Effect on Tobacco Use and Subjective Measures of Including E-cigarettes in a Simulated Ban of Menthol in Combustible Cigarettes

Abstract

Introduction

Bans of menthol characterizing flavor in tobacco products have been proposed; however, there is limited data regarding the impact on current menthol cigarette smokers of including e-cigarettes in such bans.

Methods

In this six-week pilot study, 47 menthol smokers were randomized to receive all tobacco products from an experimental marketplace simulating either no menthol ban, a menthol ban for cigarettes but not e-cigarettes, or a ban for both (“total menthol ban”).

Results

At the first visit, all but one participant selected cigarettes with e-cigarettes selected by 38%, 69%, and 40% of participants in the no ban, menthol cigarette ban, and total menthol ban groups, respectively. Over the study period, the total menthol ban group smoked more than the menthol cigarette ban group (estimated mean ratio [EMR] in cigarettes per day = 1.38; 95% CI: 1.1, 1.75; p = .006). Compared to the no ban condition, the menthol cigarette ban group smoked slightly fewer (EMR = 0.87; 95% CI: .68, 1.11) and the total menthol ban group smoked slightly more (EMR = 1.20; 95% CI: 1.00, 1.45) although neither difference reached statistical significance. In both menthol ban conditions, ratings were lower (vs. no ban) on several measures of craving and cigarette effects and liking.

Conclusions

Menthol bans that include e-cigarettes may result in different patterns of tobacco use than if only combustible cigarettes are included, although e-cigarettes were not extensively used in any group. Larger studies are needed to determine policies most likely to provide the largest public health benefit.

Implications

Bans of menthol characterizing flavor have been proposed, however, the effects on menthol cigarette smokers of including e-cigarettes in such bans are not clear. This study found that smokers randomized to a simulated ban on menthol in both cigarettes and e-cigarettes smoked more cigarettes per day over the 6-week study period than those randomized to a simulated ban on menthol in only cigarettes suggesting that smoking patterns among current menthol smokers differ depending on which products are included in a menthol ban. Larger studies are needed to determine the policies most likely to provide the largest public health benefit.

Introduction

Reports examining the effects of menthol characterizing flavor in cigarettes have concluded that banning menthol in combustible cigarettes would benefit public health.^1–4 Some municipalities have, therefore, banned characterizing menthol flavorants (in either some or all tobacco products) and the FDA has recently taken the initial steps necessary to enact a national ban on menthol characterizing flavor in combustible cigarettes and cigars in the United States.^5–7

Naturalistic and experimental studies suggest that banning menthol in combustible cigarettes is associated with decreased cigarette use in current menthol smokers.⁸ For example, cohort studies in Canada have found that after the implementation of a menthol ban, smoking cessation rates were higher in those who prior to the ban smoked menthol (vs. non–menthol) cigarettes.^9,10 Experimental studies that have simulated menthol cigarette bans have generally found that smoking decreases when menthol smokers switch to non–menthol cigarettes.^11–13 Nonetheless, these studies have found that following either an actual or simulated menthol cigarette ban, most menthol cigarette smokers switch to non–menthol cigarettes, and continue to smoke combustible cigarettes. It is possible that some menthol smokers, rather than switching to non–menthol combustible cigarettes, would switch to using flavored non–cigarette tobacco products were they available.^14–16 There is currently limited data for current menthol cigarette smokers comparing the effects of banning menthol from only combustible cigarettes relative to a broader ban that also includes e-cigarettes. Since e-cigarettes are thought to be less harmful than combustible cigarettes,¹⁷ if continued availability of menthol-flavored e-cigarettes resulted in substitution to non–combustible products rather than combustible non–menthol cigarettes that would likely be of benefit to current menthol cigarette smokers. Conversely, if the availability of menthol-flavored e-cigarettes undermined the effects of a menthol cigarette ban on smoking behavior that would be detrimental to current menthol cigarette smokers. Effect on current smokers is one factor to be considered in determining the overall public health effects of a policy that restricts characterizing flavors from e-cigarettes.¹⁸

The primary aim of this study was to provide initial data regarding the impact on current menthol cigarette smokers of either including or excluding e-cigarettes from a ban on menthol characterizing flavor. We hypothesized that menthol cigarette smokers would smoke fewer cigarettes, vape more e-cigarettes, have higher motivation to quit smoking, and be more supportive of a ban on menthol cigarettes when given the option of using menthol-flavored e-cigarettes than when no such option was present.

Methods

Design

In this six-week study, menthol cigarette smokers were asked to only use tobacco/ nicotine products that they received from an experimental marketplace set up for this project. Participants were randomly assigned to one of three experimental marketplaces which differed in the products that were available to choose from. The marketplaces simulated conditions in which there were: 1. no ban on menthol in either cigarettes or e-cigarettes (ie, “no ban” condition); 2. a ban on menthol in cigarettes but not e-cigarettes (ie, “menthol cigarette ban” condition); and 3. a ban on menthol in both cigarettes and e-cigarettes (ie, “total menthol ban” condition). Outcome measures assessed included tobacco product use, biomarkers of exposure, and subjective measures.

Participants

To be eligible, participants needed to be between the ages of 18 and 64 years at the time of enrollment (minimum eligible age increased to 21 years after the legal age to purchase cigarettes increased), smoke ≥ 5 cigarettes per day, and report that they purchase only menthol cigarettes. Excluded were those who regularly used tobacco products other than cigarettes (> 9 times per month), reported current serious and unstable medical or psychiatric conditions, used medication that could interfere with the study outcomes (eg, smoking cessation medications), or were pregnant, or breastfeeding. To reduce heterogeneity in this pilot study, we only enrolled participants who at the time of study entry rated their motivation to quit smoking “at this time” as < 7 (out of 10). Eligibility criteria were assessed by self-report except for pregnancy status (confirmed via a urine pregnancy test) and smoking status (confirmed by an exhaled carbon monoxide (CO) concentration of > 8 ppm or if not meeting the exhaled CO criteria, a measure of 6 on NicAlert urine cotinine test strips).

Experimental Marketplace Description

The experimental marketplace was designed to allow the evaluation of product preferences in a systematic and ecologically valid manner. The marketplace included two types of e-cigarettes commonly used at the time of study initiation—a cigarette-like e-cigarette system (ie, Vuse Solo) and a tank-like e-cigarette system (ie, Joyetech eGO AIO). In the no menthol ban condition, participants had available to them menthol, and non–menthol cigarettes as well as menthol and tobacco-flavored e-cigarettes (other e-cigarette flavors were not available since the focus of this study was specifically on menthol characterizing flavor). In the menthol cigarette ban condition, only non–menthol cigarettes were available, but both menthol and tobacco-flavored e-cigarettes were also available. In the total menthol ban condition, only non–menthol cigarettes, and tobacco-flavored e-cigarettes were available. Nicotine gum and lozenge in various flavors were available in all three conditions. The cigarette brands available were the participant’s usual brand of menthol cigarettes in the no–ban condition and their preferred brand of non–menthol cigarettes in all conditions. For those with no brand preference for non–menthol cigarettes, the non–menthol version of their usual brand was used.

Following a screening visit, four visits occurred separated by two-week intervals (ie, baseline, week 2, week 4, and week 6 visits). At the baseline, week 2 and week 4 visits each participant was given 84 credits each “worth” $1.00 that could be exchanged for products in their marketplace. To discourage needless buying, unused credits could be used at subsequent visits or exchanged for cash at the end of the study. Similar to other experimental marketplace studies, products were discounted to approximately 66% of their retail market value to discourage purchasing from outside the marketplace and to prevent hoarding credits for payment at the end of the study.^19,20 Sufficient points were provided to approximately allow a one pack per day smoker to replace their cigarettes with either cigarettes, e-cigarettes, or some combination. The products were provided with the packaging that normally accompanies them with no specific additional instructions provided. No penalties were levied for outside purchases in order to encourage truthful reports.

Procedures

This study was approved by the University of Minnesota Institutional Review Board. At a screening visit, written informed consent was obtained and eligibility confirmed. All visits occurred between February 2018 and March 2020. In-person visits were suspended in March 2020 due to Coronavirus Disease 2019 (COVID-19) restrictions and for those enrolled at the time of the suspension (n = 4), a final phone call was attempted to collect data. Randomization schedules were generated using permuted block randomization, stratified by gender with the assignment revealed to research staff prior to a participant’s baseline visit and to participants at the baseline visit. After the screening visit, participants were sent a link daily to a survey asking about the number of cigarettes smoked that day. Following the baseline visit, the daily surveys also asked how many times an e-cigarette was used and for each use, an estimate of the duration of use, and how many puffs were taken. At each in-person visit, exhaled CO was measured, and at all visits, questionnaires assessing subjective measures were completed. Urinary cotinine was assayed from urine collected at the baseline and week 6 visits.²¹

The Minnesota Nicotine Withdrawal Scale (MNWS) and the short form of the Tobacco Craving Questionnaire (TCQ) were completed at each visit. MNWS craving was scored by the “craving for cigarettes” item and withdrawal symptoms by the sum of the other 7 items.^22,23 From the TCQ, a total score, and scores for factors reflecting emotionality, expectancy, compulsivity, and purposefulness were calculated.²⁴ At each visit, participants rated (on a 10-point scale) each of 17 measures of drug liking or drug effects for the cigarettes smoked and e-cigarettes used with a scale adapted from items in instruments previously described.^25,26 Motivation and self-efficacy for quitting smoking were assessed at all visits.²⁷ Perceived health risks (on a 10-point scale) for nine health problems were rated at the baseline visit based on usual cigarettes smoked and at subsequent visits based on all of the tobacco products they have been using and separately on the e-cigarettes they have been using (for those that used e-cigarettes).²⁸ Nicotine dependence (using the Fagerstrom Test for Nicotine Dependence) and support for a ban on menthol (on a 10-point scale from “Not Supportive” to “Very Supportive”) in commonly used categories of tobacco products (ie, cigarettes, e-cigarettes, smokeless tobacco products) and in all tobacco products were assessed at baseline and week 6.²⁹

Statistical Analysis

Participants’ reported daily cigarette (total cigarettes per day) and e-cigarette use (total minutes and total puffs per day) were analyzed as weekly means for each individual. If no report was provided for all days in a given week, that week’s average was considered missing. Otherwise, it was defined as the mean product use over all reported days that week. The day of the baseline visit was not included. Baseline product use was defined as the average over all reported days in the seven days prior to the baseline visit.

The primary analysis of cigarette use was conducted with longitudinal regression models using generalized estimating equations (GEE) to account for repeated measurements, assuming a Poisson error distribution and exchangeable correlation for weekly observations from the same participant. The independent variables were the randomization group (three categories), weeks since baseline visit (continuous), and baseline cigarettes per day (continuous). Post-hoc analyses evaluating if the use of e-cigarettes was associated with changes in cigarette smoking were conducted in the entire sample and separately in the menthol cigarette ban group (the group in which the largest proportion of participants selected e-cigarettes from the marketplace) using the same model but replacing randomization group with e-cigarette use group (ie, none, any menthol e-cigarette use, any non-menthol e-cigarette use).

For e-cigarettes, because only about one-half of participants reported any use and there was high variation among those who did, descriptive statistics were calculated and rank-based Kruskal-Wallis tests were used to compare groups. Baseline adjustment was not necessary due to almost no baseline use (a total of 4 participants reported any e-cigarette use in the week before the baseline visit with the total number of puffs used during the week for each participant being 2, 5, 12, and 58 puffs). The three study groups were compared for each week and with a combined measure of average product use (minutes or puffs) over the entire six-week study period.

Exhaled CO concentrations, urinary cotinine concentrations, and subjective measures were analyzed with longitudinal regression models using GEE, assuming a normal error distribution and exchangeable correlation for weekly observations from the same participant. Independent variables were as described for the cigarette use analysis. Exhaled CO and cotinine concentrations were log-transformed. Analysis was done using R software.^30,31

Results

Among the 47 participants who completed the baseline visit, 16 were randomized to the no menthol ban condition, 16 to the menthol cigarette ban condition, and 15 to the total menthol ban condition. In the no menthol ban condition, 44% of participants were female, 31% were African American, 56% were White, the average (SD) age was 44.1 (10.4) years and at the time of enrollment 81% (n = 13) reported having ever used e-cigarettes in the past (25% once or a couple of times, 38% some days, 13% every day and 6% not providing information about the extent of past use). In the menthol cigarette ban condition, 50% of participants were female, 44% were African American, 44% were White, the average (SD) age was 45.8 (9.8) years and 63% (n = 10) reported having ever used e-cigarettes in the past (31% once or a couple of times, 6% some days and 25% every day). In the total menthol ban condition, 47% of participants were female, 27% were African American, 67% were White, the average (SD) age was 44.9 (11.2) years and 80% (n = 12) reported having ever used e-cigarettes in the past (40% once or a couple of times, 7% some days and 33% every day). Thirty-nine participants completed the entire study (14 in the no menthol ban condition, 13 in the menthol cigarette ban condition, and 12 in the total menthol ban condition—see Supplementary Figure).

Choice of Products

At the baseline visit, all participants in the no menthol ban group selected menthol cigarettes with 38% also selecting an e-cigarette with menthol-flavored e-liquid and no participants selecting tobacco flavored e-liquid or medicinal nicotine. In the menthol cigarette ban group, 94% selected non-menthol cigarettes with 69% also selecting an e-cigarette (all of whom selected menthol flavored e-liquid) and 13% also selecting medicinal nicotine. One participant selected only e-cigarettes at the baseline visit (but selected cigarettes at subsequent visits). In the total menthol ban group, all participants selected non–menthol cigarettes with 40% also selecting an e-cigarette with tobacco-flavored e-liquid (as that was the only available flavor) and 13% selecting medicinal nicotine. Among all 47 participants, at the baseline visit a similar number selected each of the two e-cigarette devices (21% selected Vuse, 23% selected Joyetech, 4% selected both). Relatively few participants at the week 2 or week 4 visit chose the alternate e-cigarette device (n = 3), selected additional e-liquid (3 in the no menthol ban group, 3 in the menthol cigarette ban group, and 4 in the total menthol ban group) or selected medicinal nicotine (1 in the no menthol ban group and 1 in the total menthol ban group).

Tobacco Use and Biomarkers of Exposure

Participants assigned to the no menthol ban condition reported that 2.1% of cigarettes smoked were non–menthol (total reported non–menthol cigarettes smoked divided by total cigarettes smoked). Among those in the menthol cigarette ban condition, 1.2% of reported cigarettes smoked were menthol-flavored and among those in the total menthol ban condition, 3.3% of reported cigarettes were menthol. Over the six-week study period, those randomized to a total menthol ban smoked significantly more total cigarettes than those randomized to a menthol cigarette ban (estimated mean ratio in cigarettes smoked per day = 1.38; 95% CI: 1.1, 1.75; p = .006). Relative to the no menthol ban group, those in the menthol cigarette ban group smoked slightly fewer (estimated mean ratio = 0.87; 95% CI: .68, 1.11; p = .260) and those in the total menthol ban group smoked somewhat more cigarettes per day (estimated mean ratio = 1.20; 95% CI: 1.00, 1.45; p = .057) although neither difference reached statistical significance (Figure 1). Exhaled CO concentrations followed a similar pattern with CO numerically higher in the total menthol ban group compared to the menthol cigarette ban group (estimated mean ratio = 1.23; 95% CI: .98, 1.56; p = .08) (Table 1). Urinary cotinine concentration (normalized for creatinine) was not higher in the total menthol ban condition versus the menthol cigarette ban condition (estimated mean ratio = 0.94; 95% 0.72, 1.21) suggesting that those in the menthol cigarette ban condition may have been obtaining more nicotine from e-cigarettes (Table 1). Post-hoc analysis suggests that menthol-flavored e-cigarette use was associated with decreased cigarette smoking. The estimated mean ratio of cigarettes smoked per day in those who used menthol e-cigarettes (n = 17) relative to no e-cigarette use (n = 24) was 0.89 (95% CI: .71, 1.10) among all participants and 0.59 (95% CI: .37, .96) among those in the menthol cigarette ban group. Among all participants, the estimated mean ratio of cigarettes smoked per day in those who used non–menthol e-cigarettes (n = 6) relative to no e-cigarette use was 1.12 (95% CI: .84, 1.51). As these were not pre-specified outcomes measures, they should be considered exploratory.

Table 1.

Raw mean (SD) of smoking biomarkers and subjective measures reported at each visit. In the comparison between groups for smoking biomarkers, a geometric mean ratio (95% CI) is indicated; for all other measures, the mean difference between groups (95% CI) is indicated. These represent the average difference between randomized groups over 6 weeks, adjusting for baseline values, using a linear regression model (one linear regression model per measure/ outcome). Shaded cells are those in which the 95% confidence interval does not cross zero (or one for the smoking biomarkers).

Measure	No Menthol Ban				Menthol Cigarette Ban					Total Menthol Ban					Mean difference (total ban vs cig ban)
	Week 0	Week 2	Week 4	Week 6	Week 0	Week 2	Week 4	Week 6	Mean difference vs. no ban (95% CI)	Week 0	Week 2	Week 4	Week 6	Mean difference vs. no ban (95% CI)
Exhaled CO	16.1 (7.7)	16.2 (6.6)	16.9 (9.1)	14.8 (7.3)	18.9 (10.7)	15.6 (9.0)	13.9 (6.1)	15.3 (6.8)	0.82* (0.64,1.06)	13.3 (5.4)	12.4 (5.1)	16.2 (8.8)	15.5 (8.6)	1.02* (0.79, 1.31)	1.23* (0.98, 1.56)
Urinary cotinine (ng/mg creatinine)	2668 (1375)			3231 (1860)	3072 (1355)			3125 (1188)	0.92* (0.66, 1.27)	2912 (1486)			2855 (1643)	0.86* (0.62, 1.19)	0.94* (0.72, 1.21)
Dependence/ Craving/ Withdrawal**
FTND	5.2 (1.5)			4.9 (1.7)	5.0 (1.6)			4.8 (1.6)	0.12 (-0.83, 1.06)	5.6 (1.9)			5.2 (2.1)	0.15 (-0.70,1.0)	0.03 (-0.76, 0.82)
Craving(from MNWS)	2.4 (1.0)	2.3 (0.6)	2.5 (1.1)	2.1 (0.9)	2.4 (1.1)	2.2 (1.1)	1.9 (1.2)	2.0 (0.9)	-0.30 (-0.76, 0.16)	2.5 (0.8)	2.2 (1.1)	2.4 (0.7)	2.0 (0.8)	-0.18 (-0.59, 0.23)	0.12 (-0.32, 0.56)
Withdrawal(from MNWS)	6.8 (7.2)	6.4 (6.5)	6.3 (6.2)	5.9 (6.2)	5.6 (6.1)	6.2 (5.8)	5.9 (6.4)	4.7 (5.7)	-0.82 (-3.80, 2.17)	5.8 (3.7)	3.6 (5.2)	3.6 (5.0)	3.6 (4.2)	-2.71 (-5.52, 0.09)	-1.90 (-4.56, 0.77)
TCQ Total Score	46.9 (13.5)	49.1 (16.5)	48.9 (17.2)	46.8 (13.7)	50.5 (12.8)	47.7 (8.8)	43.7 (13.4)	39.7 (12.3)	-7.04 (-12.00, -2.08)	47.4 (14.7)	44.9 (13.7)	42.8 (12.0)	38.8 (13.1)	-6.91 (-12.63, -1.19)	0.13 (-4.49, 4.74)
TCQ Emotional	9.4 (4.2)	9.7 (9.0)	10.7 (5.1)	10.2 (4.8)	9.2 (5.6)	8.9 (4.3)	8.1 (5.3)	7.2 (4.1)	-2.11 (-4.17, -0.04)	8.9 (4.4)	8.8 (4.3)	7.9 (4.4)	7.0 (3.8)	-2.19 (-3.98, -0.40)	-0.09 (-1.99, 1.82)
TCQ Expectancy	15.5 (3.9)	16.1 (4.4)	15.9 (4.5)	15.6 (4.1)	16.9 (3.8)	16.5 (2.9)	15.1 (3.7)	15.0 (3.5)	-1.00 (-2.62, 0.61)	17.0 (3.9)	15.3 (3.9)	15.2 (2.9)	14.2 (3.8)	-1.85 (-3.28, -0.42)	-0.84 (-2.52, 0.83)
TCQ Compulsivity	8.6 (4.3)	9.5 (6.0)	10.0 (5.0)	8.5 (4.4)	9.7 (3.9)	8.6 (4.5)	8.4 (4.7)	7.2 (5.2)	-2.18 (-4.07, -0.29)	8.6 (5.4)	8.2 (4.1)	7.7 (3.8)	7.1 (4.0)	-2.12 (-3.97, -0.26)	0.06 (-1.69, 1.82)
TCQ Purposefulness	13.4 (4.6)	13.8 (4.7)	12.3 (5.2)	12.5 (4.5)	14.6 (2.9)	13.7 (2.7)	12.1 (3.1)	10.3 (4.9)	-1.52 (-3.58, 0.54)	12.9 (3.8)	12.6 (3.4)	12.1 (3.3)	10.6 (3.8)	-0.88 (-3.06, 0.43)	0.64 (-1.04, 2.32)

* Geometric mean ratio.

** Range (min—max) of possible scores for each scale: FTND (0 – 10); Craving from MNWS (0 – 4); Withdrawal from MNWS (0 – 28); TCQ Total (12 – 84); TCQ Emotional (3 – 21); TCQ Expectancy (3 – 21); TCQ Compulsivity (3 – 21); TCQ Purposefulness (3 - 21).

Figure 1.

Total Number (Raw Mean ± SD) of Cigarettes Smoked in Each Group.

Although there were no statistically significant differences among groups in self-reported e-cigarette use as measured by either number of puffs used per day or the number of minutes of e-cigarette use per day, relatively few participants reported any e-cigarette use. During week 1, six participants in the no menthol ban group, nine participants in the menthol cigarette ban group, and 6 participants in the total menthol ban group reported any e-cigarette use with fewer reporting use in subsequent weeks (Supplementary Table 1). Mean ± SD e-cigarette use per randomized participant over the 6 week study period was 5 ± 9 puffs and 17 ± 34 minutes per day in the no menthol ban group; 57 ± 180 puffs and 113 ± 194 minutes per day in the menthol cigarette ban group; and 11 ± 25 puffs and 75 ± 256 minutes per day in the total menthol ban group. The greatest differences were observed over the first two weeks with use in the menthol cigarette ban group subsequently declining (Supplementary Table 1). Removing one outlier in the menthol cigarette ban group who reported substantially greater e-cigarette use than any other participant (the same participant who did not select any cigarettes from the marketplace at the baseline visit) decreased use per participant in that group to 12 ± 15 puffs per day and 89 ± 177 minutes per day.

A total of 7 participants reported any medicinal nicotine use during the study with only 1 participant using more than a total of 10 doses during their participation (that participant reported using a total of 18 doses of nicotine gum).

Subjective Measures

Several differences were observed among groups in measures of craving/ withdrawal symptom severity. Specifically, total and factor scores on the TCQ for emotionality (ie, smoking in anticipation of relief from withdrawal or negative mood) and compulsivity (an inability to control tobacco use) were lower in both menthol ban groups than in the no ban group. The expectancy factor score (ie, the anticipation of positive outcomes from smoking) was only significantly lower in the total ban group with scores numerically lower for all other factor scores in both menthol ban groups compared to the no menthol ban group (Table 1). No substantial differences among groups were found in measures of motivation or self-efficacy to quit smoking, in nicotine dependence, support for a ban on menthol in tobacco products, or on measures of perceived health risks from the products used (Table 2). Participants in both menthol ban groups rated their cigarettes significantly lower on a number of measures from the effects and liking scale than did those in the no menthol ban group (Table 3). Perceived health risks of e-cigarette use and e-cigarette effects and liking measures were generally rated lower than those of combustible cigarettes; however, the sample size was small as the analysis was limited only to those who used e-cigarettes (Supplementary Table 2).

Table 2.

Raw mean (SD) of subjective measures reported at each visit. In the comparison between groups, mean difference between groups, (95% CI) is indicated. These represent the average difference between randomized groups over 6 weeks, adjusting for baseline values, using a linear regression model (one linear regression model per measure/ outcome).

Measure	No Menthol Ban				Menthol Cigarette Ban					Total Menthol Ban					Mean difference (total ban vs cig ban)
	Week 0	Week 2	Week 4	Week 6	Week 0	Week 2	Week 4	Week 6	Mean difference vs. no ban (95% CI)	Week 0	Week 2	Week 4	Week 6	Mean difference vs. no ban (95% CI)
Support for Menthol Ban in Tobacco Product Categories (each item on a 1 – 10 scale)
Ban in cigarettes	1.5 (1.5)			1.5 (1.0)	1.1 (0.2)			1.4 (1.2)	0.12 (-0.68, 0.91)	2.1 (2.5)			1.9 (1.8)	0.30 (-0.62, 1.22)	0.18 (-0.73, 1.10)
Ban in e-cigarettes	2.6 (2.1)			2.7 (2.4)	2.6 (2.4)			4.1 (3.3)	1.34 (-0.52, 3.20)	3.0 (3.2)			2.7 (2.4)	1.20 (-0.87, 3.26)	-0.14 (-2.43, 2.14)
Ban in smokeless tobacco	3.5 (3.7)			3.2 (3.2)	3.0 (2.9)			3.9 (3.3)	0.84 (-1.13, 2.80)	2.7 (2.6)			2.9 (2.8)	0.19 (-1.62, 2.00)	-0.65 (-2.69, 1.39)
Complete menthol ban	1.2 (0.7)			1.6 (1.1)	1.6 (2.2)			1.8 (2.4)	-0.22 (-0.89, 0.45)	1.7 (1.5)			2.3 (2.8)	0.33 (-0.68, 1.34)	0.55 (-0.36, 1.47)
Motivation and Self-Efficacy to Quit (each item on a 1 – 10 scale)
Motivation to quit smoking	2.6 (1.8)	3.1 (2.3)	3.3 (2.3)	3.5 (2.2)	2.5 (1.8)	3.7 (2.1)	3.8 (2.5)	3.7 (3.0)	0.60 (-0.45, 1.65)	3.1 (2.1)	2.9 (2.1)	3.2 (2.5)	3.5 (2.6)	-0.27 (-1.21, 0.67)	-0.87 (-1.92, 0.19)
Confidence to quit smoking	3.4 (1.7)	3.3 (2.2)	3.3 (1.8)	3.4 (1.7)	3.2 (2.5)	3.5 (2.2)	3.9 (2.4)	3.8 (3.4)	0.44 (-0.44, 1.32)	2.6 (1.8)	3.4 (2.8)	3.7 (3.1)	2.8 (2.1)	0.68 (-0.41, 1.78)	0.24 (-0.80, 1.29)
Effectiveness to quit smoking	3.9 (2.4)	4.2 (2.2)	3.2 (2.0)	4.0 (2.1)	3.6 (2.2)	3.3 (2.5)	3.4 (2.6)	4.0 (3.4)	-0.25 (-1.42, 0.92)	3.4 (2.1)	3.0 (2.5)	3.8 (3.0)	3.2 (2.9)	-0.22 (-1.43, 1.00)	0.04 (-1.14, 1.22)
Effort to quit smoking	2.6 (2.0)	2.7 (2.1)	3.1 (2.6)	3.8 (2.8)	2.9 (2.4)	3.9 (2.5)	3.3 (2.3)	3.3 (2.9)	0.30 (-0.88, 1.48)	2.7 (1.7)	2.6 (1.8)	3.0 (2.2)	3.2 (2.3)	-0.04 (-1.02, 0.94)	-0.34 (-1.38, 0.70)
Perceived Health Risk for All Tobacco Products Being Used (each item on a 1 – 10 scale)
Lung cancer	6.6 (2.3)	7.0 (2.5)	6.9 (2.5)	6.6 (2.5)	8.1 (1.9)	7.0 (2.8)	7.6 (2.4)	6.6 (3.1)	-1.13 (-2.56, 0.30)	6.7 (2.6)	6.4 (3.5)	6.5 (3.4)	6.2 (3.4)	-0.90 (-2.45, 0.65)	0.23 (-1.42, 1.88)
Emphysema	6.2 (2.0)	6.1 (3.1)	7.0 (2.7)	6.2 (2.7)	7.4 (2.5)	6.0 (2.9)	6.6 (2.8)	5.8 (2.6)	-1.33 (-2.93, 0.27)	6.3 (3.1)	6.6 (3.5)	5.6 (3.5)	5.8 (3.2)	-0.85 (-2.56, 0.85)	0.47 (-1.20, 2.14)
Bronchitis	5.9 (2.3)	6.1 (3.1)	6.8 (2.8)	6.1 (2.7)	5.9 (2.7)	5.8 (2.8)	6.0 (3.0)	6.5 (2.0)	-0.29 (-1.78, 1.19)	5.5 (2.6)	5.6 (3.1)	6.0 (3.2)	5.8 (3.3)	-0.41 (-1.88, 1.05)	-0.12 (-1.57, 1.33)
Other cancers	5.6 (2.3)	5.6 (3.2)	5.8 (3.1)	5.1 (2.5)	7.6 (2.4)	5.9 (3.0)	6.3 (2.5)	5.7 (3.0)	-1.21 (-2.75, 0.32)	4.8 (2.5)	5.2 (3.1)	5.0 (2.9)	5.4 (3.0)	-0.01 (-1.49, 1.46)	1.20 (-0.46, 2.86)
Heart disease	5.9 (2.8)	6.5 (2.3)	6.6 (2.3)	5.4 (2.7)	7.4 (2.3)	5.9 (2.7)	6.4 (2.4)	5.8 (2.8)	-1.13 (-2.47, 0.21)	5.7 (2.4)	6.1 (3.3)	5.2 (3.2)	5.5 (2.9)	-0.54 (-1.85, 0.77)	0.59 (-1.04, 2.23)
Risk of addiction	8.6 (2.0)	8.7 (2.0)	8.4 (2.1)	7.6 (3.2)	9.2 (1.2)	7.3 (3.5)	8.1 (1.8)	8.8 (2.1)	-0.80 (-1.79, 0.19)	8.5 (2.6)	7.6 (3.0)	8.5 (2.5)	8.2 (2.6)	-0.34 (-1.29, 0.60)	0.46 (-0.38, 1.29)
Stroke	5.6 (2.6)	5.7 (2.7)	5.6 (2.4)	5.1 (2.1)	6.1 (2.5)	5.3 (2.7)	5.6 (2.2)	5.4 (2.8)	-0.61 (-1.93, 0.72)	4.7 (2.7)	5.6 (3.0)	5.8 (2.6)	5.2 (2.9)	0.35 (-0.79, 1.50)	0.96 (-0.39, 2.31)
Mouth cancer	5.6 (2.8)	5.4 (3.4)	5.0 (3.2)	4.6 (2.6)	5.5 (3.0)	5.8 (2.9)	6.4 (3.2)	5.9 (3.1)	0.66 (-0.95, 2.27)	4.6 (2.8)	5.1 (2.9)	4.8 (2.9)	4.8 (3.1)	0.09 (-1.08, 1.26)	-0.57 (-2.13, 1.00)
Tooth loss	4.6 (3.2)	4.7 (3.7)	4.3 (3.2)	4.1 (3.1)	5.0 (3.4)	5.0 (3.3)	5.6 (3.4)	5.2 (3.7)	0.34 (-1.27, 1.94)	4.3 (2.4)	5.2 (3.2)	3.8 (3.2)	4.2 (2.9)	-0.05 (-1.28, 1.18)	-0.39 (-1.89, 1.12)

Table 3.

Raw mean (SD) of measures reported at each visit. In the comparison between groups, the mean difference between groups (95% CI) is indicated. These represent the average difference between randomized groups over 6 weeks, adjusting for baseline values, using a linear regression model (one linear regression model per measure/ outcome). Shaded cells are those in which the 95% confidence interval does not cross zero.

	No Menthol Ban				Menthol Cigarette Ban					Total Menthol Ban
Measure	Week 0	Week 2	Week 4	Week 6	Week 0	Week 2	Week 4	Week 6	Mean difference vs. no ban (95% CI)	Week 0	Week 2	Week 4	Week 6	Mean difference vs. no ban (95% CI)	Mean difference (total ban vs cig ban)
Cigarette Effects and Liking Measures (each item on a 1 – 10 scale)
Reduced craving	7.9 (2.4)	9.0 (1.6)	7.8 (2.1)	7.8 (1.8)	7.3 (2.9)	6.5 (3.1)	6.3 (2.6)	5.5 (3.2)	-1.95 (-3.36, -0.55)	7.9 (2.3)	5.1 (3.5)	6.5 (3.1)	6.8 (2.3)	-2.36 (-3.81, -0.91)	-0.41 (-2.11, 1.30)
Good Effects	7.3 (2.3)	7.6 (1.7)	6.5 (2.5)	6.9 (1.8)	6.8 (2.2)	6.1 (1.9)	4.3 (2.6)	4.4 (2.5)	-1.98 (-2.82, 1.13)	6.5 (2.2)	4.4 (2.1)	6.0 (2.6)	5.2 (2.4)	-1.68 (-2.75, -0.60)	0.30 (-0.91, 1.51)
Bad Effects	2.9 (2.1)	2.3 (2.3)	4.1 (3.0)	3.3 (2.4)	3.9 (2.7)	3.7 (2.6)	3.1 (2.4)	2.2 (1.6)	-0.52 (-1.82, 0.77)	3.6 (1.8)	3.1 (2.1)	4.0 (1.8)	3.7 (1.8)	0.08 (-1.17, 1.33)	0.61 (-0.47, 1.68)
Satisfying	8.4 (1.4)	8.6 (1.4)	8.2 (1.7)	8.0 (1.2)	8.2 (2.3)	6.9 (2.6)	6.6 (1.8)	6.5 (2.0)	-1.56 (-2.59, -0.53)	9.1 (1.5)	6.6 (2.3)	6.0 (2.3)	6.6 (2.6)	-2.02 (-3.17, -0.86)	-0.46 (-1.79, 0.87)
Like	8.4 (1.4)	8.7 (1.4)	7.8 (1.9)	8.1 (1.1)	8.0 (2.4)	5.1 (3.4)	5.0 (2.3)	5.9 (3.2)	-2.65 (-3.90, -1.41)	8.9 (1.2)	5.6 (2.6)	5.3 (2.9)	5.3 (3.0)	-3.01 (-4.42, -1.60)	-0.36 (-2.12, 1.41)
Desire	8.2 (1.3)	7.9 (2.0)	7.7 (2.5)	7.4 (1.6)	8.3 (1.2)	5.9 (2.5)	5.6 (2.8)	5.8 (2.4)	-2.00 (-3.16, -0.85)	8.3 (1.4)	5.9 (2.2)	6.1 (2.1)	5.5 (2.9)	-1.92 (-3.13, -0.70)	0.09 (-1.24, 1.41)
Strong	7.1 (2.1)	6.5 (2.6)	7.1 (2.6)	6.5 (2.6)	6.5 (1.8)	6.1 (2.5)	5.9 (2.2)	5.8 (1.3)	-0.60 (-1.79, 0.60)	6.7 (2.5)	5.7 (1.9)	5.5 (1.9)	5.0 (2.2)	-1.25 (-2.28, -0.21)	-0.65 (-1.77, 0.47)
Alert	5.6 (2.8)	5.6 (3.3)	6.3 (2.5)	4.8 (3.3)	4.8 (2.6)	3.5 (2.9)	3.4 (2.7)	3.2 (2.8)	-1.82 (-3.16, -0.47)	5.8 (3.2)	3.7 (2.4)	4.8 (2.6)	3.3 (2.7)	-1.80 (-3.08, -0.53)	0.01 (-1.41, 1.44)
Relaxed	8.9 (0.8)	7.4 (2.4)	7.3 (2.3)	6.8 (2.6)	8.5 (1.3)	5.7 (2.7)	4.9 (3.0)	5.1 (2.8)	-1.71 (-3.13, -0.29)	8.1 (2.2)	5.5 (2.3)	6.2 (2.3)	4.7 (2.2)	-1.37 (-2.74, -0.01)	0.34 (-1.00, 1.67)
Head Rush	4.2 (2.7)	3.1 (3.2)	3.8 (3.0)	2.4 (1.9)	2.7 (1.5)	2.7 (2.2)	2.1 (2.1)	1.3 (0.6)	-0.12 (-1.05, 0.81)	3.0 (1.7)	2.1 (1.8)	2.2 (2.0)	1.6 (1.1)	-0.50 (-1.45, 0.44)	-0.38 (-1.26, 0.49)
Racing Heart	3.4 (1.8)	2.7 (2.5)	3.5 (3.0)	2.3 (1.8)	2.7 (1.8)	2.1 (1.5)	1.9 (1.7)	1.4 (1.1)	-0.75 (-1.82, 0.31)	2.5 (1.8)	1.8 (1.5)	2.2 (2.3)	1.9 (1.6)	-0.60 (-1.73, 0.52)	0.15 (-0.57, 0.87)
Tremor	2.1 (2.2)	1.6 (1.8)	2.1 (1.9)	1.6 (1.2)	1.6 (1.5)	1.2 (0.4)	1.5 (1.1)	1.0 (0.0)	-0.32 (-0.77, 0.13)	1.2 (0.6)	1.1 (0.3)	1.7 (2.2)	1.5 (1.4)	0.05 (-0.44, 0.53)	0.37 (-0.12, 0.85)
Light Headed	2.1 (1.5)	1.9 (1.6)	2.9 (2.5)	2.1 (2.0)	2.2 (1.6)	1.7 (1.4)	1.6 (1.1)	1.2 (0.4)	-0.77 (-1.77, 0.24)	1.6 (0.8)	1.8 (1.4)	2.3 (1.8)	1.3 (1.2)	-0.44 (-1.47, 0.59)	0.33 (-0.39, 1.04)
Drowsy	1.4 (1.3)	1.4 (1.1)	2.1 (2.1)	1.3 (0.7)	1.4 (0.6)	1.3 (0.6)	1.4 (0.9)	1.0 (0.0)	-0.57 (-1.01, -0.13)	1.3 (0.6)	1.3 (0.8)	1.7 (1.3)	1.4 (1.4)	-0.31 (-0.81, 0.18)	0.25 (-0.17, 0.68)
Energetic	4.1 (2.3)	3.8 (3.1)	4.5 (2.7)	3.9 (2.6)	3.4 (2.4)	3.4 (2.3)	3.0 (2.2)	2.8 (2.9)	-0.75 (-2.03, 0.53)	3.7 (1.9)	2.3 (2.1)	3.3 (2.0)	2.6 (1.7)	-1.23 (-2.35, -0.11)	-0.48 (-1.62, 0.66)
Jittery	2.4 (2.0)	1.7 (1.7)	3.1 (2.9)	1.6 (1.4)	1.5 (1.3)	1.5 (0.8)	1.4 (1.6)	1.2 (0.4)	-0.27 (-0.77, 0.23)	1.4 (0.7)	1.3 (0.8)	1.8 (1.7)	1.4 (1.0)	-0.07 (-0.65, 0.50)	0.20 (-0.23, 0.63)
High	1.8 (1.5)	1.4 (0.8)	2.3 (2.2)	1.5 (1.0)	1.8 (1.3)	1.7 (1.4)	1.4 (1.1)	1.2 (0.8)	-0.22 (-1.00, 0.57)	1.1 (0.5)	1.1 (0.5)	1.5 (1.5)	1.4 (1.2)	-0.19 (-0.93, 0.55)	0.03 (-0.52, 0.58)

Discussion

Our study found that when menthol cigarettes were no longer available, smokers of these products switched to non-menthol cigarettes. The use of e-cigarettes was modest; however, if e-cigarettes with menthol-flavored e-liquid were available, most participants selected them and more tried using them relative to when only tobacco flavored e-liquid was available. In the context of menthol cigarettes not being available, the number of cigarettes smoked was lower when menthol flavored e-liquid was available than when only tobacco flavored e-liquid was available.

Studies evaluating the potential effects of e-cigarette flavor bans (including menthol) on combustible cigarette use among current smokers have found indications that such bans could result in less substitution for e-cigarettes and higher continued combustible cigarette use. For example, a study using a discrete choice experiment found that a greater proportion of adult smokers and recent quitters chose combustible cigarettes when flavors were banned in both combustible cigarettes and e-cigarettes compared to when flavors were only banned in combustible cigarettes.¹⁵ A study using a behavioral economics approach found that as the price of menthol cigarettes increased, most menthol smokers substituted with non-menthol cigarettes; however, a large proportion substituted with menthol-flavored e-cigarettes regardless of the availability of other menthol-flavored combustible products.¹⁴ A study in which smokers not interested in quitting were asked to substitute e-cigarettes for a 6-week period found that smoking decreased to a greater extent in those who used menthol e-cigarettes relative to other flavors.³² Surveys or focus groups suggest that smokers believe that the availability of flavored e-cigarettes aids in the reduction or cessation of combustible product use and that if flavored e-cigarettes were banned, users of these flavors would decrease e-cigarette use and increase combustible cigarette use.^33,34 Our study expands on previous research by evaluating actual tobacco product use during a simulated menthol ban and specifically assesses the effect of a ban on menthol in combustible cigarettes combined with a ban on all flavors except for menthol in e-cigarettes.

An additional finding in our study was that e-cigarette use was low across groups and decreased over time. Possible explanations for this finding include that e-cigarettes, in general, are of limited appeal to the smokers recruited for this study (ie, in this study participants were on average approximately 45 years of age, at baseline smoked approximately one pack per day, did not regularly use e-cigarettes and were not highly motivated to quit smoking) or the specific e-cigarette products used in this study were of limited appeal. The products selected were among the more commonly used at the time the study was started but it’s possible that different patterns of use would have been observed with products that subsequently became popular (such as JUUL). Future studies are needed to determine if certain subgroups of smokers are more likely than others to switch to e-cigarettes and if newer e-cigarette products lead to greater substitution. Nonetheless, differences in smoking were found between the two menthol ban conditions suggesting that e-cigarette use was sufficient to alter tobacco use patterns with exploratory analysis suggesting that menthol e-cigarette use was associated with decreased smoking. Further research is needed to confirm these observations. Our observed trend toward a higher number of cigarettes smoked per day in the total ban condition compared to the no ban group is somewhat inconsistent with studies finding decreased smoking when menthol smokers switch to non–menthol cigarettes; however, the focus of those studies was not on noncombustible product use.^11–13 Further research is needed to better characterize these findings. The results in this study for subjective measures are generally consistent with other studies reporting lower craving or withdrawal symptom severity and lower ratings in at least some subjective measures after menthol smokers switched to non–menthol cigarettes which could potentially be an effective intermediate step towards smoking cessation although more research is needed in this area.^{11–13,35,36} In contrast to most studies showing moderate to strong support for a menthol cigarette ban, support in the current study was low for a ban on menthol across tobacco product categories (Table 1).^{11,12,37–39} A recent study reported that support for a menthol ban is lowest among users of flavored products and among those not currently intending to quit.⁴⁰ Since this study recruited only menthol cigarette smokers not motivated in quitting at this time, this may explain (at least in part) the low support for a menthol ban.

The primary limitation of this study is the relatively small sample size studied. Although larger studies are needed to confirm any trends observed, differences among groups were found despite the small sample size. Other limitations of this study are that the impact of flavors other than menthol was not considered and recruitment was limited to only those not motivated to quit smoking. Future studies are needed regarding the extent to which these results generalize to smokers more broadly and to assess the impact of bans that either include or exclude other flavorants in e-cigarettes. Additionally, due to the small sample, we did not conduct subgroup analyses to examine if menthol bans differentially impact tobacco use among racial groups. Considering that menthol cigarette use is particularly prevalent among African Americans, such research is critical.⁴¹

Conclusions

This study provides preliminary data that menthol bans that include e-cigarettes may result in different patterns of tobacco use than a menthol ban that only includes combustible cigarettes. Specifically, continuing to allow menthol e-cigarettes to be available in the event that menthol is banned in combustible cigarettes may lead to lower levels of smoking among current menthol smokers than banning menthol in both products. Larger studies are needed to further describe these differences in order to determine which policies are most likely to provide the largest public health benefit.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, is available online at https://academic.oup.com/ntr.

Clinicaltrials.gov # NCT03402243

Funding

This work was supported by the National Institutes of Health’s National Institute on Drug Abuse (grant R03DA045150), by the National Center for Advancing Translational Sciences (grant UL1TR002494), and by the National Institutes of Health’s National Cancer Institute (grant P30CA077598) utilizing the Biostatistics and Bioinformatics Core shared resource of the Masonic Cancer Center at University of Minnesota

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Declaration of Interests

None

Data availability

De-identified data may be made available to other investigators upon request. Proposed uses of requested data from qualified investigators will be evaluated taking into account ethical considerations, scientific integrity, and compliance with institutional guidelines