Article Text

Download PDFPDF

Effect of flavour manipulation on ENDS (JUUL) users’ experiences, puffing behaviour and nicotine exposure among US college students
  1. Mayra Vargas-Rivera1,
  2. Mohammad Ebrahimi Kalan1,
  3. Melissa Ward-Peterson2,
  4. Olatokunbo Osibogun1,
  5. Wei Li1,
  6. David Brown3,
  7. Thomas Eissenberg4,
  8. Wasim Maziak1
  1. 1 Epidemiology, Robert Stempel College of Public Health, Florida International University, Miami, Florida, USA
  2. 2 Community Based Research Institute, Robert Stempel College of Public Health, Florida International University, Miami, Florida, USA
  3. 3 Family and Community Medicine, Florida International University College of Medicine, Miami, Florida, USA
  4. 4 Psychology and Center for the Study of Tobacco Products, Virginia Commonwealth University, Richmond, Virginia, USA
  1. Correspondence to Dr Wasim Maziak, Epidemiology, Robert Stempel College of Public Health, Florida International University, Miami, FL 33199, USA; wmaziak{at}


Significance Electronic nicotine delivery system (ENDS) use has continued to increase exponentially among young people in the USA, with unique flavours being one of the most cited reasons for use. Yet, controlled studies examining the effects of restricting flavour are lacking. This study evaluates the impact of ENDS flavour manipulation on user’s puffing behaviour, subjective experience, harm perception and nicotine exposure among college-aged ENDS users.

Methods JUUL users (n=30, age 18 to 24 years) attended two 60 min ad libitum ENDS use sessions (JUUL preferred flavour vs JUUL classic tobacco flavour) in a cross-over design. Puff topography and plasma nicotine concentration were measured, and participants completed subjective experience questionnaires.

Results Increases were observed on measures of satisfaction, taste, enjoyment, urges to vape/smoke, pleasure, product appeal and increased concentration following using the preferred flavour pod (p values <0.05). Compared with preferred flavour, participants in the tobacco flavour were less motivated to use it in the future (70.9 vs 19.1 scores, p<0.001), even if it was the only product on the market (75.8 vs 30.7 scores, p<0.001). While nicotine levels significantly increased in both conditions from pre to post session (p values <0.001), no significant differences were observed in nicotine boost levels or on puff topography parameters when comparing both flavour conditions.

Conclusions This pilot study provides evidence that ENDS flavours have a substantial effect in enhancing young current ENDS users’ experiences, product appeal and motivation to use the product in the future. It highlights that limiting flavours could play a potential role when designing strategic policies to reduce the appeal of ENDS use among young people.

  • electronic nicotine delivery devices
  • nicotine
  • non-cigarette tobacco products
  • smoking topography
  • harm reduction

Data availability statement

Data are available upon reasonable request. De-identified data is available upon reasonable request to: Dr Wasim Maziak, MD, PhD, Professor and Chair, Department of Epidemiology, Robert Stempel College of Public Health and Social Work, Florida International University, Miami, FL, Phone: 305-348-4501, Email:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The use of electronic nicotine delivery systems (ENDS) has continued to increase dramatically among youth and young adults in the USA.1–5 Reports from the 2019 National Youth Tobacco Survey (NYTS) show an increasing trend in high school students’ current (past 30-day) use of ENDS to 27.5% in 2019, compared with 20.8% in 2018 and 11.7% in 2017.1 Current patterns of ENDS use among young adults are also alarming, with results from the Monitoring the Future (MTF) national survey revealing that the past 30-day use of ENDS among college students more than doubled between 2017 and 2018 (from 6.1% to 15.5%), and also increased from 6.5% to 10.6% for all young adults.5 The increase in ENDS use among college students was categorised among the ‘greatest 1 year increase we have seen for any substance since MTF began over 40 years ago’.5 These epidemiological trends are worrisome, given the emerging evidence about ENDS potential to harm and addict young users.6–13

ENDS features such as flavour, are important drivers of ENDS uptake and use among young people.3 14–20 Most young ENDS users use flavoured products,1–3 19–21 and flavour is one of the most commonly cited reason for ENDS use among young people.14 19–21 The 2019 NYTS found that 97% of ENDS users among youth reported using a flavoured product in the last month, and 70% cited flavours as the reason for their use.3 22 In the USA, the majority of young ENDS users cited the use of popular fruit and menthol/mint as their preferred ENDS flavours.19 A 2017 market study of English-language websites found an astonishing 15 586 unique ENDS liquid flavours available online,23 with many flavours such as cotton candy, gummy bear, unicorn and graham cracker, that appear to target young people.24 Because so many young ENDS users use flavours, and because so many flavours appear marketed to them, restricting or eliminating ENDS flavours has been identified as a promising policy strategy for curbing the ENDS epidemic among young people in the USA.14 16 25

To build a robust foundation for such a strategy clear evidence about the effectiveness of flavour limitation on youth ENDS use is needed. To date, few studies have explored ENDS flavour preference among young people, and most have been qualitative, cross-sectional or observational studies.19–21 26–29 Of the couple studies that used an experimental design to assess flavour manipulation on users' experience and use pattern, none was designed to answer the regulatory-specific question of what is likely to happen if instead of preferred flavour, there was only a tobacco-flavoured ENDS option.30 31 This study tries to answer an important regulatory question about the potential short-term effects of ENDS flavours to tobacco in terms of curbing use among young people who show as a group a predominant preference to flavoured products.


Participants, design and procedures

Given the wide variability, designs and generations of ENDS, we opted to recruit pod-based JUUL ENDS users in our study to standardise lab study parameters, and be able to isolate the effect of flavour manipulation. JUUL—a small, easy to charge USB-shaped device—is currently the most popular brand of ENDS in the USA, especially among young people.32–34 At the time the study was conducted, its disposable e-liquid pods were available in a variety of flavours such as fruit, mint and cucumber.35 At the time of this study, JUUL pods were available with two nicotine concentrations by weight of 5% (about 40 mg per pod, based on 59 mg/mL) and 3% (about 23 mg per pod, based on 35 mg/mL) in the USA.36 37 In addition, JUUL e-liquid was originally designed with a new formula containing nicotine salts instead of free-base nicotine, which results in a faster delivery and absorption of nicotine with less irritation than other ENDS products.37 38

We compared participants’ responses to their own preferred JUUL pod flavour to the classic tobacco-flavoured JUUL pod. Participants were 32 healthy young adults between the ages 18 to 24 years who self-reported as current JUUL ENDS users (regularly used at least once per month for 6 months or more) recruited from the Miami metropolitan area via flyers, word of mouth and online advertisements (eg, Craigslist).

Participants provided informed consent and attended the lab for two sessions that differed by flavour condition (preferred flavour vs tobacco-flavoured control). Session order was counterbalanced. Because all participants used 5% nicotine JUUL pods, that concentration was kept constant in the two conditions. Sessions were separated by a 48-hour washout period to avoid carryover effects and preceded by ≥24-hour tobacco/nicotine abstinence validated by pre-session plasma nicotine levels. Participants provided an exhaled breath sample that was analysed immediately for carbon monoxide concentration (Vitalograph, Lenexa) and this measure was used for biochemical verification of combustible tobacco product use status. The assigned flavour pod was weighed prior and following the end of the session, in which participants used their session-designated pod flavour ad libitum for up to 60 min. Blood samples were drawn immediately prior to the session, and immediately after the end of the ad libitum use period. Subjective responses were assessed before and after the ad libitum period, while puff topography and cardiovascular indices were monitored continuously during use. During each session, participants were seated in a comfortable recliner chair in a private room and were given the option to watch movies of their choice or use their mobile devices. Given that the study involves the use of a potentially harmful product, exclusion criteria included self-reported history of chronic health problems (eg, asthma, high or low blood pressure, diabetes) or psychiatric conditions, as well as regular use of prescription medications (other than vitamins or birth control). Additionally, similar to other tobacco studies,39 40 individuals that self-reported current use of >5 cigarettes/month or other tobacco/nicotine products (>5 times in the past month) were also excluded since concurrent regular use of other tobacco/nicotine products would bias measurement of ENDS associated outcomes. Females were excluded if they were pregnant (verified by urinalysis), or breastfeeding. Participants were compensated a total of US$150 for their time, effort and expense of attending study sessions.


Plasma nicotine

Blood was drawn 5 min before and immediately after the ENDS use period at each laboratory visit. Blood samples were centrifuged, plasma stored at −80°C and analysed for nicotine concentration. Plasma nicotine was quantified using the gas chromatography-mass spectrometry method.41 The assay’s limit of detection was 0.2 ng/mL and limit of quantitation was 1.0 ng/mL.

ENDS puff topography

ENDS use puff topography was assessed using a topography instrument that was adapted to ENDS, following the validated waterpipe/cigarette topography device designed and tested by the same pioneering research team in the American University at Beirut.42 Its software converted pressure signals to air flow (mL/sec) data, producing measures of total ENDS use time, total puff time, puff duration, inter-puff interval, number of puffs, total volume inhaled and average puff volume.

Subjective measures

Subjective ENDS use experience related to items such as pleasantness, enjoyment, pleasure from use and interest in continued use was assessed using a visual analogue scale (VAS) adapted from Leavens and colleagues,43 while subjective puff liking was measured using the Duke Sensory Questionnaire (DSQ).44 45 VAS scores ranged from 0 (‘not at all’) to 100 (‘extremely’), while the DSQ ranged from 1 (‘not at all’) to 7 (‘extremely’). Harm perception was assessed by using items adapted from Popova and Ling46 that ranged from a score of 1 (‘not at all harmful’) to 7 (‘extremely harmful’), while intention to quit and motivation to vape were measured using items that ranged within a score of 1 (‘not at all likely’) to 100 (‘extremely likely’), adapted from work by Jaber and colleagues.47

Data preparation and analysis

Plasma nicotine concentration values that were below the limit of quantitation (LOQ) were replaced with the LOQ (1 ng/mL) because this approach is more conservative than assuming that values below the LOQ were zero.48 49 Means for pre-session and post-session plasma nicotine were summarised with means and SEM. In order to account for nicotine change over time, nicotine boost was calculated by subtracting the pre-session measurement from the post-session measurement. Means, SD and percentages were used to describe the baseline characteristics of the study sample. Means and SD were calculated and compared by ENDS flavour condition (preferred and non-preferred flavour) for all subjective measures, exhaled carbon monoxide and topography. Planned comparisons using two-tailed paired samples t-tests were used to examine differences between means for which p value <0.05 are considered significant. In addition, to test for significant differences of ENDS flavour condition on plasma nicotine levels, analysis of variance with two within-subject factors: ENDS flavour condition (preferred and non-preferred) and time (pre-session and post-session) was conducted. All analyses were performed in IBM SPSS V.25.50


A total of 32 participants originally completed both study sessions. However, two participants whose results showed they were not abstinent from tobacco use (pre-session plasma nicotine value >10 ng/mL) were removed from the overall analysis. Thus, a total of 30 participants were eligible to be included in the analysis. The mean age for ever trying a nicotine/tobacco product was 16.5±1.5 years, while the mean age for first trying any ENDS product was 18.4±2.3 years. For the complete details of participant baseline characteristics, see table 1.

Table 1

Baseline characteristics for the overall sample (n=30)

Plasma nicotine

The mean pre-session plasma nicotine concentration was 1.8±0.36 ng/mL and 2.02±0.41 ng/mL, respectively, for the preferred condition and the non-preferred flavour conditions. Plasma nicotine increased to 10.85±1.5 ng/mL at post-session in the preferred condition, whereas in the non-preferred flavour condition, it increased to 10.35±1.6 ng/mL. While increases in mean pre-session to post-session plasma nicotine were significant individually in each condition (p<0.001 for each), mean post-session nicotine did not differ significantly between flavour conditions (p=0.68). Though the nicotine boost, or overall change of nicotine level in the plasma over time, was slightly higher in the preferred flavour (9.50±1.5 ng/mL) than in the non-preferred flavour condition (8.86±1.4 ng/mL), this difference was not significant (p=0.60). No significant interaction of ENDS flavour condition by time was observed (F=0.378, p=0.54); thus, changes in nicotine over time did not depend on the flavour condition.

ENDS puff topography

Table 2 presents the statistical analysis summary for puff topography outcomes by flavour conditions. Between the two study conditions (preferred flavour vs tobacco flavour), no significant differences were observed for average smoking time, cumulative puffing time, puff duration, inter-puff interval, number of puffs, total volume, puff volume and maximum puff volume.

Table 2

JUUL use topography measured by flavour condition (n=21)*

Subjective measures

Subjective measures showed significant differences between the two conditions. In the ENDS use experience VAS, participants reported less positive experiences when using the tobacco-flavoured product compared with their own flavour (eg, led to a reduction in pleasure, willingness to use in the future, liking and satisfaction or urge to use; p values <0.05 for all). Significant differences (p<0.05) between both flavour conditions were also found in the DSQ in four indices: puff liking, puff satisfaction, similarity of product to their own brand and strength of puffs on the nose. For these items, all scores were greater after vaping the preferred flavour, except for the strength of the puffs in the nose, which was higher in the non-preferred flavour condition. Regarding harm perception, there were no significant differences between both ENDS flavour conditions for any of the three items (p>0.05 for all). However, scores for the preferred flavour were significantly higher in the motivation to vape items than in the non-preferred flavour condition (p<0.05). Figure 1, table 3 and figure 2 illustrate the mean post-session ratings for the subjective ENDS use experience VAS, DSQ and motivation to vape responses for the two flavour conditions, respectively.

Figure 1

Means for post-session subjective responses for the ENDS use experience visual analogue scale by flavor condition (n=30).

Figure 2

Means for post-session subjective responses for ENDS motivation to vape responses by flavor condition (n=30).

Table 3

Means for post-session subjective responses for the Duke Sensory Questionnaire (n=30)


This is the first cross-over laboratory study to explore the effects of ENDS flavour manipulation on current users’ subjective experiences, nicotine levels and puffing behaviour. It shows that flavour is likely to have an important role in ENDS users’ satisfaction, use pleasure and product appeal. In addition, limiting flavour to the tobacco variant decreased intention to use in the future, even it was the only product on the market. Flavour, however, did not seem to play a significant role in shaping user’s harm perception, puffing behaviour and exposure to nicotine. Together, our results reveal that limiting characteristic flavours other than tobacco may be a potentially effective strategy to reduce ENDS use among young people, since the use of tobacco flavour was associated with a reduction of the appeal, subjective experiences and positive sensory perceptions of our college-aged participants.

Based on our study of young JUUL users, our results indicate that moving from preferred to tobacco-flavoured products was associated with a reduction in users’ positive experiences (eg, satisfaction, taste, puff liking and motivation to use the product in the future). Other studies have documented the importance of flavour in attracting youth to use ENDS and sustaining ENDS use, mostly applying qualitative study designs.3 14–21 26–28 The strength of our data, however, is that they are not based on users’ opinion, but their actual responses to flavour manipulation in an experimental, tightly controlled, study design. Few other studies used an experimental design to assess the effect of flavour manipulation on user’s experiences,26 31 but the conditions were not tailored to specifically address the regulatory context of removing flavour. One experimental study addressed the issue of flavour manipulation effect on users across three conditions (assigned flavour vs preferred vs tobacco), and found different puffing patterns of use between the three conditions.31 However, this was a limited sample and assessment study, where the puffing patterns of 14 inpatient subjects who used open system ENDS were assessed visually.

Our study is the first laboratory study to provide puff topographical data from JUUL users, and our hypothesis behind such measurement was that the tobacco-flavour condition will be associated with a reduction in puffing patterns compared with preferred flavour. However, our analysis revealed no differences in puffing behaviour according to condition. Interestingly, we have also not found a difference in the nicotine boost between the two conditions, which corroborates the puffing data. Apparently, as we have seen with other tobacco use methods such as hookah, nicotine dependence rather than flavour is likely to be the main determinant of puffing behaviour and nicotine exposure among users.40 As for harm perception, our results suggest that flavour is not likely to have substantial influence, taking into consideration that even the JUUL tobacco flavour is likely to have been designed to give sweet and smooth ‘vape’.51 These qualities have been shown to increase the appeal of the product by masking the harsh taste of tobacco, and ameliorating harm perception by users.27 51 52

While our study provides evidence into the influence of flavour on ENDS use experience and motivation, some limitations need to be considered. First, our study involved a sample of 30 mostly Hispanic college students in the Miami-Dade, Florida, area. It is possible that we were unable to detect significant differences between flavours on puff topography, nicotine boost and harm perception due to decreased statistical power, or because particular characteristics of the JUUL device (ie, USB design, nicotine salt formula) may be different from other ENDS. Future study designs with ample power will be needed to confirm or clarify these results. However, we were able to detect significant differences and important trends that future research and policy initiatives can build on. We also could not blind participants to the study conditions, but they were blinded to the specific hypotheses to be tested, and the use sessions’ order was counterbalanced to control for carry-over effect. While puff topography measurements were lost for nine participants at random due to Microsoft Windows/eTop topography software compatibility issues, these have now been resolved. Still, our analyses yielded similar results even after all measures except topography were analysed for the entire 30 participant study sample. Inevitably, the laboratory environment differs from participant’s actual use environment. Nonetheless, we tried to minimise this effect and facilitate a natural use experience by offering participants with a comfortable private space, with access to movies of their choice, in order to facilitate a natural use experience. Finally, this study is not about comparing preferred flavour to tobacco flavour because this comparison is biassed to begin with since participants who have their preferred, non-tobacco, flavour have already declared their preference. Rather, this study tries to answer an important regulatory question that is being discussed widely about the potential effects of limiting ENDS flavours in terms of curbing use among young people.

Notwithstanding these limitations our study provides empirical evidence that limiting flavour may be a potentially effective strategy to limit ENDS use among young people, at least in the short-term. Studies of flavour limitation with prospective follow-up and extended outcomes are needed to gauge the full potential of flavour-limiting regulatory action on young people ENDS use. Such evidence eventually will be instrumental to support such regulatory action in the face of expected opposition.

What this paper adds

  • Switching from preferred to tobacco-flavoured ENDS is strongly associated with a reduction in users’ subjective positive experiences.

  • This study demonstrates that ENDS flavour is likely to have an important role in current ENDS users’ satisfaction, use pleasure and product appeal.

  • Highlights the potential role of limiting flavour as a strategy to reduce the appeal of ENDS and curb ENDS use among young people.

Data availability statement

Data are available upon reasonable request. De-identified data is available upon reasonable request to: Dr Wasim Maziak, MD, PhD, Professor and Chair, Department of Epidemiology, Robert Stempel College of Public Health and Social Work, Florida International University, Miami, FL, Phone: 305-348-4501, Email:

Ethics statements

Ethics approval

This study was approved by Florida International University Institutional Review Board.


We gratefully acknowledge the help from Melanie Engels in ascertaining the raw data for this study, from Stephan Mertens in discussing the manuscript and from Sandra Goldbeck-Wood in editing the manuscript.



  • Correction notice This paper has been updated since first published to update 'Competing interests' statement and correct first author's name to 'Mayra Vargas-Rivera'.

  • Contributors WM, MW-P and MVR contributed to the conception and design of the study, supervised data collection, contributed to interpretation of the data and critically revised the manuscript for important intellectual content. MVR, MEK and WL collected the data. MVR performed the initial analyses and drafted the initial manuscript. TE, DB, OO, MEK and WL contributed to interpretation of the data and critically revised the manuscript for important intellectual content. This manuscript was written by MVR with input from all co-authors who read and approved the final version.

  • Funding This project was funded intramurally by the Florida International University Office of Research and Economic Development (FIU-RCMI) Pilot Study Program. MV is supported by the National Institute of Health Fogarty International Center (NIH FIC) under award R01TW010654. MEK is supported by the National Institute on Drug Abuse at National Institutes of Health (NIDA NIH) under award R01DA042477. MW-P is currently supported by the National Institute of Minority Health and Health Disparities under award U54MD012393 for FIU-RCMI. OO is supported by the NIDA NIH under award T32DA043449. TE is supported by the NIDA NIH under award number U54DA036105 and the Center for Tobacco Products of the US Food and Drug Administration. WM is supported by the NIH FIC under award R01TW010654 and the NIDA NIH under award R01DA042477.

  • Disclaimer The content is solely the responsibility of the authors and does not necessarily represent the views of the NIH or the FDA.

  • Competing interests TE is a paid consultant in litigation against the tobacco industry and also the electronic cigarette industry and is named on a patent for a device that measures the puffing behaviour of electronic cigarette users.

  • Provenance and peer review Not commissioned; externally peer reviewed.