Constraining Electronic Nicotine Delivery Systems (ENDS) Nicotine Dose by Controlling Nicotine Flux at a Limited Puff Duration

Abstract

Nicotine flux, the rate of electronic nicotine delivery system (ENDS) nicotine emission, is important in determining ENDS abuse liability. However, flux does not account for user behavior, including puff duration. Along with nicotine flux, puff duration limits the dose of nicotine that can be inhaled. Controlling both flux and puff duration allows regulators to constrain nicotine dose effectively. This study examined the effects of differing levels of ENDS nicotine flux (by manipulating liquid nicotine concentration and holding device power constant), with user puff duration limited to two seconds. Participants (N=32) completed four sessions, each session differing by nicotine flux (no flux, low flux, cigarette-like flux, and high flux conditions). Participants completed two ENDS use bouts in each session, while puff duration was limited to two seconds. Plasma nicotine concentration, heart rate, and subjective effects were measured. At higher flux, higher plasma nicotine concentration and higher heart rate were observed. Moreover, higher levels of flux decreased ratings of craving and urge to use nicotine and increased positive subjective effects, such as calmness. This study demonstrates that by manipulating nicotine flux and limiting puff duration, nicotine dose can be controlled. Subsequent research should demonstrate the effects of manipulating puff duration systematically. Results underscore the importance of targeting both flux and puff duration for ENDS regulation, intended to reduce abuse liability while maintaining the potential to facilitate transitions from cigarettes to ENDS.

Introduction

Since their introduction to the United States (US) market in the 2000s, electronic nicotine delivery systems (ENDS; Bandi et al., 2023; Brouwer et al., 2023; Romeh et al., 2023; Selya et al., 2023), a diverse class of products intended to deliver aerosolized nicotine to people who use them, have experienced considerable uptake. Between 2016 and 2019, ENDS sales increased roughly 300% in the US (Ali et al., 2020). Some people who use ENDS consider them to be less harmful than combustible tobacco cigarettes (Huang et al., 2019; Li et al., 2022; Nyman et al., 2019), although ENDS deliver nicotine and other toxicants (Goniewicz et al., 2018). Effective regulation of ENDS devices and liquids can help address possible harms from ENDS use, limit the abuse of these products (i.e., use by otherwise nicotine-naive individuals), and ultimately protect public health.

Regulating ENDS is challenging because ENDS products vary substantially in terms of their liquid nicotine concentration, electrical power output, and other characteristics, like system type (open versus closed) and liquid solvent concentration. These characteristics are relevant to ENDS effects, especially nicotine delivery. ENDS liquid nicotine concentration is related to plasma nicotine concentration in people who use ENDS (Dawkins et al., 2016; Hiler et al., 2017). ENDS liquid nicotine concentration is often considered an important regulatory target. For instance, since 2014, the European Union (EU) has limited ENDS liquid nicotine concentration to ≤20 mg/mL with the stated intent of ensuring that ENDS nicotine delivery is comparable to that of combustible cigarettes (European Commission, 2014). However, liquid nicotine concentration regulations like the EU’s do not account for other characteristics of ENDS that influence nicotine delivery, such as device power and user puff duration, both of which can be manipulated by ENDS users or ENDS manufacturers.

For example, many refillable ENDS provide the user with the ability to control device power. When liquid nicotine concentration is held constant, higher device power leads to greater nicotine delivery (e.g., Eversole et al., 2023; Hiler et al., 2020). Indeed, people who use ENDS exclusively worked harder to obtain puffs in a low nicotine/high power condition relative to a high nicotine/low power condition (Hoetger et al., 2022). Taken together, results suggest that the ability to manipulate device power may undermine regulatory limitations of liquid nicotine concentration. Lower nicotine concentration liquids can deliver substantial amounts of nicotine to the blood of people using ENDS when paired with a high power device (e.g., Wagener et al., 2017). For this reason, regulators may want to target nicotine emission rate (i.e., nicotine “flux” expressed in μg/sec; Shihadeh & Eissenberg, 2015; Eissenberg & Shihadeh, 2015) in order to control nicotine delivery rather than liquid nicotine concentration alone. An advantage of nicotine flux as a regulatory target is that it simultaneously accounts for a variety of relevant device and liquid characteristics, including device power and liquid nicotine concentration (Eissenberg et al., 2021). Another advantage is that nicotine flux can be predicted mathematically (Talih et al., 2017), allowing regulators to know the flux of any ENDS, based upon its design, operating characteristics, and liquid constituents, including nicotine concentration (Talih et al., 2017).

While nicotine flux accounts for the various device and liquid characteristics, it cannot account for user behavior. One key behavioral variable relevant to ENDS nicotine delivery to the blood is puff duration. When other factors are controlled, longer puff duration is associated with increased plasma nicotine concentration (Farsalinos et al., 2015; Hiler et al., 2017; Eversole et al., 2022). Puff duration can vary widely; people who are experienced with ENDS take longer puffs than people who are ENDS-naïve (Farsalinos et al., 2015; Spindle et al., 2017). Even interindividual puff duration varies (Kosmider et al., 2018). From a regulator’s perspective, this variability in puff duration may be problematic. Even when ENDS nicotine flux is controlled in a way that would seem to limit nicotine delivery effectively, people who take longer puffs can increase the amount of nicotine they receive. Indeed, nicotine “dose” – the amount of nicotine delivered to the user after a single puff – is a function of both nicotine flux and puff duration. For example, if an ENDS device/liquid combination has a nicotine flux of 50 μg/sec, and an individual takes a 5 sec puff from that ENDS, that puff delivered 250 μg nicotine. Thus, for regulators interested in controlling the dose of nicotine delivered to people using ENDS (i.e., to limit abuse liability; Carter et al., 2009; Henningfield et al., 2011), both nicotine flux and puff duration must be controlled.

Controlling puff duration may seem difficult, but it already has been incorporated into several ENDS devices. JUUL ENDS limits puff duration to 5.9 sec in devices available in the US and the UK (Talih et al., 2020). Moreover, the Vuse Solo ENDS available in the US limits puff duration to 3.7 seconds, while the Vuse Alto ENDS available in the US limits puff duration to 4.9 sec (Talih et al., 2023). Thus, the technology exists to regulate ENDS nicotine dose on a per puff basis. However, while previous research has focused on the effects of different nicotine concentrations and different power settings, none (to our knowledge) have focused on different nicotine fluxes with limits on puff duration. Better understanding how varying nicotine flux, along with a limited puff duration, relates to outcomes such as plasma nicotine concentration, heart rate, and subjective effects could help scientists and policymakers determine if these manipulations could inform regulatory changes. The purpose of this study is to examine the effects of varying nicotine fluxes while user puff duration is limited to less than two seconds. Outcome measures include plasma nicotine concentration and a variety of subjective effect measures.

Method

We report how we determined our sample size, all data exclusions, all manipulations, and all measures in the study. Data were analyzed using SPSS, version 28. This study’s design and its analysis were pre-registered on www.clinicaltrials.gov (NCT04378907).

Participants

Individuals were eligible to participate if they were healthy (based on self-report), 18 to 55 years of age, self-reported daily use of cigarettes, ENDS, or both products, with daily use of at least one product. All of these groups of individuals were recruited as polytobacco use has become more common and discerning a specific group for an individual has become more challenging. In addition, because ENDS use is so common among people who smoke cigarettes, we felt that including people who primarily smoke cigarettes was relevant. Individuals were excluded if, at screening, they weighed less than 110 pounds, self-reported a current, diagnosed chronic illness or psychiatric condition, current use of a psychotropic medication, alcohol use more than 25 days out of the past 30 days, cannabis use more than 15 days out of the past 30 days, or use of cocaine, opioids, benzodiazepines, methamphetamine, or other illegal substances in the past 30 days. Urine samples were tested for cotinine (Accutest, Jant Pharmacal Corporation) to verify nicotine use for all participants. Females were excluded if they reported currently breastfeeding or if they tested positive for pregnancy by urinalysis at screening. Individuals who reported any intention to quit tobacco/nicotine use in the next 30 days were excluded and referred to cessation treatment. This study was approved by the Institutional Review Board at Virginia Commonwealth University.

Procedures

Participants completed four, approximately 3-hour sessions that were ordered by Latin-square and separated by at least 48 hours. Each session differed by the liquid nicotine concentration, 0, 6, 15 or 30 mg/ml. Participants were blinded to study condition and instructed to abstain from tobacco and nicotine-containing products for ≥12 hours prior to each session. In order to verify overnight abstinence from combustible tobacco products, expired air carbon monoxide (CO; Vitalograh) had to be either less than or equal to half of the CO value at the in-person screening visit. Alternatively, if CO value was 10 ppm or less at screening, it had to be less than or equal to 5 ppm at each session. For participants who primarily used ENDS, where CO is not an indication of ENDS use/abstinence, baseline plasma nicotine concentration was inspected retrospectively using a criterion of ≤5 mg/ml to verify compliance with overnight abstinence (as in Hiler et al., 2017). Plasma analysis occurred after participants completed the study; therefore, a 1-hour observation period before ENDS administration ensured that participants were abstinent for at least one hour before each session (as in Hiler et al., 2020; Spindle et al., 2018).

Experimental sessions included two ENDS-use bouts separated by approximately 20 minutes: a 10-puff, directed use period (with one puff every 30 seconds) followed by an approximately 30-min ad libitum use period. In both use periods, custom-built hardware and software (American University of Beirut; Beirut, Lebanon) or a visual display limited puffs to 2 seconds. The 10-puff directed bout always occurred before the 30-minute ad lib bout. After the two use bouts, participants were allowed to smoke their own brand cigarette or use their own brand ENDS during a 5-minute period, and the latency to puff was measured, in addition to the number of puffs taken.

An intravenous catheter was used to sample approximately 7 ml of blood before and after each ENDS use period, as well as after the own brand use period, for a total of five samples per session (results from the first four blood samples are included here). Subjective measures were administered a total of five times per session, immediately after each blood sampling (results from the first four administrations are included here). Heart rate was recorded throughout each session. Finally, participants were asked to complete a follow-up survey about product use three months after completion of the study sessions (data not included here).

Materials

In each session, participants used a Kanger Sub Box Mini with a 0.5-ohm coil and a glass tank, set to 30 watts, which contained either 0, 6, 15 or 30 mg/ml nicotine-containing liquid. These liquid concentrations, combined with the 30 W setting, created nicotine flux (see Shihadeh & Eissenberg, 2015) conditions that reflected no flux (0 μg/s nicotine), low flux (36.01 μg/s nicotine), cigarette-like flux (90.03 μg/s nicotine), and high flux conditions (180.06 μg/s nicotine). Flux was calculated based on a mathematical model (Talih et al., 2017). All liquids used consisted of 30% propylene glycol and 70% vegetable glycerin with >99% nicotine in protonated form (AVAIL Vapor, Richmond VA). Nicotine content for each liquid was verified independently to be within 2 mg/ml of labeled concentration by VCU’s Bioanalytical Analysis Core Laboratories.

Outcome Measures

Physiological measures

Blood samples were centrifuged and stored at −80°C until analysis of plasma nicotine concentration conducted by VCU’s Bioanalytical Analysis Core Laboratories using a limit of quantitation (LOQ) of 0.1 ng/ml (similar to Eversole et al., 2022). Participants’ heart rates were measured continuously during each session (Criticare Technologies, Inc., North Kingstown, RI).

Puff topography

Puff topography was recorded continuously during each ENDS use bout using mouthpiece-based puff topography equipment, which was developed and manufactured at the American University of Beirut (AUB). Puff topography variables included puff count, duration, and volume, as well as inter-puff interval and average flow rate. This equipment was designed specifically to measure the puff topography of people who use ENDS and has been described previously (e.g., Hiler et al., 2020; Spindle et al., 2018). Mouthpieces were manufactured to fit the device used in this study. Prior to each session, puff topography measurement software was calibrated using an automatic digital flow calibrator. Additional custom-built hardware and software (also developed at AUB) or a visual display limited puffs to 2 seconds.

Subjective effects

Five subjective measures were administered at various times during each study session. Nicotine abstinence symptoms were evaluated using the Hughes-Hatsukami Withdrawal Questionnaire (11 items; Hughes & Hatsukami, 1986). Nicotine-related effects were assessed using the Direct Effects of Nicotine Questionnaire (11 items; Evans et al., 2006), and ENDS-specific effects were assessed using a modified version of the Direct Effects of ENDS Use Questionnaire (10 items; adapted from Foulds et al., 1992; Pickworth, Bunker, & Henningfield, 1994). These three questionnaires (the Hughes-Hatsukami Withdrawal Questionnaire, Direct Effects of Nicotine Questionnaire, and Direct Effects of ENDS Use Questionnaire) were administered via computer using the Visual Analog Scale (VAS) format; for each item, a word or phrase was centered on a horizontal line with “not at all” on the left of the line and “extremely” on the right. Participants clicked on any point of the line and response scores reflected the percentage of total line length measured from the left anchor.

Specific sensations associated with ENDS product use (e.g., harshness) were assessed using the general Labeled Magnitude Scale (gLMS; Rosbrook & Green, 2016) and the Labeled Hedonic Scale (LHS; Lim et al., 2009). The gLMS uses sensation level descriptions at the following numerical values: 0 (“No Sensation At All”), 1 (“Barely Detectable”), 6 (“Weak”), 16 (“Moderate”), 35 (“Strong”), 53 (“Very Strong”), and the highest possible rating of 100 (“Strongest Imaginable Sensation of Any Kind”). Participants were instructed to look at the scale and type the corresponding number in a box. The LHS is similar to the gLMS except that participants were asked to respond to each question with a scale anchored by “most disliked sensation imaginable” and “most liked sensation imaginable.”

The Hughes-Hatsukami Withdrawal Questionnaire and Direct Effects of Nicotine Scale were administered before and after the two ENDS use bouts, as well as after the own brand use bout, for a total of five times each session (results for the first four times are presented here, as these time points were the most relevant to the overall research questions). The Direct Effects of ENDS use, the gLMS, and the LHS were administered after each ENDS use bout, for a total of two times each session.

Amount of liquid consumed

Liquid consumption was measured by weighing the ENDS before and after each ENDS use bout. The difference between ENDS mass before and after use represented the mass of liquid consumed (grams) during each use bout. Total liquid consumed (by volume, in milliliters) during each use bout was computed by dividing the mass consumed by the liquid density.

Sample Size, Data Preparation, and Analysis

Sample size was determined using effect size data from a previous, similar study, which included both people who used ENDS and people who smoked cigarettes. Results revealed that 32 participants were sufficient to achieve power >80% given a Type I error rate of 0.05.

Plasma nicotine concentration values lower than the LOQ (0.1 ng/ml; see Physiological measures) were replaced with 0.1 ng/ml (similar to Eversole et al., 2022). Heart rate data were averaged for 10 minutes before (“baseline”) and throughout each ENDS use period (three values in total). Data for each topography measure were averaged for each ENDS use bout, except for puff number.

All data were analyzed using SPSS Version 28.0. For variables with pre- and post- values for which time was of interest (plasma nicotine, and subjective effects items for the Hughes-Hatsukami and Direct Effects of Nicotine questionnaires), repeated-measures analysis of variance (ANOVA) tests with condition (4 levels) and time (4 levels) were conducted. Tukey’s post hoc tests were used to compare baseline means to subsequent time points as well as all means across conditions at each time point. For these outcomes, interactions of condition by time are the results of most interest and are described below.

For variables that were only administered after each bout (i.e., items in the Direct Effects of Vaping and gLMS questionnaires) or during each bout (i.e., topography variables), separate ANOVAs for each bout by condition (4 levels) were conducted. For the amount of liquid consumed, an ANOVA with one factor (condition, 4 levels) was conducted. In all other respects, these analyses were the same as the analysis described above. For puff duration, in addition to ANOVAs, one-sample t-tests were conducted to determine if puff duration was constrained effectively to no more than 2 seconds.

For plasma nicotine, data were analyzed with and without participants who had baseline concentrations above 5.0 ng/ml (indicating non-abstinence). No differences were observed for the interaction or main effects and thus all participants were included in the final analysis. Several outcome measures had missing data points, and participants’ data were excluded in those cases: for plasma nicotine, one participant had several missing plasma samples; for the Hughes-Hatsukami questionnaire and the Direct Effects of Nicotine questionnaire, one participant was missing one datapoint. For puff duration, puff volume, average flow rate, and puff number, three participants were removed from the analysis due to equipment malfunction. For inter-puff interval, four participants’ data were removed due to equipment malfunction. Final Ns for these measures were as follows: plasma nicotine, Hughes-Hatsukami questionnaire, Direct Effects of Nicotine questionnaires (N = 31); puff duration, puff volume, average flow rate, and puff number (N = 27); inter-puff interval (N = 26).

Data Availability

This study was pre-registered on www.clinicaltrials.gov, NCT04378907. Data and study materials are not yet available publicly but can be made available upon request.

Results

Participant Characteristics

In total, 54 participants provided informed consent and completed the in-person screening visit. Of those, 39 were eligible to participate and 15 were ineligible. Of the 39 participants who were eligible to participate, 32 completed all sessions and are included in the final analysis, with the exceptions noted above.

Participants were people who primarily smoked cigarettes (N = 9), people who primarily used ENDS (N = 20) and people who used both cigarettes and ENDS (N = 3). Overall, participants were on average 29.41 (SD=10.18) years old. For sex assigned at birth, 21 were male and 11 were female. For race, 17 identified as White, 5 as Black or African American, 5 as Asian, 3 as Other, and 2 as multiple races, from a list that included White, Black or African American, Native Hawaiian or Other Pacific Islander, American Indian/Alaskan Native, Asian, and Other. Racial categories were based on National Institute of Health reporting requirements, with the exception of the addition of “Other” by the authors. Most participants (31) were not Hispanic or Latino. For participants who reported daily cigarette use (N = 11), the mean number of cigarettes per day reported smoking was 15 (SD=7.80), mean score on the Fagerström Test for Nicotine Dependence was 4.36 (SD=2.62; Heatherton, Kozlowski, Frecker, & Fagerstrom, 1991), and mean expired air CO concentration at screening was 16.36 (SD=7.80). For participants who reported any cigarette use (N=13), most reported using non-menthol cigarettes (N = 9). For participants who used ENDS in the past 30 days (N = 23), a variety of device types and flavors were reported. For participants who did not report daily cigarette use (N = 21), mean CO at screening was 1.29 (SD=0.46), and mean score on the E-cigarette Dependence Scale was 2.61 (SD=0.62; Morean et al., 2018; range=0–4).

Plasma nicotine concentration

A summary of all main effects and interactions is shown in Tables 1 and 2. A significant condition by time interaction was observed for plasma nicotine. As shown in Figure 1, plasma nicotine concentration increased significantly in the cigarette-like flux condition from a mean baseline value of 1.64 ng/ml (SD=2.34) to 8.8 ng/ml (SD=7.41) after the directed bout and to 8.61 ng/ml (SD=7.47) after the ad lib bout. Plasma nicotine concentration also increased significantly in the high nicotine flux condition from 1.62 ng/ml (SD=2.53) at baseline to 12.42 ng/ml (SD=12.21) after the directed bout, to 7.62 ng/ml (SD=5.75) before the ad lib bout, and to 11.05 ng/ml (SD=8.71) after the ad lib bout. Across conditions, significantly higher nicotine concentrations were observed in the cigarette-like flux condition after the directed bout at 8.80 (SD=7.41) and after ad lib bout at 8.61 ng/ml (SD=7.47) compared to the no flux condition after the directed bout at 1.93 ng/ml (SD=2.80), and after the ad lib bout at 1.62 ng/ml (SD=2.15). Significantly higher nicotine concentrations were observed in the high-flux condition after both bouts, compared to the low flux condition at 4.75 ng/ml (SD=3.44) after the directed bout and 5.39 ng/ml (SD=3.32) after the ad lib bout (Tukey’s HSD; all ps<.05).

Table 1.

Summary of Results for Plasma Nicotine, Heart Rate (HR), Latency to Puff Own Brand, the Hughes-Hatsukami Withdrawal Questionnaire, Direct Effects of Nicotine Questionnaire.

Outcome measure	Condition			Time			Condition by Time
	F	p	Ƞp2	F	p	Ƞp2	F	p	Ƞp2
Plasma nicotinea	26.18	<.001	.47	34.25	<.001	.53	18.64	<.001	.38
Heart Rateb	8.72	<.001	.22	81.46	<.001	.72	18.81	<.001	.38
Latency to puff own brandc	1.16	.326	.04	N/A	N/A	N/A	N/A	N/A	N/A
Hughes-Hatsukamia
Anxious	2.56	.065	.08	12.52	<.001	.30	1.95	.076	.06
Craving	11.09	<.001	.27	45.23	<.001	.60	11.00	<.001	.27
Depression	2.51	.082	.08	3.63	.029	.11	1.98	.067	.06
Difficulty Concentrating	1.30	.280	.04	17.66	<.001	.37	1.96	.048	.06
Drowsy	2.26	.092	.07	2.08	.109	.07	.42	.89	.01
Hunger	1.64	.190	.05	8.76	<.001	.23	.30	.941	.01
Impatient	4.18	.008	.12	11.22	<.001	.27	1.44	.192	.05
Irritable	2.93	.04	.09	20.67	<.001	.41	1.29	.26	.04
Restless	1.91	.135	.06	8.73	<.001	.23	.97	.462	.03
Sweets	.16	.910	.01	4.71	.006	.14	.54	.757	.02
Urge	9.19	<.001	.23	38.75	<.001	.56	7.38	<.001	.20
Direct Effects of Nicotine Questionnairea
Confused	1.12	.330	.04	1.56	.209	.05	.92	.474	.03
Dizzy	6.16	.002	.17	8.62	<.001	.22	7.90	<.001	.21
Headache	.13	.902	.00	1.23	.301	.04	1.00	.431	.03
Heart Pounding	4.46	.008	.13	3.57	.028	.11	4.54	<.001	.13
Lightheaded	3.64	.022	.11	6.58	<.001	.18	6.16	<.001	.17
Outcome measure	Condition			Time			Condition Time
	F	p	Ƞp2	F	p	Ƞp2	F	p	Ƞp2
Nauseous	4.24	.015	.13	4.25	.012	.12	3.09	.007	.09
Nervous	3.68	.025	.11	3.85	.014	.11	1.72	.126	.05
Salivation	.64	.560	.02	2.05	.151	.06	.97	.447	.03
Sweaty	.66	.547	.02	3.02	.039	.09	1.07	.381	.03
Weak	2.34	.105	.07	1.61	.198	.05	2.33	.032	.07

^adf = 3,90; 3,90; 9,270

^bdf = 3, 93; 2, 62; 6, 186

^cdf = 3, 93

Table 2.

Summary of Results for Direct Effects of ENDS Use Questionnaire, General Labeled Magnitude Scale, and Labeled Hedonic Scale.

Outcome measure	Condition
	Directed bout			Ad lib bout
	F	p	Ƞp2	F	p	Ƞp2
Direct Effects of ENDS Use Questionnairea
Awake	7.27	<.001	.19	3.30	.026	.10
Calm	5.40	.002	.15	3.78	.01	.11
Concentrate	1.54	.22	.05	2.31	.08	.07
Dizzy	15.28	<.001	.33	4.41	.006	.12
Pleasant	1.51	.218	.05	1.34	.266	.04
Reduce hunger	4.03	.010	.12	1.74	.168	.05
Right now	20.07	<.001	.39	5.60	.003	.15
Satisfy	2.73	.049	.08	1.14	.336	.04
Sick	7.65	<.001	.20	5.55	.004	.15
Taste good	3.92	.011	.11	4.30	.007	.12
General Labeled Magnitude Scalea
	Directed bout			Ad lib bout
	F	p	Ƞp2	F	p	Ƞp2
Flavor	1.21	.311	.04	1.04	.368	.03
Throat hit	15.85	<.001	.34	15.38	<.001	.33
Harshness	33.93	<.001	.52	8.68	<.001	.22
Labeled Hedonic Scalea
Flavor	2.93	.044	.09	4.94	.003	.14
Harshness	1.53	.213	.05	.85	.470	.03
Throat hit	.46	.663	.02	.31	.764	.01
Warmth	1.37	.262	.04	3.42	.020	.10

^adf = 3, 93

Figure 1.

Mean plasma nicotine concentrations for N = 31 participants who took two second puffs. Filled symbols indicate significant difference from baseline; an asterisk (*) indicates significant difference from no flux; a symbol (a) indicates significant difference from low flux.

Heart Rate

A significant condition by time interaction was observed for heart rate. During the directed bout, heart rate increased significantly in all conditions: from a baseline of 68.80 bpm (SD=10.31) to 72.34 bpm (SD=9.15) in the no flux condition; from 70.56 bpm (SD=10.51) to 78.95 bpm (SD=9.96) in the low flux condition; from 68.42 bpm (SD=7.70) to 82.27 bpm (SD=11.28) in the cigarette-like flux condition; and from 66.25 bpm (SD=7.29) to 83.11 bpm (SD=10.91) in the high flux condition. Compared to baseline, heart rate also increased significantly in the ad lib bouts, to 74.94 bpm (SD=8.74) in the low flux condition; to 75.05 bpm (SD=9.87) in the cigarette-like flux condition; and to 74.75 bpm (SD=9.01) in the high flux condition. Compared to the no flux condition, heart rate was higher in the directed and ad lib bouts for the low flux cigarette-like flux, and the high flux conditions. Heart rate at baseline was lower in the high flux condition at 66.25 bpm (SD=7.29) compared to the low flux condition at 70.56 bpm (SD=10.51), and heart rate was higher in the directed bout in the high flux condition compared to the directed bout in the low flux condition.

Latency to Use Own Brand Product

There was no significant effect of condition for this measure. Mean latencies to use own brand product in each session are as follows: in the no flux condition, 7.74 s (SD=4.59); in the low flux condition, 10.18 s (SD=12.27); in the cigarette-like flux condition, 10.51 s (SD=9.57); in the high-flux condition, 11.64 s (SD=12.10).

Hughes-Hatsukami Questionnaire

Means are shown in Supplementary Table 1. Significant interactions of condition by time were observed for the items “Craving a cigarette and/or e-cigarette/nicotine,” “Urges to smoke cigarettes and/or vape e-cigarettes,” and “Difficulty Concentrating”. Means for “Craving a cigarette and/or e-cigarette/nicotine are shown in Figure 2. Mean score significantly decreased after the directed and ad lib bouts in all conditions. Compared to the no flux condition, scores were significantly lower after the directed bout in the cigarette-like flux and high flux conditions and remained lower for these conditions before and after the ad lib bout. In addition, compared to the low flux condition, scores in the cigarette-like flux and high flux conditions were also lower after the directed about, and lower for the high flux condition before and after the ad lib bout.

Figure 2.

Mean cigarette craving ratings for N = 31 participants who took two second puffs. Filled symbols indicate significant difference from baseline; an asterisk (*) indicates significant difference from no flux; a symbol (a) indicates significant difference from low flux.

A similar pattern was observed for the item “Urges to smoke cigarettes and/or vape e-cigarettes,” with significant reductions in scores after the directed bout (and all timepoints thereafter) in the low flux, cigarette-like flux, and high flux conditions. Compared to the no flux condition, scores were lower in the cigarette-like flux and high flux conditions after the directed and ad lib bouts. Compared to the low flux condition, scores were significantly lower in both the cigarette-like flux and high flux conditions, after the directed bout (all ps <.05; Tukey’s HSD).

For the item “Difficulty Concentrating,” significant reductions in scores were observed after the ad lib bout in the low flux condition, as well as before and after the ad lib bout in the cigarette-like flux and high flux conditions. Compared to the no flux condition, after the ad lib bout, scores were lower in the high flux condition (all ps <.05; Tukey’s HSD).

Direct Effects of Nicotine Questionnaire

Means are shown in Supplementary Table 1. Significant interactions of condition by time were observed for the items “Dizzy,” “Heart Pounding,” “Lightheaded,” “Nauseous,” and “Weak.” For the items “Dizzy,” “Heart Pounding,” “Lightheaded,” and “Nauseous,” in the high flux condition, significant increases were observed after the directed bout, compared to baseline. In addition, for all items, mean scores were significantly higher after the directed bout in the high flux condition compared to the no flux and low flux conditions. For the items “Dizzy” and “Heart pounding,” scores were significantly higher in the high flux condition compared to the cigarette-like flux condition (all ps <.05; Tukey’s HSD).

Direct Effects of ENDS Use Questionnaire

Means are shown in Supplementary Table 2. In the directed bout, a significant main effect of condition was observed for “Awake,” “Calm,” “Dizzy,” “Right now,” “Satisfy,” Sick,” and “Taste good.” For “Awake,” mean scores in the low flux, cigarette-like flux, and high flux conditions were significantly higher than in the no flux condition. For “Calm,” mean scores in the cigarette-like flux and high flux conditions were significantly higher than those in the no flux condition. For “Dizzy,” scores in the cigarette-like flux and high flux conditions were significantly higher than in the no flux and low flux conditions. For “Right now,” scores in the cigarette-like flux and high flux conditions were significantly lower than in the no flux and low flux conditions. Also, mean score in the high flux condition was significantly lower than the mean score in the cigarette-like flux condition. For “Sick,” the mean in the high flux condition was significantly higher than in the no flux and low flux conditions. For “Taste good,” mean score in the high flux condition was significantly lower than in the low flux condition. For “Satisfy,” post-hoc tests did not reveal any significant differences across means.

In the ad lib bout, a significant main effect of condition was observed for “Awake,” “Calm,” “Dizzy,” “Right now,” Sick,” and “Taste good.” For “Awake,” mean score in the high flux condition was significantly higher than in the no flux condition. For “Calm,” mean score in the cigarette-like flux condition was higher than in the no flux condition. For “Dizzy,” mean score was significantly higher in the high flux condition than in the no flux condition. For “Right now,” mean scores in the cigarette-like and high flux conditions were significantly lower than in the no flux condition. For “Sick,” mean score in the high flux condition was significantly higher than in the no flux and low flux conditions. For “Taste good,” mean score in the high flux condition was significantly lower than in the no flux and low flux conditions.

General Labeled Magnitude Scale

Means are shown in Supplementary Table 2. In the directed bout, a significant main effect of condition was observed for “Throat hit” and “Harshness.” For “Throat hit,” mean scores in the cigarette-like flux and high flux conditions were significantly higher than in the no flux condition. Also, mean score in the high flux condition was significantly higher than in the low flux condition. For “Harshness,” mean scores in the low flux, cigarette-like flux, and high flux conditions were significantly higher than in the no flux condition. Also, mean score in the high flux condition was significantly higher than in the low flux condition and in the cigarette-like flux condition.

In the ad lib bout, a significant main effect of condition was observed for “Throat hit” and “Harshness.” For “Throat hit,” mean scores in the cigarette-like flux and high flux conditions were significantly higher than in the no flux condition. Also, mean score in the high flux condition was significantly higher than in the low flux condition. For “Harshness,” mean score in the high flux condition was significantly higher than in the no flux and low flux conditions.

Labeled Hedonic Scale

Means are shown in Supplementary Table 2. In the directed bout, a significant main effect of condition was observed for “Flavor,” although post-hoc tests did not reveal any significant differences across means. In the ad lib bout, a significant main effect of condition was observed for “Flavor” and for “Warmth.” For “Flavor,” mean score in the high flux condition was significantly lower than in the no flux, low flux, and cigarette-like flux conditions. For “Warmth,” mean score in the high flux condition were significantly lower than in the low flux condition.

Puff topography

Means for all puff topography outcomes are shown in Supplementary Table 3. First, one-sample t-tests were conducted for puff duration, to determine if puff duration was constrained effectively to no more than 2 seconds. Mean puff duration was significantly longer than 2 seconds in the no flux directed bout [t(30) = 3.73] and ad lib bout [t(29)=2.26], low flux directed bout [t(30) = 3.46], and cigarette-like flux directed bout [t(30) = 2.16; all ps < .020], though not longer than 2.31 seconds on average.

In the directed bout, main effects of condition were observed for puff duration and puff volume. Mean puff duration in the cigarette-like flux and high flux conditions was significantly lower than mean puff duration in the no flux and low flux conditions. In addition, mean puff duration in the high flux condition was significantly lower than in the cigarette-like flux condition. Mean puff volume was significantly lower in the high flux condition than in the no flux and low flux conditions.

In the ad lib bout, main effects of condition were observed for puff duration, puff volume, inter-puff interval, puff number, and average flow rate. Mean puff duration in the high flux condition was significantly lower than in the no flux condition. Mean puff volume was significantly lower in the high flux condition than in the no flux condition. Mean inter-puff interval was significantly higher in the cigarette-like flux and high flux conditions compared to the no flux condition. Mean puff number was significantly lower in the cigarette-like flux and high flux conditions, compared to the no flux and low flux conditions. Post hoc tests did not reveal any significant differences across conditions for average flow rate.

Puffs in Own Product Use Period

A significant effect of condition was observed for the number of puffs taken in the own product use period. Post hoc tests revealed that participants took fewer puffs from their own brand product in the cigarette-like flux condition, at 7.44 puffs (SD=4.59), compared to 8.69 puffs (SD=4.98) in the no flux condition, 8.84 puffs (SD=5.0) in the low flux condition, and 8.06 puffs (SD=5.0) in the high flux condition (all ps <.05; Tukey’s HSD).

Amount of liquid consumed

Means are shown in Supplementary Table 3. A significant effect of condition was observed for the amount of liquid consumed in the ENDS ad lib use period. While means appeared to show less liquid consumed as nicotine concentration increased, no significant differences between means were observed.

Discussion

The present study examined the effects of varied nicotine flux conditions (via varying nicotine concentration and holding power constant) on plasma nicotine concentration and subjective measures of ENDS use when puff duration was limited to two seconds. Results show that overall, with the combination of known nicotine flux and limited puff duration (and thus dose), higher flux led to higher plasma nicotine concentration. Significantly higher mean nicotine plasma concentration was observed after the directed and ad lib bouts for the cigarette-like flux condition and the high flux condition compared to no and low flux conditions. This effect highlights the importance of nicotine flux as a regulatory target. Increased nicotine dose increases rewarding effects, and thus the abuse liability of nicotine delivery products (Jaffe & Jaffe, 1989), although adequate nicotine delivery is also important for suppressing abstinence symptoms. With these results in mind, policies that limit flux are one step towards regulating nicotine dose overall, with the aim of decreasing abuse liability but still suppressing abstinence symptoms.

Consistent with the plasma nicotine results, higher flux led to increased heart rate. Heart rate was significantly higher after both bouts for the low, cigarette-like, and high flux conditions, compared to the no flux condition, as well as in the high flux condition compared to the low flux condition during the directed bout. These findings are expected, given that one effect of nicotine exposure is increased heart rate (Hiler et al., 2017; Jensen et al., 2020; Lopez et al., 2016; Maloney et al., 2019).

The subjective effects of nicotine were also explored. Increased nicotine flux was associated with a reduction in self-reported craving for tobacco product/nicotine, with the greatest decrease in craving observed in the high flux condition. Similarly, participants reported reduced urges to use tobacco product/nicotine in the low, cigarette-like, and high flux conditions, with the greatest decreases in the cigarette-like flux and high flux conditions. Finally, participants reported less difficulty concentrating as nicotine flux increased. These results suggest that higher levels of nicotine flux are related to significantly suppressed tobacco abstinence symptoms, such as craving and urge to use tobacco product/nicotine, even when puff duration is limited. Findings are in line with past research, showing that increased nicotine delivery is related to greater abstinence symptom suppression (e.g., Evans et al., 2006; Hiler et al., 2017). Importantly, suppression of these symptoms is key for reducing the likelihood that people would use cigarettes instead of ENDS but are also related to abuse liability (Carter et al., 2009; Henningfield et al., 2011; Jasinski & Henningfield, 1989). Determining the ideal nicotine dose could meet both goals (suppressing abstinence symptoms but not increasing abuse liability; Shihadeh & Eissenberg, 2015).

Other subjective effects were also measured. Individuals in the higher flux conditions reported feeling calmer and more awake in both bouts, but also reported increased feelings of sickness and dizziness. Individuals in the cigarette-like and high flux conditions also reported increased harshness and increased throat hit of the ENDS, along with decreased warmth and flavor ratings. These findings indicate that while some of the positive subjective sensations increase with higher flux, so too do some of the negative sensations that come with increased nicotine dose. These findings are unsurprising, considering the effects of nicotine dose on subjective effects and sensations (Kalman, 2002).

Puff duration also differed by condition, as did other measures of puff topography. Mean puff duration and volume were shortest in the high flux condition, and in the ad lib bout, inter-puff intervals were longer and puff number was lower for both the cigarette-like and high flux conditions. Results regarding puff topography show that nicotine flux influences puffing behavior. The present work also shows that puff duration in ENDS can be controlled. Given that longer puff duration is associated with increased plasma nicotine concentration (Farsalinos et al., 2015; Hiler et al., 2017; Eversole et al., 2022) and that there is substantial variability in puff duration among users (Farsalinos et al., 2015), puff duration is another target for regulation, in addition to nicotine flux.

However, our findings related to puff duration should be interpreted with caution. Given that puff duration was limited, we did not test how results would have differed if longer allowable puff durations, consistent with current products on the market, were included. Similar patterns of effects, such as increased plasma concentration with varying nicotine concentration of the liquid have been reported in other studies where puff duration was not limited (e.g., Dawkins et al., 2016; Eversole et al., 2023; Hiler et al., 2017). More specifically, in a previous study with similar methodology and a few similar conditions (Eversole et al., 2023), after 10 puffs from an ENDS with 30 mg/ml nicotine and set to 30 W (similar to our high flux condition), we observed plasma nicotine concentrations of 20.04 (SD=22.13) ng/ml in people who smoked cigarettes and 25.11 (SD=21.32) ng/ml in people who used ENDS, compared to 12.42 (SD=12.21) ng/ml after 10 puffs in the high flux condition in this study. Also, in that previous study, after 10 puffs from an ENDS with 15 mg/ml nicotine and set to 30 W (similar to our cigarette-like flux condition), we observed plasma nicotine concentrations of 15.50 (SD=14.01) ng/ml in people who smoked cigarettes and 16.56 (SD=11.54) ng/ml in people who used ENDS, compared to 8.8 (SD=7.41) ng/ml in the present study.

Although this study provides preliminary evidence for the importance of limiting puff duration as a regulatory target, additional work is needed, such as a direct comparison of the effects of limited puff durations with unconstrained puff durations. Nonetheless, findings suggest that, when puff duration is constrained, varying nicotine flux has impacts on the biological and subjective effects of people who use tobacco products. Together, controlling nicotine flux (via nicotine concentration and power settings) and limiting puff duration has the potential to constrain nicotine dose, and thus abuse potential, for people who use ENDS.

We should also note that the findings presented are short-term, acute effects. The extent to which similar findings would persist with longer-term use of ENDS under these conditions is unclear. With longer-term use of these products, individuals may learn and adapt their behaviors to receive the most nicotine possible. Future studies should consider exploring the long-term effects of varying nicotine flux and limiting puff duration on both biological and subjective outcomes. This work has some limitations. First, many more participants in the study sample used ENDS regularly compared to those who used cigarettes exclusively. Individuals who are more experienced with ENDS can have different puffing behaviors than those who are not, taking longer puffs and thereby likely obtaining more nicotine per puff (Farsalinos et al., 2015; Hiler et al., 2017; Spindle et al., 2017). In addition, we found that while puff duration was mostly held to two seconds and under, in some conditions, it significantly exceeded two seconds (although mean puff duration did not exceed 2.31 seconds). This finding may be a result of participants continuing to inhale after the two-second limit, despite the fact that no power was delivered to the coil after two seconds; that is, they were inhaling air, not aerosol. Subsequent work should explore puff duration limits further. For instance, different puff durations should be manipulated and explored, both independent of and in conjunction with nicotine flux, to better understand how puff duration is related to nicotine dose and other effects.

Taken together, the findings presented here suggest that when puff duration is controlled, nicotine is still delivered, and abstinence symptoms are suppressed, especially in higher flux conditions. These results suggest that regulators interested in controlling nicotine dose – the product of flux multiplied by duration – can do so in an effort to limit abuse liability while preserving effects that facilitate cigarette-to-ENDS transitions, such as nicotine delivery and abstinence symptoms suppression.

Table 3.

Summary of Results for Puff Topography Measures.

Outcome measure	Directed bout			Ad lib bout
	F	p	Ƞp2	F	p	Ƞp2
Puff durationa, b	13.75	<.001	.31	3.35	.028	.11
Puff volumec, d	8.31	<.001	.21	4.87	.006	.15
Inter-puff intervala,e	.96	.357	.03	3.22	.031	.11
Puff numbera,b	.74	.510	.02	7.59	<.001	.22
Average flow ratea,b	.51	.576	.02	3.16	.041	.11
Puff number—own brandc	3.66	.015	.11	N/A	N/A	N/A
Liquid usedc	0.23	.663	.01	2.98	.045	.09

^adf = 3, 90

^b= 3, 81

^cdf = 3, 93

^d= 3, 84

^e= 3, 75

Public health significance:

Electronic nicotine delivery systems (ENDS) may facilitate transitions from combustible cigarettes, but also may have high enough abuse liability to lead some nicotine-naive individuals to initiate and maintain ENDS use. Limiting the dose of nicotine delivered by ENDS may reduce abuse liability while preserving effects that help people who smoke transition away from cigarettes. This study examines the effects of limiting dose by varying ENDS nicotine emission rate and controlling puff duration. Results suggest that puff duration can be controlled, demonstrate that high flux suppresses tobacco abstinence effects, and highlight the importance of targeting both flux and puff duration for ENDS regulation.