes. One important population affected by this policy would be early young adults who smoke. We aimed to determine the impact of cigarette nicotine content, vaping device nicotine concentration, and vaping device flavors on choices to smoke, vape, or abstain. Lower nicotine in cigarettes, along with non-tobacco flavors and lower nicotine concentration in the vaping device, were associated with the fewest choices to smoke. Regulators should consider that the availability and characteristics of alternative tobacco products can moderate the product standard’s impact.

Introduction

While combustible cigarette smoking has declined considerably among U.S. adults and youth, leading many to believe the problem of smoking has been “solved,” smoking remains the leading cause of death, disability, and disease in the United States.^1,2 While vaping is a more popular form of nicotine intake in young people,³ it is less dangerous than smoking,⁴ which remains a primary preventable cause of death.⁵ The burden of smoking-related disease is inequitably distributed; smoking is disproportionately concentrated in at-risk subgroups such as those with mental health concerns and sexual and gender minorities,^6,7 and these disparities begin early.⁸ Thus, arresting the trajectory of smoking early, despite reductions in cigarette smoking in young people, is a critical public health priority in order to reduce death and disease from tobacco use.

A national nicotine reduction policy, which would mandate a reduction in nicotine to very low levels in all commercially available cigarettes, has been proposed by the U.S. Food and Drug Administration.⁹ Data from over a decade’s worth of laboratory studies and clinical trials with adults, including individuals from priority populations (eg, those with mental health conditions), have demonstrated that when people who smoke are randomized to very low nicotine content (VLNC) cigarettes, they reduce their smoking relative to those who are randomized to normal nicotine content (NNC) control cigarettes.^10–13 Young people are also an important population to consider when determining the implementation of a tobacco product standard. Research has consistently shown that such a policy would lead to a lower addiction potential of cigarettes and reduced smoking.^14–17

However, the tobacco product landscape includes more than just cigarettes, and reducing nicotine in cigarettes may lead to changes in the use of other tobacco products, such as e-cigarettes and related products (herein referred to as ‘nicotine vaping devices’).¹⁸ This may be especially true for adolescents and early young adults (referred to here collectively as ‘youth’, as studies referenced focus on participants ranging in age from adolescence to ~25 years of age), who use these products at high rates.¹⁹ Furthermore, the extent to which a cigarette nicotine reduction standard reduces cigarette smoking may depend on the characteristics (ie, nicotine concentration, flavor) of the available nicotine vaping devices. If the appeal of and reinforcement from combusted tobacco products with reduced nicotine content is low, smokers will likely seek alternative sources of nicotine that may partially substitute for cigarettes.²⁰ If reducing cigarette nicotine content reduces the reinforcing effects of cigarettes relative to nicotine vaping devices, youth may partially or completely switch from cigarettes to nicotine vaping devices. Given that both cigarettes and nicotine vaping devices are potential targets of FDA regulation, it is important to understand how specific nicotine vaping device characteristics influence their reinforcement value relative to that of NNC and VLNC cigarettes in youth.

The aims of this project are twofold. First, this project compares the impact of cigarette nicotine content on combusted cigarette smoking when only cigarettes are available versus when an alternative product (nicotine vaping device) is concurrently available. Second, we aimed to examine the extent to which specific nicotine vaping device characteristics (nicotine concentration and flavors) influenced the relative reinforcement value of nicotine vaping devices compared to cigarettes, when NNC cigarettes are available (mirroring the current environment) versus when VLNC cigarettes are available (mirroring a future reduced nicotine standard environment). We hypothesized that VLNC cigarettes would reduce cigarette choices to a greater extent than NNC cigarettes and that non-tobacco flavors and higher nicotine concentration in vaping devices would be the most effective at reducing cigarette choices. The study was registered via Clinicaltrials.gov, NCT03194256.

Methods

Recruitment, Screening, and Consent

Participants were recruited between September 2018 and September 2022 for a 6-session laboratory study that required overnight abstinence from nicotine. Participants were recruited throughout Rhode Island and the surrounding geographical area using flyers and social media advertisements, as well as ads in newspapers, on websites, school listservs, on public transportation, and on community-based in-person recruitment events. Interested individuals provided contact information and subsequently completed a telephone screening interview to determine initial eligibility. If the participant was eligible at the phone screen and was under 18, the staff asked permission to call the participant’s parent or guardian. If the parent verbally consented to their child’s participation over the phone, a parental consent form was either mailed to the parent to sign and for their child to bring to their first session or was signed electronically via an emailed REDCap link. At the first session, participants aged 18 or over provided informed consent, and those under 18 provided written assent. All procedures were approved by the Brown University IRB. A waiver of prosecution from Rhode Island’s Attorney General’s office allowed for the provision of study tobacco products in the context of research to those under 21.

Participants

Participants were eligible if they were 15–20 years of age, reported smoking cigarettes daily or nearly every day for at least the prior 3 months, and reported lifetime use of nicotine vaping devices on at least two occasions to ensure that the study did not introduce vaping to never vaping youth, and provided a breath carbon monoxide (CO) of ≥ 5 ppm (or a positive reading on a urine cotinine test). Participants were excluded if they were pregnant or breastfeeding, intending to quit smoking in the next 30 days, used nicotine-containing products other than cigarettes or vaping devices ≥15 days in the past 30 days, used illicit drugs (excluding cannabis), or reported binge drinking more than 9 out of the last 30 days. Participants who reported active intent or plans for suicide within the past 30 days or a suicide attempt within the past 2 years were excluded after speaking with a study medical monitor. Participants also had to be willing to use the study products but had not participated in another research study during the past year in which they were switched to reduced nicotine cigarettes for longer than 1 week. Participants reporting prior adverse reactions to vaping products or food allergies that could potentially be exacerbated by e-liquid flavorings were excluded from the study. All participants were approved by a study medical monitor who ensured their heart rate and blood pressure readings were within an acceptable range, and that they did not have unstable medical or psychological conditions.

Research Cigarettes

Study research cigarettes (Spectrum, 22nd Century Group, Inc.; produced for NIDA; NOT-DA-14-004) were used to model a VLNC standard. The VLNC cigarette contained 0.4 mg/g nicotine, whereas the NNC cigarette contained 15.8 mg nicotine/g tobacco. Participants were provided menthol or non-menthol research cigarettes to sample, based on their preference. Cigarette nicotine condition was masked to participants and investigators.

Nicotine Vaping Devices and E-liquid

The study nicotine vaping device provided was a Kangertech vape pen with a 2 mL size clearomizer, and a 650 mAh 3.7 V EVOD battery with 2.2 ohm coil resistance. E-liquids, obtained from American E-Liquids, had a base of 70% propylene glycol and 30% vegetable glycerin. Participants randomized to the “moderate” nicotine concentration condition received e-liquids with 1.8% freebase nicotine and those randomized to the “low” nicotine concentration condition received e-liquids with 0.3% freebase nicotine; these nicotine concentrations have been shown to yield distinct plasma nicotine concentrations among people naïve to vaping.^21,22 The “tobacco flavors” condition included two options reflecting the available flavors of combusted cigarettes (tobacco and tobacco/menthol). The “non-tobacco flavors” condition included four options (chocolate mint, mango, vanilla, and watermelon). E-liquid nicotine concentration was masked to participants and investigators, whereas e-liquid flavor was by necessity open-label.

Baseline Session

At the first session, consent was obtained, eligibility was confirmed biochemically, and baseline assessments were conducted. Following baseline assessments, participants learned how to use the study nicotine vaping device, sampled all of the available e-liquid flavors, and selected two (one tobacco flavor and one non-tobacco flavor) to be used in the corresponding future experimental sessions. In this initial flavor-sampling session, e-liquids containing no nicotine were used.

Experimental Session 1

Cigarette Only Session.

This and all following sessions were conducted following overnight smoking abstinence, which was confirmed via a breath CO sample of ≤5 ppm, or a reduction of 50% from their baseline CO reading. Sessions were generally conducted 2–7 days apart in an 8’ × 10ʹ laboratory room that was ventilated with a single-pass air handling system; session time of day varied across participants but was held constant within participants to the extent possible. A research assistant monitored adherence to study procedures from outside of the lab room by observing participants through a one-way mirror and communicating via intercom.

Randomization.

Each participant was randomized in a 1:1 ratio to either the NNC or the VLNC cigarettes in a triple-blind fashion. Cartons of study cigarettes were re-labeled with research codes indicating nicotine content and shipped to our research site by a centralized administrative core. Then, participants in each of the two cigarette groups were assigned to all four combinations of factors involving vaping devices (flavor and nicotine concentration each at 2 levels) in counter-balanced order. Vaping materials for each session were prepared by research staff who never interacted with participants; clearomizers were prefilled with assigned e-liquids and labeled only to indicate in which session they were to be administered. Thus, throughout the study, neither participants, research staff, nor investigators knew the level of nicotine being administered in either cigarettes or e-liquid (flavors were not masked).

Cigarette Sampling Phase.

After completing presampling assessments, participants were introduced to the choice-task procedure. They were asked to take 4 puffs of the research cigarette (VLNC or NNC depending on group assignment), with puffs timed 30 seconds apart. Specifically, a computer screen instructed participants to light the cigarette without taking a puff. Then, the screen instructed participants to “Take your first puff now”; a 30-second countdown timer then started. At the end of 30 seconds, participants were instructed to take a second puff. This was repeated until four puffs were taken. Participants then completed postsampling product rating forms; there was a 15-minute delay between the sampling and preference phase.

Preference Phase.

Participants were shown a computer screen with two buttons labeled “2 Cigarette Puffs” and “0 Puffs”. Participants were shown instructions stating that clicking 10 times on the “2 cigarette puffs” button would earn them two puffs from their study cigarette while clicking the “0 puffs’ button 10 times would indicate that they did not wish to smoke on that trial. Participants were instructed to light a new cigarette each time they earned puffs. Participants made a total of 10 choices consisting of a 1-minute consumption phase, with a 2-minute inter-trial interval. Participants then completed post-task assessments.

Experimental Sessions 2–5

Cigarette and Nicotine Vaping Device Sessions Sampling phases.

At each subsequent visit, participants again completed the sampling procedures for the study cigarette and then completed sampling for the nicotine vaping device available for that session, with a 15-minute period between the cigarette sampling and nicotine vaping device sampling phases. In each of these experimental sessions, study cigarette assignment stayed the same (VLNC or NNC, as in Session 1); however, the vaping device e-liquid was one of the four combinations of flavor and nicotine concentration (tobacco 3 mg/mL, non-tobacco 3 mg/mL, tobacco 18 mg/mL, or non-tobacco 18 mg/mL), with order counterbalanced across participants.

Vaping Device Sampling Phase.

After completing the cigarette sampling as described above and following a 15-minute delay, participants completed a similar procedure for vape sampling. The computer screen instructed participants to “pick up the vaping device and press the button,” then 3 seconds later they were instructed to “put the vaping device to your mouth but don’t inhale yet.” After another 2 seconds, they were instructed to “Start puffing now.” After 3 seconds, they were instructed to “stop puffing now.” After a 30-second break between puffs, this procedure was completed three more times. Participants then completed the Product Rating Forms and there was another 15-minute delay before the start of the Choice Task.

Choice Task

Participants were shown a computer screen with three buttons labeled “2 Cigarette Puffs,” “0 Puffs,” and “2 Vape Puffs.” Instructions stated that clicking 10 times on the “2 cigarette puffs’ button would earn them two puffs from their study cigarette, clicking the “0 puffs’ button 10 times would indicate that they did not wish to smoke or vape on that trial, and clicking 10 times on the “2 vape puffs” button would earn them two puffs from the vaping device. Participants made a total of 10 choices consisting of a 1-minute consumption phase, with a 2-minute inter-trial interval, and completed assessments after the end of the task.

Descriptive Measures

Demographics.

Participants were queried on their age, race, sex assigned at birth, and gender identity.

Carbon Monoxide.

Exhaled breath carbon monoxide (ppm) was measured via Bedfont carbon monoxide monitor.

Biomarkers.

Participants submitted urine samples at baseline. Urine was analyzed for the presence of cotinine and total nicotine equivalents, adjusted for creatinine.²³

Timeline Follow-Back.

Cigarette, nicotine vaping device, and other tobacco product use in the past 14 days at baseline was assessed using a calendar-assisted timeline follow-back (TLFB) interview, which was updated between sessions to account for use during the study. From the TLFB, average cigarettes per day and the percentage of days on which vaping was reported were calculated.²⁴

Dependence.

Nicotine dependence was assessed via the six-item Fagerström Test for Nicotine Dependence. The items are summed to yield a total score of 0–10 with higher scores indicating greater physical dependence on nicotine.²⁵

Outcomes

Primary Outcomes.

Outcomes from the Choice Task formed the three primary outcomes. Choices to smoke or vape resulted in 2 puffs per choice, such that each choice was ‘worth’ 2 puffs. Choices to abstain were placed on the same scale, such that choices to abstain were multiplied by 2.

Secondary Outcomes

Cigarette Evaluation Scale.

At baseline, participants were asked to complete the validated Cigarette Evaluation Scale (CES) for their usual brand cigarette. At baseline, participants were instructed to think about their experience with their usual brand of cigarette. After the preference phase and each of the experimental sessions, participants were asked to evaluate their assigned study cigarette on a 7-point scale ranging from “not at all” to “extremely” for each item. The scale comprises 5 subscales: Smoking Satisfaction, Psychological Reward, Aversion, Enjoyment of Respiratory Sensations, and Craving Relief.²⁶

Vaping Device Evaluation Scale.

A modified version of CES, adapted for vaping for the purposes of this study, was administered after each session. The scale comprises the same subscales as the CES described above.

Data Analysis Plan

Although we used a 2 × 2 × 2 design, due to the impact of the COVID-19 pandemic on recruitment, we did not achieve our targeted sample size of 120 due to a prolonged COVID-19-related pause in in-person research. However, at our final sample size of 80, we were able to detect numerous significant main effects with important tobacco regulatory implications, which is what we powered the study to detect and on which we primarily focus. Interactive effects were tested but should be interpreted with caution in light of this limitation.

Choice-task data were analyzed as the total number of choices to smoke cigarettes (session 1) or the total number of choices to smoke cigarettes, the total number of choices to vape, and the total number of choices to abstain (sessions 2–5). Data for Session 1 were analyzed using linear regression. Initially, we fit a model that included the main effect of cigarette nicotine content, e-liquid nicotine concentration, and e-liquid flavors, and adjusted for gender, menthol preference, cigarettes smoked per day, and percent of days reported vaping at baseline, with a random intercept for the subject. In addition, we also fitted a second model that also included all two-way interactions.

Data for sessions 2–5 were analyzed similarly, except that data were analyzed using a linear mixed model with a random effect for participants to account for repeated measurements within an individual. For sessions 2–5, separate models were fitted for the number of choices to smoke, number of choices to vape, and the number of choices to abstain. In each, we fit a model that only included main effects for the three factors, as well as a model that included all two-way interactions.

The analysis of CES subscales was analogous to the analysis of the choice-task data. CES subscales were analyzed using linear mixed models with a random effect for participants to account for repeated measures on a participant. In all cases, we first fit a model that included the main effects for each of the three factors, adjusted for the baseline value of the subscale (for precision) and gender and menthol preference. In addition, we also fitted a model that also included the two-way interactions between the three factors. Finally, Vaping Device Evaluation Scale (VDES) data were analyzed following the same approach with the exception that no baseline levels of the VDES were included, as this was not assessed.

Results

Sample Characteristics

Descriptive characteristics of the sample can be found in Table 1 and a CONSORT diagram of participant flow is shown in Figure 1. All recruited participants were between 18 and 20 years of age (no participants were successfully recruited between the ages of 15–17). Of 124 participants screened, 80 were randomized to the VLNC (n = 39) or NNC (n = 41) cigarette condition. Of those randomized, 85% of the VLNC group and 93% of the NNC group completed the study (Fisher’s exact test p = .31). The COVID-19 pandemic halted recruitment for seven months (March to October, 2020). A summary of differences in enrolled sample characteristics from pre- to post-COVID recruitment can be found in Table S1.

Table 1.

Baseline Characteristics by Group

	NNC (N = 41)	VLNC (N = 39)	Total (N = 80)	p-value
Age (years)				.22
Mean (SD)	18.9 (0.7)	18.7 (0.8)	18.8 (0.8)
Sex at birth				>.99
Female	16 (39.0%)	15 (38.5%)	31 (38.8%)
Male	25 (61.0%)	24 (61.5%)	49 (61.2%)
Gender				.69
Female	13 (31.7%)	15 (38.5%)	28 (35.0%)
Male	24 (58.5%)	22 (56.4%)	46 (57.5%)
Transgender male to female	1 (2.4%)	1 (2.6%)	2 (2.5%)
Transgender female to male	0 (0.0%)	0 (0.0%)	0 (0.0%)
Genderqueer/gender non-conforming	2 (4.9%)	0 (0.0%)	2 (2.5%)
Different identity	0 (0.0%)	1 (2.6%)	1 (1.2%)
Prefer not to answer	1 (2.4%)	0 (0.0%)	1 (1.2%)
Unknown	0 (0.0%)	0 (0.0%)	0 (0.0%)
Ethnicity				.48
Non-Hispanic	32 (78.0%)	34 (87.2%)	66 (82.5%)
Hispanic	7 (17.1%)	5 (12.8%)	12 (15.0%)
Don’t know	2 (4.9%)	0 (0.0%)	2 (2.5%)
Race				.20
White only	27 (65.9%)	24 (61.5%)	51 (63.8%)
Black only	2 (4.9%)	0 (0.0%)	2 (2.5%)
AI or AN only	0 (0.0%)	0 (0.0%)	0 (0.0%)
Asian only	4 (9.8%)	8 (20.5%)	12 (15.0%)
NH/PI only	0 (0.0%)	0 (0.0%)	0 (0.0%)
More than one	8 (19.5%)	5 (12.8%)	13 (16.2%)
Don’t Know	0 (0.0%)	2 (5.1%)	2 (2.5%)
CO				.54
Mean (SD)	7.4 (5.9)	8.1 (4.8)	7.8 (5.3)
CPD				.32
Mean (SD)	5.8 (5.6)	4.7 (3.6)	5.3 (4.7)
FTND				.55
Mean (SD)	2.6 (2.2)	2.2 (1.7)	2.4 (2.0)
Cotinine (ng/mL)	Missing = 0	Missing = 1	Missing = 1
Geometric mean	2426.7 (1426.0, 3427.4)	2577.5 (1330.1, 3824.9)	2499.3 (1723.6, 3274.9)	.85
Total nicotine equivalents (TNEs, ng/mL)	Missing=0	Missing =1	Missing =1	.49
Geometric mean	19.0 (12.8, 28.2)	22.9 (15.1, 34.6)	20.8 (15.6, 27.5)

CO = carbon monoxide; CPD = cigarettes per day; TNEs = total nicotine equivalents; FTND = Fagerstrom Test for Nicotine Dependence; VLNC = very low nicotine content cigarettes; NNC = normal nicotine content cigarettes.

Figure 1.

CONSORT diagram.

Cigarette Preference Phase Outcomes

When only cigarettes were available, there were no significant differences in the mean number of choices to smoke (mean difference −1.2 puffs, 95% CI: −3.4, 0.93; p = .25) between those who were assigned to smoke VLNC and NNC cigarettes.

Choice to Smoke, Vape, or Abstain as a Function of Cigarette Nicotine Content, E-lquid Flavor, and E-liquid Nicotine Concentration

When the nicotine vaping device was concurrently available with cigarettes, participants randomized to the VLNC group smoked an average of 3.5 fewer cigarette puffs than participants randomized to the NNC group (95% CI: −5.5, −1.5; p < .001), while participants randomized to the non-tobacco flavors group smoked an average of 0.75 fewer cigarette puffs than participants randomized to only have access to tobacco flavors (95% CI: −1.5, 0.03; p = .040). There were no significant differences in the number of puffs based on nicotine vaping device nicotine concentration, nor were any of the interactions significant.

When the outcome was choices to vape, VLNC cigarette assignment (95% CI: 2.2, 6.6; p < .001), non-tobacco flavors (95% CI: −2.1,−0.53; p < .001), and 3 mg/mL nicotine concentration (95% CI: 1.0, 2.5; p < .001) were significantly associated with more choices to vape (Table 2, Figure 2). No interaction terms were significant (Table S2). When the outcome was choosing to abstain, lower vaping device nicotine concentration (95% CI: −2.3, −0.95; p < .001) and greater percent of vaping days (95% CI: −0.08, −0.02; p < .001) were significantly associated with fewer choices to abstain. In the interaction models, the interaction term for vaping device flavor and nicotine concentration was significant, such that the combination of tobacco flavor and 3 mg/mL nicotine concentration in the vaping device was associated with significantly fewer choices to abstain (95% CI: −2.7, −0.01; p = .049); no other terms were significant.

Table 2.

Main Effects Models Predicting Each Listed Outcome From Cigarette Assignment Group and Nicotine Vaping Device Characteristics. CES and VDES Subscales Not Included as there were no Significant Effects. Bold text indicates statistical significance, p < 0.05.

	Cigarette nicotine content group (REF = NNC)		Nicotine vaping device flavor (REF = non-tobacco)		Nicotine vaping device nicotine concentration (REF = 18 mg/mL)		TLFB CPD		TLFB % vaping days
	Mean difference beta (95% CI)	p	Mean difference beta (95% CI)	p	Mean difference beta (95% CI)	p	Mean difference beta (95% CI)	p	Mean difference beta (95% CI)	p
Cigarette puffs	−3.5 (−5.5, −1.5)	.00	0.75 (0.03, 1.5)	.04	−0.14 (0.03, 1.5)	.69	0.21 (−0.01, 0.43)	.066	0.02 (−0.01, 0.05)	.12
Nicotine vaping device puffs	4.4 (2.2, 6.6)	.00	−1.3 (−2.1, −0.53)	.00	1.8 (1.0, 2.5)	.00	−0.02 (−0.26, 0.23)	.901	0.03 (0.00, 0.06)	.08
Abstains	−0.90 (−2.9, 1.1)	.36	0.54 (−.15, 1.2)	.12	−1.6 (−2.3, −0.95)	.00	−0.19 (−0.42, 0.03)	.083	−0.05 (−0.08, −0.02)	.00
CES SS	−1.3 (−1.8, −0.78)	.00	−0.11 (−0.27, 0.05)	.18	0.20 (0.04, 0.36)	.01	n/a		n/a
CES PR	−0.88 (−1.3, 0.41)	.00	−0.04 (−0.17, 0.08)	.49	0.04 (−0.09, 0.17)	.53	n/a		n/a
CES ERTS	−1.2 (−1.7, −0.62)	.00	−0.14 (−0.34, 0.05)	.14	0.05 (−0.15, 0.25)	.61	n/a		n/a
CES CR	−1.0 (−1.6, −0.48)	.00	0.00 (−0.20, 0.19)	.96	−0.03 (−0.23, 0.16)	.73	n/a		n/a
VDES SS	0.02 (−0.51, 0.54)	.947	−0.26 (−0.47, −0.04)	.019	0.48 (0.27, 0.69)	<.001	n/a		n/a
VDES A	0.01 (−0.23, 0.26)	.905	−0.09 (−0.22, 0.04)	.160	−0.39 (−0.51, −0.26)	<.001	n/a		n/a
VDES ERTS	−0.03 (−0.56, 0.50)	.912	0.03 (−0.22, 0.28)	.800	0.48 (0.23, 0.72)	<.001	n/a		n/a

TLFB = timeline follow back; CPD = cigarettes per day; CES = Cigarette Evaluation Scale; VDES = vaping device evaluation scale; SS = smoking satisfaction subscale, PR = psychological reward subscale, ERTS = enjoyment of respiratory tract sensations subscale; CR = craving reduction subscale; AV = aversion subscale.

Figure 2.

Choices for cigarette puffs (top panel), nicotine vaping device puffs (bottom left panel), and to choose neither option (bottom right panel), by condition. The y-axes show the maximum range of choices possible (2 puffs over 10 choice sessions). ***A black bracket indicates a significant main effect of cigarette assignment; *a gray bracket indicates a significant main effect of flavor; and ***a dashed line bracket indicates a main effect of nicotine concentration in the nicotine vaping device. VLNC = very low nicotine content cigarettes; NNC = normal nicotine content cigarettes.

CES Outcomes

VLNC cigarette assignment was predictive of lower scores on the psychological reward (95% CI: −1.3, −0.41; p < .001), smoking satisfaction (95% CI: −1.8, −0.78; p < .001), enjoyment of respiratory satisfaction (95% CI: −1.7, −0.62; p < .001), and craving reduction subscales (95% CI: −1.6, −0.48; p < .001; see Table 2 and Table S2). Higher nicotine concentration in the nicotine vaping device was also associated with greater smoking satisfaction (95% CI: 0.04, 0.36; p < .001). The sole significant interaction term was an interaction between VLNC assignment and tobacco flavor, which was associated with lower craving reduction (95% CI: −0.87, −0.09; p < .001).

VDES Outcomes

Lower nicotine concentration in the nicotine vaping device was associated with greater vaping satisfaction (95% CI: 0.27,0.69; p < .001), lower enjoyment of respiratory sensations (95% CI: −0.51, −0.26; p < .001), and lower aversion in main effects models (95% CI: −0.51, −0.26; p < .001; Table 2). Tobacco flavor was associated with lower vaping satisfaction in the main effects model. No interaction terms were significant (Table S2).

Discussion

Given the changing landscape in tobacco product use among youth, studies testing the effects of potential cigarette product standards in early young adults who smoke must consider tobacco products in the context of nicotine vaping device availability and characteristics. The current study assessed the effects of a reduced-nicotine cigarette product standard in youth on choices to smoke, vape, or abstain when the provided (or available) nicotine vaping device varied by nicotine concentration (3 or 18 mg/mL) and flavor (tobacco only or non-tobacco flavors) in a laboratory choice procedure. We first compared choices for cigarette smoking vs. abstinence when only the VLNC- or NNC-assigned cigarette was available. Under these conditions, nicotine content in cigarettes did not significantly affect puffing frequency. In other words, VLNC cigarettes were equally as able to maintain responding as NNC cigarettes in early young adults who did not have access to alternative products. This finding comports with findings from previous acute, short-term VLNC versus NNC choice studies in which each type of cigarette is presented alone^27,28; though in longer-term studies when VLNCs are dispensed, they do reduce smoking in early young adults.¹⁵ In that trial, which excluded alternative tobacco product users and did not provide access to alternatives, there was a significant reduction in smoking but not elimination of smoking or increased quit attempts; thus underscoring that though VLNCs are less preferred and therefore may be smoked less, but to maximize their benefit and increase cessation, alternative options may be necessary.¹⁵

In contrast to when cigarettes were presented alone, concurrent availability of the nicotine vaping device reduced choices to smoke VLNC cigarettes relative to NNC cigarettes, consistent with our hypotheses. This overall effect shows the importance of including an alternative product in the choice task, which can shift behavior in ways that may not be predictable from single-option tasks; this is particularly critical for studying tobacco use behaviors which occur in the context of many possible sources of nicotine that can affect use of cigarettes.^27,29 Furthermore, studies with adults have also shown that providing access to alternative products in conjunction with VLNC cigarettes is associated with the greatest reduction in smoking.^30–32

Furthermore, we found that the nicotine vaping device with non-tobacco flavors was associated with fewer choices to smoke in the main effects models. Flavored nicotine vaping devices are very popular among youth, including many nonsmoking youth^33–35; leading some U.S. state and local governments to enact flavor bans for products popular with youth.³⁶ Given this popularity, it is not surprising that participants made more choices to vape and fewer choices to smoke when allowed to choose among several dessert and fruit flavors. This aligns with data from observational and laboratory studies showing that sweet flavors are preferred among youth,^33,34,37,38 young adults, and older adults.^39,40 Our data suggest that eliminating non-tobacco flavors in nicotine vaping devices may lead to lower vaping among youth but could lead to higher rates of continued smoking in the context of a nicotine reduction standard. We saw few interactions between nicotine concentration and flavor, which is consistent with other work suggesting that flavor and nicotine are unlikely to interact and affect the reward from vaping; however, we were underpowered to examine interactions, and these findings should therefore be considered exploratory.^39,41

Interestingly, the 3 mg/mL lower nicotine e-liquid was also associated with greater choices to vape and fewer choices to abstain, indicating that this e-liquid was preferred to the 18 mg/mL dose. This was contrary to our hypothesis, and may be explained by the finding that lower nicotine vaping device nicotine concentration was associated with less aversion and greater subjective reward on the VDES. This could also be due to the freebase nicotine preparation, rather than a nicotine salt preparation which allows for greater nicotine concentrations with less harshness.⁴² Participants who had fewer days of vaping at baseline also made fewer choices to abstain. This outcome suggests that this combination of vape characteristics may have been most palatable, particularly for those who vaped more frequently to begin with. Other data from laboratory studies have also shown that higher freebase nicotine concentration decreases positive subjective response to nicotine vaping devices.⁴³

These data are somewhat in contrast to data from a recent study with adults, in which participants were assigned to smoke either VLNC or NNC cigarettes and assigned nicotine vaping devices of a combination of either 3 or 18 mg/mL nicotine and tobacco or non-tobacco flavors in a 2 × 2 × 2 factorial design for 12 weeks.⁴⁴ That study found that higher nicotine concentration was associated with fewer cigarettes smoked per day in the VLNC condition; this is in line with reviews indicating that adults generally prefer higher nicotine concentrations.⁴⁵ This difference may also be due to the acute nature of the current study, such that initial aversiveness of higher nicotine concentration may dissipate as users learn to better titrate their nicotine concentration with novel vaping devices to obtain optimal nicotine levels.^46,47 There are differing opinions on the potential benefits and drawbacks of limiting nicotine concentration in nicotine vaping device products. While limiting nicotine may lead to less use in adults, it also may lead to greater overall exposure to toxicants from increased exposure to toxicants in e-liquid⁴⁷; whereas for youth, our data suggests that lower nicotine levels, at least in older freebase formulations, may be associated with greater initial abuse liability. The rapid changes in vaping devices and the nicotine formulations in them pose a significant challenge for regulators as they seek to determine product standards.

This study is novel in being the first to model the impact of nicotine vaping device characteristics in the context of a nicotine reduction standard in early young adults. However, the study had several limitations. First, the nicotine vaping device we chose, a pen-style device, was a preferred product at the time the study was funded; however, this pen-style device was subsequently supplanted in popularity by pod-based devices and, currently, disposable devices; importantly, these newer devices often use nicotine salt-base preparation which can affect appeal.^48,49 While this is a significant limitation, it is an inherent challenge of conducting regulatory research in the context of a rapidly evolving tobacco product landscape. An alternative design might have allowed for this evolving product landscape by changing the e-cigarette device to a pod-based device while the study was ongoing; however, this would have reduced the internal validity of the study. Moreover, the controlled nature of the study allowed us to isolate the importance of flavors in reducing the choice to smoke, which would likely remain true in newer devices. The reduction in aversion at higher doses in salt-based nicotine vaping devices; however, complicates the generalizability of the finding that lower nicotine dose is associated with fewer choices to smoke; future research should use salt-based devices. Another limitation is that we did not enroll anyone younger than 18 years; although we were approved to do so. The necessity of parental consent for a study that required disclosure of an illegal behavior may have hindered the recruitment of this younger group. Finally, we enrolled participants who reported smoking daily or nearly daily to reduce risk to participants, as participants were necessarily exposed to cigarettes in the lab; however, many young people smoke intermittently,^50,51 and therefore these results may not generalize to all young people who smoke.

Overall, these results suggest that a VLNC standard, implemented in a marketplace similar to the current one, may reduce smoking among early young adults. However, the availability and characteristics of nicotine vaping devices, and potentially other alternative tobacco products, influence choices to smoke VLNC cigarettes. As such, regulators should consider that new product standards occur in the context of a complex tobacco marketplace, and the availability and characteristics of alternative products can moderate the impact of the product standard.¹⁸ The potential benefits of a nicotine reduction policy on young people, and public health in general, are greatly influenced by whether young tobacco users respond by increasing their use of alternative combustible or noncombustible tobacco products, or by quitting. As any nicotine use by young people can be harmful, it will be critical to understand these moderating effects of combusted and non-combusted alternative tobacco products in order to reduce the overall harm from tobacco products in the event of a nicotine reduction policy, and to encourage shifting from smoking to complete cessation of nicotine use.

Supplementary material

Supplementary material is available at Nicotine and Tobacco Research online.

Funding

This work was supported by U54DA036114 (PI Donny/Hatsukami). Research reported in this publication was supported by NIDA and the FDA Center for Tobacco Products (CTP). The content is solely the responsibility of the authors and does not necessarily represent the official views of NIH or FDA. Research cigarettes were supplied by NIDA.

Declaration of Interests

The authors have no conflicts to declare.

Author Contributions

Rachel Cassidy (Conceptualization [equal], Investigation [equal], Methodology [equal], Project administration [equal], Writing—original draft [equal]), Jennifer Tidey (Conceptualization [equal], Funding acquisition [equal], Investigation [equal], Methodology [equal], Resources [equal], Supervision [equal], Writing—review & editing [equal]), Julissa Godin (Project administration [equal]), Robert Swift (Investigation [equal]), Connor Demorest (Data curation [equal], Formal analysis [equal]), Mariel Bello (Writing—review & editing [equal]), Rachel Denlinger-Apte (Investigation [equal], Methodology [equal], Project administration [equal], Writing—review & editing [equal]), Christine Goodwin (Project administration [equal]), Patricia Cioe (Investigation [equal], Methodology [equal], Project administration [equal]), Joseph Koopmeiners (Data curation [equal], Formal analysis [equal], Methodology [equal]), Eric Donny (Conceptualization [equal], Data curation [equal], Formal analysis [equal], Funding acquisition [equal], Investigation [equal], Methodology [equal], Project administration [equal]), Dorothy Hatsukami (Conceptualization [equal], Funding acquisition [equal], Investigation [equal], Methodology [equal], Project administration [equal]), and Suzanne Colby (Conceptualization [equal], Funding acquisition [equal], Investigation [equal], Methodology [equal], Project administration [equal])

Data Availability

The primary data set can be made available to researchers upon request, consistent with the data sharing plan in place as part of the grant funding.