Sweet Flavors and the Addictive Potential of E-Cigarettes: A Systematic Narrative Review

Abstract

Introduction

Population-based surveys indicate that sweet and fruity flavors in the e-liquids are key drivers of attractiveness of e-cigarettes (ECIGs). However, their role in the addictive potential of ECIGs is less well established. This systematic review aimed to synthetize the available evidence on the impact of sweet and fruity flavors on indicators of nicotine dependence.

Methods

A systematic review was conducted in PubMed, Embase, and Google Scholar in accordance with PRISMA guidelines. Studies were eligible if they reported the effects of sweet/fruity flavors in ECIGs on (1) consumption rate, (2) vaping topography and nicotine intake, (3) reward and reinforcement, and (4) indicators of nicotine dependence. Studies limited to qualitative assessments of flavor properties were excluded. Initially, 454 studies were identified and screened independently by two reviewers, resulting in 23 eligible studies. Risk of bias and study quality were evaluated using the Johanna Briggs Institute Critical Appraisal Tools.

Results

Sweet/fruity flavors were consistently associated with increased consumption rates and a higher puff number and volume, leading to more nicotine exposure. Several studies also demonstrated that sweet flavors enhanced the reward and reinforcement of ECIG use. However, findings regarding their effects on craving and withdrawal were inconsistent.

Conclusion

Sweet/fruity flavors appear to increase the abuse liability of ECIGs, but the evidence remains inconclusive. Additional high-quality studies that adequately control for known covariates are required to clarify their contribution and the independent role of flavors in nicotine dependence. These findings, together with emerging toxicological data on ECIGs, should be carefully considered in regulatory decisions regarding flavor bans in ECIGs.

Introduction

Initially, electronic nicotine delivery devices (electronic cigarettes [ECIGs]) were primarily marketed and used as smoking cessation aids [1]. However, their use has increasingly shifted toward recreational purposes. ECIGs have become particularly popular among youth, including children and adolescents [2]. This trend is largely driven by the perception of many that ECIGs constitute a safer alternative to conventional tobacco smoking [3, 4] and the perceived benefits of abstinence from smoking (less coughing, improved breathing, better physical fitness) [5]. Such reports are largely endorsed by the experts’ overall judgment that conventional tobacco smoking is more harmful to users than vaping ECIGs [6, 7] and supported to some extent by toxicological evidence. Compared to combustible tobacco smokers, ECIG users are exposed to substantial lower levels of carcinogens [8] and toxicants [9], as most toxicants are generated during the pyrolysis of tobacco.

The liquids used in ECIG are now available in thousands of flavors, with estimated 250 new flavors entering the online market each month [10]. Currently, approximately 90% of young adult ECIG users (aged 18–24) report using flavored products, most commonly menthol/mint, fruit, candy/sweet, and tobacco flavors [11]. Sweet/fruity and minty flavors are consistently reported as the most preferred [12, 13]. Since 2012–2013, the market share of unflavored and menthol/mint e-liquids has declined, while fruit and “other” flavors have expanded substantially [14]. This shift parallels changes in user preferences, which have moved away from traditional tobacco and menthol flavors toward sweet and candy-like flavors [13, 15–18]. At present, the ten most popular flavors are dominated by sweet or fruity variants [19].

From a chemosensory perspective, adolescents may be especially susceptible to sweet-flavored ECIG products, as their preference for sweet-tasting substances is stronger than that of adults and decreases with age [20, 21]. Sweet and fruity flavors are among the most influential product attributes driving ECIG appeal [22, 23], owing to their pleasant sensory characteristics [23–26]. In contrast, the harsh and bitter taste of nicotine-containing aerosols reduces hedonic appeal [27]. Sweet flavors can mask these aversive properties [28], and perceived sweetness is positively associated with both overall product liking [29] and initial puff satisfaction [30]. Adolescents have demonstrated more than a six-fold higher interest in trying an ECIG when it is fruit-flavored compared with tobacco-flavored [31]. Moreover, the fruity flavor “berry” elicited higher subjective reward ratings than tobacco flavor [32]. Collectively, these findings indicate that sweet and fruity flavors represent a major driver of ECIG initiation, particularly among adolescents [33, 34].

Nicotine is a weak primary reinforcer and a strong secondary reinforcer. Consequently, self-administration of nicotine via smoking or vaping often leads to sustained use [35]. As the principal addictive constituent of e-liquids, nicotine contributes substantially to dependence among ECIG users, with dependence levels comparable to those observed in cigarette smokers [36], despite users frequently perceiving ECIGs as less addictive [37]. Recent reviews [38, 39] concluded that the risk of ECIG dependence increases with nicotine concentration in e-liquids, more frequent use, and longer duration of vaping.

Flavors may further contribute to the abuse liability and dependence potential of ECIGs, although the extent and mechanisms remain unclear. As noted above, large-scale surveys and other subjective assessments consistently show that sweet and fruity flavors enhance ECIG appeal and consumption [22, 23, 33, 34]. This suggests that such flavors may also increase the dependence potential of ECIGs. The present systematic review aimed to evaluate whether evidence from experimental and controlled studies, using validated instruments and measures, supports this hypothesis. The following objective indicators of ECIG addictive potential were selected: (1) consumption rate (or frequency of use), (2) vaping topography (puff number and puff volume), (3) nicotine dose, (4) objective markers of reward and reinforcement, and (5) dependence-related criteria (craving and withdrawal).

Methods

Literature Search

The study protocol for this systematic review was registered in the International Prospective Register of Systematic Reviews (PROSPERO) with registration number CRD420251147447 and conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. PubMed and Google Scholar together have been shown to capture approximately 93% of the published literature on a given research question [40]. A systematic review was therefore performed in PubMed, Embase and Google Scholar on November 25, 2025, updated on March 11, 2026, to identify relevant studies, including those published ahead of print. Studies were eligible if they examined indicators of nicotine dependence in relation to sweet and/or fruity flavors. Eligible studies included controlled designs comparing sweet and/or fruity-flavored ECIGs with unflavored ECIGs in terms of (1) consumption rate, (2) vaping topography and nicotine intake, (3) objective measures of reward and reinforcement, and (4) dependence-related outcomes, such as craving and withdrawal. All outcomes were required to be assessed using validated measures. Studies were excluded if they (1) used a qualitative design; (2) relied solely on subjective, indirect measures of abuse liability (e.g., preferred flavor, appeal, or liking of certain flavors); (3) focused on smoking satisfaction, smoking cessation, or transition from ECIGs to cigarette smoking; (4) limited to minty flavors; or were (5) case reports, poster presentations, editorials, and commentaries. Studies exclusively on minty flavors were excluded to reduce heterogeneity of the studies as their mechanism of action differs from that of sweet/fruity flavors. Specifically, sweet/fruity flavors primarily induce olfactory and orosensory effects, whereas minty flavors increase the inhalability of ECIG aerosol. The selection of eligible studies was independently performed by two reviewers (author 1 and author 2) in two stages.

The initial search yielded 603 studies. After removal of duplicates, 492 unique records remained. Titles and abstracts were screened against inclusion and exclusion criteria, resulting in 18 eligible studies. Additional five studies were identified through Google Scholar and reference list searches of retrieved articles, yielding a total of 23 studies for inclusion. Table 1 summarizes the characteristics and outcomes of the included studies. The PRISMA flow diagram is presented in Figure 1, and the search strategy and PRISMA checklist are provided in the online supplementary material (for all online suppl. material, see https://doi.org/10.1159/000553008).

Table 1.

Results of experimental studies to the impact of sweet/fruity flavors on consumption rate, nicotine delivery, reinforcing effects, and rewarding effects in subjects using ENDSs (ECIGs)

Design and subjects	Comparators	Main outcome	References
Sweet flavors and consumption rate
Randomized study among young adult smokers (n = 32, aged 18–30) using a within-subjects design	Assessment of puff number using fruit- or dessert-flavored or unflavored ECIGs	In a 90-min ad libitum vaping session, young adult smokers (n = 32) took twice as many flavored puffs (fruit- or dessert-flavored) compared with unflavored ECIG puffs (40 vs. 23 puffs; IRR = 2.028, p = 0.01)	[41]
Prospective study among adolescents who have ever used ECIGs at baseline aged 14–15 (n = 354)	Videotaped ad libitum vaping session of unflavored and flavored ECIGs with nicotine, and initial use of a flavored vs. unflavored ECIGs	During the 90-min vaping session, participants took twice the amount of flavored puffs (fruit or dessert-flavored) compared to unflavored ECIGs puffs (40 vs. 23 puffs; IRR = 2.028; 95% CI: 1.183 to 3.475, p = 0.01). Compared to unflavored ECIGs, initial use of flavored ECIGs was associated with escalation in the number of days of ECIG use (β = 0.35, z = 2.58, p = 0.01) across the following 18 months	[42]
Prospective study among ECIG using adolescents (n = 396 aged 16.2±1.2) and 590 adults (aged 34.3±9.9)	ECIG use in the past month vs. total number of flavors preferred, preferences for fruit or dessert flavor	In adolescents, but not adults, more days of ECIG use during the past 30 days were positively associated with a larger total number of flavors preferred (R2 = 1.60; ηp2 = 0.04), and preferences for fruit (R2 = 2.98; ηp2 = 0.02) and dessert flavor (R2 = 3.80; ηp2 = 0.02) in e-liquids	[43]
Longitudinal study in a convenience sample of past 6-month ECIG use (n = 478, aged 16.1±0.4)	Participants were included if information was available on use of flavor, past 30-day vaping days or episodes per day, past 30-day number of puffs per nicotine vaping episode and frequency outcomes 6 months later	Use of sweet flavors (fruit, candy, sweet or dessert, buttery or combinations) vs. exclusive use of tobacco, menthol or mint, or flavorless was positively associated with vaping continuation (64.3% vs. 42.9%; adjusted OR = 3.76 [95% CI: 1.20 to 10.31]). Flavor used was not associated with number of past 30-day vaping days or episodes per day	[44]
Cross-sectional study in past 30-day ECIG (current) users (n = 344, aged 21.2±0.4)	Flavor most often used in the past month (menthol/mint, fruit/sweet, or iceb) vs. combustible tobacco use, frequency of use, and vaping dependence symptoms	Using primarily ice flavor was associated with reporting more past 30-day vaping days (ice-fruit/sweet vs. fruit/sweet: β = 3.6; 95% CI: 0.8 to 6.4) and more episodes per vaping day vs. fruit/sweet users (β = 2.4; 95% CI: 0.5 to 4.3)	[45]
EMA study in 50 adolescent regular vapers (aged 14–17)	Data were collected on 13.4 days per participant (670 observations). Adolescents used flavors in e-liquid when they vaped nicotine on 87% of occasions (fruit = 55%; mint = 30%; tobacco = 6%; menthol = 5%, and candy, sweets, or chocolate = 5%)	More puffs per vaping episode when vaping fruit flavors (r = 0.13, 95% CI: 0.01 to 0.25, p = 0.030) or tobacco flavor (r = 0.15, 95% CI: 0.01 to 0.29; p = 0.033). However, more vaping occasions on days when tobacco flavor was vaped compared to days when fruit flavors were vaped (r = 0.20, 95% CI: 0.06 to 0.32 p = 0.005). Vaping fruit flavors was not significantly associated with more vaping occasions per day (β = 3.59; 95% CI: −0.27 to 7.46, p = 0.07). On days when they vaped because of flavor appeal, more vaping occasions were observed (r = 0.19, 95% CI: 0.08 to 0.29; p = 0.001) and more total puffs on that day (r = 0.24; 95% CI: 0.13 to 0.34; p < 0.001)	[46]
Crossover study in regular smokers (n = 32; 94% menthol smokers)	Participants self-administered ECIGs containing e-liquids differing in nicotine level (0 mg/mL, 24 mg/mL) and flavor (unflavored, menthol, fruit-flavored) ad libitum. Number of puffs was measured	In nicotine’s absence, but not when combined with 24 mg/mL nicotine, fruit flavor increased ad libitum number of puffs relative to unflavored e-liquids (nicotine-by-flavor: F2, 108.3 = 5.73, p = 0.004)	[47]
Sweet flavors, smoking topography, and nicotine intake
Crossover study in regular vapers (n = 34)	Vaped for 1 week in a natural use environment tobacco flavor, and either berry or menthol flavor in the next week	Berry and menthol flavor affected mean puff flow rate and mean puff volume compared to tobacco flavor. However, the results were inconsistent in terms of how strong berry or menthol differed from tobacco	[48]
Randomized study in men who had either smoked ≥1 cigarette in the past 30 days or ≥1 ECIG in the past 6 months (n = 39, aged 18–45)	In a controlled laboratory setting participants vaped for 10 min ad libitum with one of four e-liquids: cherry or chocolate flavor, each at 6 mg/mL and 18 mg/mL nicotine	Amount of e-liquid vaped dropped as nicotine level increased [F(1,35) = 4.67; p = 0.038], but flavor (cherry vs. chocolate) did not influence amount vaped [F(1,35) = 0.48; p = 0.49]. Conversely, the amount of total nicotine intake (in mg) did not vary by nicotine level [F(1,35) = 2.40; p = 0.13] or flavor [F(1,35) = 0.99; p = 0.33]	[30]
Randomized study in adult daily cigarette smokers (n = 14; aged 37.1±10.5)	Compared to OB combustible cigarettes: (a) nicotine-free unsweetened e-liquid, (b) nicotine-free sucralose-sweetened e-liquid, (c) 15 mg/mL nicotine, unsweetened e-liquid, and (d) 15 mg/mL nicotine, sucralose-sweetened e-liquid	Sucralose + nicotine significantly (p < 0.05) induced longer puff duration (2.34±1.4 vs. 2.07±1.2 s) and larger puff volume (200.0±142.0 vs. 168.9±108.6 mL) compared to unsweetened nicotine. 15 mg/mL nicotine, unsweetened liquid, and 15 mg/mL nicotine, sucralose-sweetened liquid increased plasma nicotine by 7.09±8.6 ng/mL and 5.85±7.0 ng/mL, respectively (p > 0.05)	[49, 50]
Longitudinal study in a convenience sample of past 6-month ECIG users (n = 478, aged 16.1±0.4)	Participant included if information was available on use of flavor, past 30-day vaping days or episodes per day, past 30-day number of puffs per nicotine vaping episode and frequency outcomes 6 months later	Use of sweet flavors (fruit, candy, sweet or dessert, buttery or combinations) vs. exclusive use of tobacco, menthol or mint flavor, or flavorless was positively associated past 30-day number of puffs per nicotine vaping episode (mean: 3.1±5.5 vs. 1.5±3.8; adjusted rate ratio = 2.41 [95% CI: 1.08 to 5.92]) 6 months later	[44]
Crossover study over 3 days in exclusive ECIG users or dual users (n = 14; aged 32.3±13.8)	Each day consisted of a 15-puff session followed by 4 h of abstinence, then 90 min ad libitum vaping with strawberry flavor, tobacco flavor, and their usual brand e-liquid	Compared to tobacco flavor, strawberry flavor significantly increased plasma AUC0–180 (p = 0.03) and peak plasma nicotine concentration was 22% higher (p = 0.17). However, the amount of nicotine inhaled and systemically retained was no longer significantly different between strawberry and tobacco flavor when normalized by the amount of nicotine systemically retained (ratio plasma nicotine AUC0–180 per nicotine retained is 1.02; 95% CI: 0.90 to 1.15, p = 0.74)	[51]
Crossover study over 3 days in exclusive ECIG users or dual users (n = 14; aged 32.3±13.8)	Per day a 15-puff session followed by 4 h of abstinence, then 90 min ad libitum vaping with strawberry flavor, tobacco flavor, and their usual e-liquid	Compared to tobacco flavor, puff duration was significantly longer for strawberry flavor (3.2±1.3 s vs. 2.8±1.1 s), but no difference in the total number of puffs. Strawberry was associated with more often vaping in small groups of puffs (1–5 puffs) compared to the tobacco flavor. The strength of the relationship between vaping topography and nicotine intake and exposure was not consistent across e-liquids	[52]
Randomized study (n = 18 adult daily smokers)	Vaped on each of 5 days one e-liquid containing nicotine (24 mg/mL) and five different flavors: cherry, classic tobacco, espresso, menthol, and vanilla	Puff frequently on cherry was higher compared to tobacco flavor (p = 0.013). Cherry flavor resulted in the highest peak plasma nicotine concentration (Cmax of 21.2 ng/mL), which was nearly 2-fold higher than tobacco flavor (Cmax of 12.5 ng/mL). Plasma nicotine AUC0–120 for cherry and tobacco flavor was 293±318 and 172±200 ng/ml/min, respectively	[53]
Double-blind crossover RCT in regular ECIG users (n = 52; 29.6±10.2, range 18–55). Current use of other tobacco products was minimal (five reported use ≤4 times/month)	On separate days, participants vaped tobacco, menthol, or fruit-flavored e-liquids with 1.2% (all flavors) or 0% (fruit only) free-base nicotine concentration for 10 prescribed puffs and 1 h ad libitum	Following ad libitum use, tobacco flavor was associated with significantly lower plasma nicotine Cmax and AUC0-60 relative to menthol or fruit (p < 0.001). Fewer puffs were taken when tobacco-flavored than menthol or fruit (p < 0.05). Fruit flavor and 0% nicotine were associated with greater puff duration (>27% higher; F = 18.96, p < 0.001) and puff volumes (>33% higher; F = 14.48, p < 0.001) than other products	[54]
Sweet flavors in paradigms of reinforcement
Randomized study among young adult smokers (n = 32, aged 18–30) using a within-subject design	Assessment of the relative reinforcing value of nicotine-containing e-liquids with fruit or dessert flavors. Unflavored was assessed at a fixed ratio FR-25 schedule; flavored was assessed at a PR-25x schedule	Using this paradigm, participants worked, despite being 12 h into nicotine withdrawal, approximately six times harder for the opportunity to self-administer flavored ECIG puffs vs. non-flavored ECIG puffs (breakpoint = 5.7; 597 responses vs. 127 responses; β = 460.733; 95% CI: 246.58 to 674.88, p < 0.0001)	[41]
Randomized study among light cigarette smokers (n = 16, mean age: 27 years, range 19–45)	Participants smoked four different ECIGs containing sweetened and unsweetened flavors in e-liquids with or without nicotine for 2 days each. Subsequently, the fMRI brain response to the sight and smell of these four ECIG types was measured	Compared to non-sweet flavors, sweet flavor-related visual and olfactory cues elicited a stronger BOLD response in the NAcc (p = 0.050), but not the nicotine-paired (p ≥ 0.05 uncorrected). The magnitude of response to the sight (p = 0.022) and smell (p = 0.017) of the ECIGs correlated with changes in liking. The sight and smell ECIGs paired with sweet + nicotine (p = 0.035) produced supra-additive NAcc responses, i.e., heightened brain cue reactivity indicating that sweet taste potentiates the reinforcing effects of nicotine in ECIGs	[55]
fMRI crossover study among young adults (n = 26; aged 18–25) who had tried an ECIG and were susceptible to future ECIG use	Participants viewed advertisements for sweet/fruit and tobacco-flavored ECIGs, and menthol and regular cigarettes. Cue reactivity was measured with fMRI in the NAcc	ECIG advertisement with images related to sweet/fruit flavors induced greater NAcc cue reactivity than those related to tobacco flavor, whereas no response was found for the cigarette condition (menthol/regular/control)	[56]
fMRI crossover study performed in nine adult smokers	Participants inhaled aerosolized e-liquid with 36 mg/mL of nicotine with and without a strawberry-vanilla flavor. Using fMRI, activation of brain regions involved in stimulus reinforcement and learned conditioning were studied	Compared to unflavored, strawberry-vanilla flavor inhibited the activation of dopaminergic addiction-related neural circuits. Compared to unflavored aerosol, strawberry-vanilla-flavored aerosol induced a stronger functional connectivity between subcortical dopaminergic and cortical brain regions involved in learning and reward value of taste	[57]
Sweet flavors and craving
Systematic review on dependence vs. e-liquid flavors	Six cross-sectional studies on this issue [45, 58–62] were cited in this systematic review. See below for details on comparators	Douglas et al. [62] (see below) found no significant difference in dependence between flavored and unflavored ECIG use. Significantly higher (p < 0.05) risk of dependence was found in ice-fruit/sweet-flavored ECIG users [45] (see below) as well as an insignificant role of flavor in dependence [61] (see below). The menthol/mint study [60] and two studies not properly specifying the flavor [58, 59] were excluded from the current review	[38]
Cross-sectional study using a convenience sample of current adult ECIG users (n = 711; aged 18–74) who “never” smoked conventional cigarettes (<100 cigarettes in lifetime)	Included in analyses when information was available on use of flavor preference, nicotine concentration, as predictors of dependence (16 validated itemsa), withdrawal, and craving. “Wisconsin Inventory of Smoking Dependence Motives” was used to assess dependence score, and seven questions related to withdrawal symptoms were used to rate withdrawal	Dependence scores were significantly higher for “tobacco flavor only” relative to sweet/fruit only and other only flavors. Higher dependence was associated with older age. Compared to tobacco flavor, higher withdrawal levels were observed for “other” flavor preference (β = 2.26±0.71, p = 0.001) and lower withdrawal levels were observed for sweet/fruit flavor(s) (β = −1.87±0.77, p = 0.016)	[62]
Cross-sectional study in past 30-day ECIG (current) users (n = 344, aged 21.2±0.4). History of combustible tobacco use was not reported	Flavor most often used in the past month (menthol/mint, fruit/sweet, or iceb) vs. combustible tobacco use and vaping dependence symptoms assessed by the modified “Hooked on Nicotine Checklist.” Nicotine concentration was not assessed	Use of primarily ice-fruit/sweet was associated greater odds of vaping dependence symptoms than primary use of sweet/fruit users (OR = 2.6, 95% CI: 1.5 to 4.4, p < 0.05). Flavored ECIG use: menthol/mint (21.2±0.3); fruit/sweet (21.2±0.4); and ice (21.2±0.4); p = 0.90. Ice-fruit/sweet flavor users reported more past 30-day vaping days than fruit/sweet (b = 3.6; 95% CI: 0.8 to 6.4). No further analysis was reported on dependence, corrected for, e.g., combustible tobacco use, age, dual use, or frequency of use	[45]
Longitudinal study in a convenience sample of past 30-day ECIG users (n = 1,430; aged 15 to 24). Information on current or previous use of other tobacco products was not available	Studies included if information was available on (a) most-used flavor (fruit, mint, menthol/ice, and tobacco) and the nicotine concentration (0–2.9%, 3–4.9%, and 5% or greater) in the e-liquid used in the past 30 days and (b) time to first vape after waking (within 30 min, longer than 30 min) as proxy for nicotine craving (dependence)	Fruit, menthol, and tobacco as most flavor used had a similar prevalence of craving (53.2%, 59.7%, and 58.3%, respectively). Also for the different nicotine strengths (0–2.9%, 3–4.9%, or 5%), fruit flavor craving rates were similar as nicotine alone. A higher prevalence of craving was observed in the tobacco flavor group using a nicotine concentration of 0–2.9% (78.8%, 95% CI: 65.4 to 94.8) compared to the fruit flavor group using a nicotine concentration of 5% or more (69.1%, 95% CI: 64.4 to 74.2). Sensitivity analyses demonstrated that these associations were relatively consistent across all age groups, suggesting no difference in effects by age	[61]
Randomized study among adult daily cigarette smokers (n = 14; aged 37.1±10.5)	Compared to OB combustible cigarettes: (a) nicotine-free unsweetened e-liquid, (b) nicotine-free sucralose-sweetened e-liquid, (c) 15 mg/mL nicotine, unsweetened e-liquid, and (d) 15 mg/mL nicotine, sucralose-sweetened e-liquid. Ratings of smoking urges for craving were evaluated using QSU-Bc	Ratings for “urges to smoke” decreased from 70.79 at baseline to 55.64 following 15 puffs of unsweetened nicotine e-liquid and from 72.50 at baseline to 57.36 following 15 puffs of the sucralose-sweetened nicotine e-liquid	[49, 50]
Randomized study among 84 adult daily smokers (aged 28.8±9.9, range 18–59)	Smoking nicotine-containing e-liquids flavored with fruit flavor (blackcurrant, strawberry, vanilla, caramel) or not for 1 week. QSU-B was used to rate craving for conventional cigarettes	No significant effect of e-liquid flavoring on average cigarette craving (β = 0.18, 95% CI: −0.44 to 0.79, p = 0.57), peak cigarette craving (β = −0.12, 95% CI: −0.59 to 0.35, p = 0.62), or cue-elicited cigarette craving (β = −0.21, 95% CI: −3.86 to 3.43, p = 0.91). The results did not differ after adjustment for age, gender, cigarettes per day, cigarette dependence, and quit motivation	[63]
Randomized study among 81 adult smokers (aged 29.8±13.2, range 18–68) without prior vaping experience, but got acquainted with using ECIGs during a 1-week try-out period	After 8-h abstinence, participants vaped for 5 min e-liquid without or with 3.6% nicotine. Half of e-liquids contained apple flavor, and the other half tobacco flavor. Using visual analogue scalesd, craving for ECIGs was measured pre- and 5 min post-vaping	Before and 5 min post-vaping, craving for cigarettes or ECIGs was independent on the flavor of ECIG (tobacco vs. apple; Χ2 (df) = 3.95 (4); p = 0.41 and Χ2 (df) = 1.77 (4); p = 0.78), respectively	[64]
Randomized study among 18 adult daily smokers (aged 44.1±7.0, range 18–55). 11 participants had prior experience with any vaping; 7 vaped more than a year ago and 4 to a few months ago	After 8-h abstinence, on five different days e-liquid with 24 mg/mL nicotine plus five different flavors (cherry, tobacco, espresso, menthol, and vanilla) were vaped. Craving assessed using the QSU-B	At 3 min post-use, smoking urges were less prominent when using cherry-flavored e-liquids compared with tobacco, menthol, espresso, and vanilla flavors. However, the level of smoking urges after using cherry-flavored e-liquids was not significantly lower compared with a combustible cigarette (p = 0.0547). Differences in smoking urges between participants who had ever vaped and those who had never vaped were not evaluated	[53]
Double-blind crossover RCT in regular ECIG users (use: ≥25 days in the past month; n = 52; aged 18–55). Current use of other tobacco products was minimal (five reported use ≤4 times/month)	On separate days, participants vaped tobacco, menthol, or fruit-flavored e-liquids with 1.2% (all flavors) or 0% (fruit only) free-base nicotine concentration for 10 prescribed puffs and 1 h ad libitum. Withdrawal was assessed using QSU-B and ASSR	Compared to the other flavors, only the use of nicotine-free fruit-flavored ECIGs was associated with significantly greater withdrawal symptoms (e.g., ENDS craving, or urge to use an ENDS): lower ratings on craving, urge to use (ASSR: p < 0.05; QSU-B: p < 0.02)	[54]

ENDS, electronic nicotine delivery device; IRR, incidence rate ratio; fMRI, functional MRI; OB, own-brand; QSU-B, Questionnaire of Smoking Urges Brief; ASSR, Abstinence Symptom Suppression Rating.

^aStrong et al. [97].

^bMarketed as a combination of fruity/sweet and cooling flavors (e.g., “blueberry ice” or “melon ice,” described here as “ice-fruit/sweet”).

^cQSU-B consisting of 10 items with a 10-point Likert scale for each item resulting in total scores ranging between 0 (strongly disagree) and 100 (strongly agree).

^dDawkins et al. [98].

Fig. 1.

PRISMA flow diagram.

Risk of Bias and Quality Assessment

The risk of bias was assessed using the Joanna Briggs Institute (JBI) Critical Appraisal Tools (CAT), with the specific checklist selected according to each study design (systematic reviews, observational cohort studies, and randomized trials) [65, 66]. The reviewers (author 1 and author 2) independently appraised all included studies. Discrepancies were resolved through discussion until consensus was reached. Each checklist item was rated as yes (no bias), no (bias present), unclear, or not applicable. As recommended by the JBI manual [65, 67], the scoring system and thresholds for classifying risk of bias were defined a priori, based on Algarni et al. [68], and agreed upon by both reviewers.

Quality assessment scores were calculated as percentage of affirmative responses relative to the total number of applicable items. Items deemed not applicable were excluded from the denominator. Studies were categorized as follows: low risk of bias (JBI score >66%), moderate risk of bias (JBI score 33–66%), and high risk of bias (JBI score <33%) [68].

Results

Online supplementary Table S1 presents the available evidence on the role of sweet/fruity flavors in e-liquids in the addictive potential of ECIGs, using different indicators of addiction risk or liability.

Sweet Flavors and Consumption Rate

In a 90-min ad libitum vaping session, young adult smokers (n = 32) took nearly twice as many puffs from flavored ECIGs (fruit- or dessert-flavored) compared with unflavored ECIG puffs (40 vs. 23 puffs; incidence rate ratio = 2.028, p = 0.01) [41]. Furthermore, initial use of flavored ECIGs was associated with greater progression to current ECIG use (β = 0.54, p = 0.04) and a steeper increase in the number of ECIG use days over an 18-month follow-up period (β = 0.35, p = 0.01) [42]. Among adolescents (n = 396; mean age 16.2 ± 1.2), but not adults, flavor preferences were significantly associated with past-month ECIG use frequency. More frequent use was linked to the number of preferred flavors (η_p² = 0.04; p < 0.001), as well as preferences for fruit (η_p² = 0.02; p < 0.01), candy/dessert (like apple pie or chocolate) (η_p² = 0.02; p < 0.01), although effect sizes were small [43].

In a longitudinal study of 478 young ECIG users (mean age 16.1 ± 0.4), exclusive use of sweet flavors was significantly associated with continued vaping compared with exclusive use of tobacco, menthol or mint, or unflavored liquids (64.3% vs. 42.9%; adjusted OR = 3.76). However, flavor was not associated with the number of past 30-day vaping days or episodes per day [44]. In a cross-sectional study of young adults (n = 344), the use of primarily “ice” flavors (fruit/sweet combined with cooling flavors) was associated with more past 30-day vaping days (β = 3.6, 95% CI: 0.8–6.4) and more daily episodes (β = 2.4, 95% CI: 0.5–4.3) compared with fruit/sweet users, though these differences were not statistically significant [45]. Ecological momentary assessment (EMA) data from adolescent regular vapers (n = 50; aged 14–17) were collected over 13.4 days per participant [46]. Participants used fruit flavors on 55% of occasions, mint on 30%, tobacco on 6%, menthol on 5%, and candy, sweets, or chocolate flavors on 5% of occasions. Fruit flavors were not significantly associated with the number of daily vaping occasions (β = 3.59, p = 0.07), although a higher number of occasions were observed on days when tobacco flavors were used (r = 0.20, p = 0.005). Adolescents took more puffs when vaping fruit (r = 0.13, p = 0.030) or tobacco flavors (r = 0.15, p = 0.033) compared with unflavored products. Finally, in a controlled laboratory study, fruit flavor increased the number of ad libitum puffs relative to unflavored e-liquids in the absence of nicotine, but not when combined with 24 mg/mL nicotine [47].

In summary, four of six studies reported that sweet flavors were positively associated with higher ECIG use, whereas the other two studies reported mixed [43] or non-significant [45] associations.

Sweet Flavors, Smoking Topography, and Nicotine Intake

Findings on the impact of sweet flavors on smoking topography and nicotine intake (e.g., nicotine peak plasma and nicotine boost) were inconsistent. In one naturalistic study, regular vapers (n = 34) used tobacco-flavored e-liquid for 1 week, followed by either berry or menthol flavor for 1 week. Flavors significantly affected mean puff flow rate and mean puff volume, but the study was underpowered to assess associations with cumulative consumption [48]. In a randomized laboratory study, adult participants (n = 39) vaped ad libitum for 10 min with one of four e-liquids: cherry or chocolate flavor, each at 6 mg/mL and 18 mg/mL nicotine. At the higher nicotine concentration, the volume of e-liquid consumed decreased (p = 0.038), but the total nicotine intake did not vary significantly by nicotine concentration (p = 0.13) or flavor (p = 0.33) [30].

In two studies, Maloney et al. [49, 50] examined the effects of sucralose (an artificial sweetener). In a within-subject design, daily smokers used (a) their own-brand cigarettes; (b) nicotine-free, unsweetened liquid ECIG; (c) nicotine-free, sucralose-sweetened liquid ECIG; (d) 15 mg/mL nicotine, unsweetened liquid, and (e) 15 mg/mL nicotine, sucralose-sweetened ECIG liquid. Compared with unsweetened nicotine condition, the sucralose + nicotine condition increased puff duration by 13% and puff volume by 18%, but did not significantly affect plasma nicotine boost (5.85 ± 7.0 ng/mL vs. 7.09 ± 8.6 ng/mL; p > 0.05) [49, 50]. A longitudinal study in a convenience sample of young ECIG users (n = 478) showed that the exclusive use of sweet flavors was positively associated with the number of puffs per vaping session over the past 30 days, compared with exclusive use of tobacco, menthol or mint, or unflavored products (adjusted rate ratio = 2.41) 6 months later [44].

In a small 3-day crossover study (n = 14), adult ECIG users completed 15-puff sessions and 90-min ad libitum sessions with strawberry, tobacco, and their usual brand e-liquids. Puff duration, nicotine inhaled/retained, and peak plasma nicotine (C_max) did not differ significantly between strawberry and tobacco flavors. Although plasma AUC_0-180 was higher for strawberry than for tobacco, values normalized to systemically retained nicotine were nearly identical (ratio: 1.02, p = 0.74), suggesting no effect of flavor on systemic nicotine absorption in the lower airways [51, 52], possibly due to limited statistical power. In a larger clinical trial, adult daily smokers (n = 18) vaped a 24 mg/mL nicotine solution with five randomized flavors (cherry, classic tobacco, espresso, menthol, and vanilla). Cherry flavor produced the highest plasma nicotine concentration (C_max = 21.2 ng/mL), nearly double that of classic tobacco (12.5 ng/mL), and approaching levels observed with combustible cigarettes (29.2 ng/mL, p > 0.05). Plasma nicotine AUC_0-120 was also higher for cherry compared with tobacco (293 ± 318 vs. 172 ± 200 ng/mL/min) [53]. Finally, a recent study reported that following ad libitum use, tobacco flavor was associated with significantly lower plasma nicotine C_max and AUC_(0-60) compared with menthol or fruit flavors (p < 0.001), likely due to fewer puffs taken (p < 0.05). Notably, the use of nicotine-free fruit-flavored ECIGs was associated with greater puff duration and puff volumes than other products (p < 0.001) [54].

In summary, sweet flavors influenced vaping topography parameters (e.g., puff number, duration, flow rate, volume) in five of seven studies. However, only one of three studies found evidence for increased systematic nicotine delivery.

Sweet Flavors and Paradigms of Reinforcement

Through associative learning mechanisms, sweet flavors may potentiate the reinforcing effects of nicotine and vaping. Using a within-subject design, the relative reinforcing value of fruit or dessert flavors, formulated in nicotine-containing e-liquids, was evaluated in young adult smokers (n = 32) [41]. After two puffs, subjective reward ratings were higher for flavored versus unflavored ECIGs (β = 0.83, p = 0.001). In a computerized progressive ratio task assessing participants’ willingness to “work” for flavored or unflavored ECIG puffs, substantially greater effort was exerted to self-administer flavored ECIG puffs than for non-flavored ECIG puffs (597 responses versus 127 responses; β = 460.733, p < 0.0001), demonstrating increased reinforcing value [41].

Kroemer et al. [55] employed a conditioning paradigm in 16 light cigarette smokers who were exposed to vaporized flavored or unflavored e-liquids, with and without nicotine, across four conditions. Functional MRI revealed that sweet nicotine-containing e-liquids supra-additively enhanced liking (r = −0.67, p = 0.005) and elicited stronger nucleus accumbens (NAcc) responses to visual and olfactory cues compared to non-sweet flavors (p = 0.050). Cue-related NAcc responses correlated significantly with changes in subjective liking (sight: p = 0.022; smell: p = 0.017). Sweet taste paired with nicotine produced supra-additive NAcc activation (p = 0.035), suggesting potentiation of nicotine reinforcement [55]. Similarly, a functional MRI study among youth adults (18–25 years) with prior ECIG exposure demonstrated greater NAcc cue reactivity to sweet/fruit versus tobacco-flavored ECIG advertisements [56]. By contrast, Hobkirk et al. [57] reported attenuated reward responses in a small sample of female daily smokers (n = 9). Compared with unflavored e-liquid (36 mg/mL nicotine), strawberry-vanilla-flavored e-liquid inhibited activation of dopaminergic reward-related neural circuits. Additionally, flavored aerosol strengthened connectivity between subcortical dopaminergic and cortical brain regions involved in taste-reward learning, potentially enhancing decision-making and behavioral regulation [69]. However, this exploratory study was limited by its small sample size and methodological constraints, including consecutive testing runs (i.e., absence of a crossover design) and the fact that participants had not abstained from smoking prior to testing.

In summary, three of four studies provided evidence that sweet flavors enhance reinforcement, either through increased effort to obtain puffs or through heightened activation of neural reward pathways. One small neuroimaging study, however, suggested that fruit flavors may attenuate reward-related activation.

Sweet Flavors and Craving

Craving, a core symptom of nicotine dependence, is commonly assessed using validated self-report questionnaires. A recent systematic review identified six cross-sectional studies comparing flavored with unflavored ECIGs [38]. Three of these cross-sectional studies were excluded from the current review because they either did not adequately specify the flavor [58, 59] or focused exclusively on menthol/mint flavor [60]. The remaining three studies cited by Kundu et al. [38] were included in the present review, together with four other additional experimental studies on craving [45, 49, 50, 53, 61, 63, 64].

The experimental studies included in the present review corroborate the mixed nature of available evidence. Among four randomized studies, two reported a positive effect of sweet flavors on craving, whereas the other two found no significant effect. Of the three cross-sectional studies, two suggested a positive association.

Cross-sectional studies likewise yielded inconsistent findings. In a convenience sample of 711 current ECIG users who virtually “never” smoked conventional cigarettes, participants preferring “sweet/fruit flavors only” or “other flavors only” had significantly (p < 0.001) lower dependence scores (1.53 ± 0.73 and 1.22 ± 0.47, respectively) compared with tobacco flavor users (2.43 ± 1.19) [62]. Interestingly, older age was associated with higher dependence scores in this study. Similarly, in a study among relatively young current ECIG users (n = 344), sweet/fruit flavor use was associated with lower odds of dependence compared to primarily ice-fruit/sweet-flavored e-liquids (OR = 2.6, p < 0.05) [45]. This finding is consistent with the observation that users of ice flavor, i.e., ice-fruit/sweet, showed higher vaping frequency than users of fruit/sweet flavors (past 30-day vaping days: 17.0 vs. 12.1; b = 3.6; 95% CI: 0.8–6.4). However, no additional analysis was performed to evaluate the association between flavors and dependence after adjustment for potential confounders such as combustible tobacco use, age, dual use, or vaping frequency. Among relatively young ECIG users, Do et al. [61] found that flavor preference (fruit, menthol, and tobacco) did not significantly affect the prevalence of vaping within 30 min after waking, a commonly used proxy for dependence. Users in the fruit flavor group with a nicotine concentration of ≥5% (69.1%, 95% CI: 64.4–74.2) showed a lower prevalence of craving than users in the tobacco flavor group using a nicotine concentration of 0–2.9% (78.8%, 95% CI: 65.4–94.8). Sensitivity analyses revealed no significant effects of age. However, information on current use of other tobacco products was unavailable, preventing assessment of this potentially important confounder. In a randomized study involving 14 adult daily cigarette smokers, Maloney et al. [49, 50] observed no significant effect of sweet flavor on craving. Similarly, in a randomized trial, daily smokers (n = 84) used either sweet/fruit-flavored or unflavored nicotine-containing e-liquid for 1 week; neither average cigarette craving (β = 0.18, p = 0.57) nor cue-elicited craving (β = −0.21, p = 0.91) differed between conditions [63]. These findings remained unchanged after adjustment for covariates including age, gender, cigarettes per day, and cigarette dependence. Likewise, van Heel et al. [64] reported no effect of flavor (tobacco vs. apple) on craving after overnight abstinence (p = 0.78). By contrast, in another randomized crossover trial involving 18 daily smokers, cherry-flavored e-liquid reduced smoking urges relative to tobacco, espresso, menthol, and vanilla flavors [53]. However, the potential influence of covariates such as age and prior vaping experience was not evaluated in this study. Finally, a recent study among regular ECIG users who rarely used other tobacco products demonstrated that, compared to the other flavors, only nicotine-free fruit-flavored ECIGs were associated with significantly greater withdrawal symptoms, including craving and vaping urges [54].

In summary, among the experimental studies reviewed, findings regarding the effect of sweet flavors on craving and withdrawal were inconsistent. Five studies demonstrated an additional effect of sweet flavors on craving, while three others found no significant differences. Cross-sectional evidence similarly suggests heterogenicity, with some studies linking sweet flavors to reduced dependence indicators and others reporting no effect.

Risk of Bias

The appraisal data for each study is provided in online supplementary Table S1, while the overall quality scores and assigned risk-of-bias categories are presented in online supplementary Table S2. Two of the 23 included studies [49, 50] reported on the same dataset and should therefore be considered a single study. In addition, one included publication was a systematic review. Of the 23 studies assessed, only two were classified as having a low risk of bias, whereas the remaining 21 were classified as having a moderate risk of bias. Several studies employed ad libitum vaping protocols, which may introduce substantial variability in exposure. Furthermore, the heterogeneity in study designs and research objectives across the four domains summarized in online supplementary Table S1 negatively affected the overall consistency of the evidence base. Accordingly, the predominance of studies with moderate to high risk of bias was not unexpected.

Discussion

This systematic review evaluated whether the addition of sweet or fruity flavors to e-liquids increases the dependence potential of ECIGs. Overall, findings were consistent with survey-based evidence indicating that preference for sweet flavors is associated with more ECIG use, including more frequent vaping days and more vaping episodes per day. Sweet flavors also appeared to alter vaping topography (e.g., greater puff number, flow rate, duration, or volume), although it remains uncertain whether these behavioral changes translate into higher systemic nicotine exposure and more severe nicotine dependence. Importantly, most – but not all – studies provided evidence that sweet flavors enhanced reinforcement. By contrast, findings regarding craving and other dependence-related indicators were inconsistent, with approximately one-third of the studies reporting an enhancing effect of sweet or fruity flavors.

The reinforcing role of sweet flavors is biologically plausible. Aversive sensory effects of nicotine and ECIG aerosols (e.g., bitterness, throat irritation) may exert a protective effect against addiction [70, 71]. Sweet/fruity flavors can counteract these aversive sensations by enhancing perceptions of sweetness and smoothness while reducing bitterness, thereby improving palatability and facilitating initiation, maintenance, and frequent use [22, 23, 28, 55, 72]. These sensory modifications may contribute to cumulative nicotine exposure and, thus, abuse liability [22, 23, 42, 43, 73]. Because the degree of physical dependence is dose-dependently related to systemic nicotine absorption [74, 75], any factor that increases puff volume or duration, such as flavor-induced changes in puff topography, could indirectly enhance nicotine reinforcement.

Nicotine flux, defined as the efficiency of nicotine aerosol transfer to the airways per unit of time (µg/s), is a key ECIG performance metric influenced by product design, heater power settings, evaporation temperature, and liquid composition [76, 77]. Puff topography strongly affects the delivery of aerosolized nicotine to the lower airways [78]. Consequently, sensory attributes such as sweet flavors may promote longer puff durations or larger puff volumes, thereby increasing systemic nicotine dose and reinforcing potential. Although several studies reported flavor-induced increases in puff parameters, plasma nicotine levels were not consistently elevated (cf. Table 1). However, nicotine pharmacokinetic assessment during vaping is methodologically challenging and subject to considerable variability (cf. Table 2). Standardized approaches, such as smoking puff analyzers and fixed-puff bout protocols, rather than ad libitum vaping, are therefore essential for more reliable measurement.

Table 2.

Pitfalls and variables in the measurement of cumulative nicotine intake during vaping

Variable	References
Subjects
Perception of bitterness depends on smoking status (smoker vs. nonsmoker)	[79]
Gender differences in sensory sensation and appeal	[80]
Experienced vs. inexperienced ECIG users	[81]
Adolescents have a higher appreciation for sweet tastes than adults	[39, 82]
Design and chemistry
Product design (rechargeable pod vapes vs. disposable)	[83]
Device power (power of heater, evaporation temperature)	[77]
Liquid level in the atomizer tank and temperature	[77]
Concentration of nicotine in e-liquid	[81]
pH of the e-liquid known to be crucial for nicotine absorption	[84]
Free/protonated nicotine ratio known to be crucial for nicotine absorption	[84]
Salt nicotine formulations are rated as sweeter and smoother, and less harsh than free-base nicotine formulations	[85]

In addition to nicotine concentration (strength), non-nicotine constituents, including flavors, may act as secondary reinforcers through conditioning mechanisms [86–89]. Flavor-related sensory cues can acquire reinforcing properties through Pavlovian associations with nicotine reward [90]. Three studies included in this review [41, 55, 56] demonstrated reinforcement-enhancing effects of sweet flavors, whereas one small neuroimaging study suggested attenuation of reward-related responses [57].

Because no gold-standard measure of ECIG dependence currently exists, dependence and craving were assessed using a variety of validated questionnaires (e.g., “urge to smoke” scales or time to first vape after waking) resulting in heterogenous findings. Five of eight experimental studies [45, 49, 50, 53, 54, 62] reported an additive effect of sweet flavors, whereas others did not. Evidence also suggests age-related variation: adolescents and young adults demonstrate a stronger preference for fruit-flavored ECIGs (adjusted OR = 3.35) [82]. In contrast, positive associations between sweet flavor preference and craving were more commonly observed in the cohorts with older participants. Moreover, in one study, older age predicted higher craving and more severe ECIG dependence [62]. Most studies did not (adequately) control for covariates known to influence the relationship between sensory perceptions and dependence (cf. Limitations). For instance, two studies examining dependence and craving [45, 61] did not control for tobacco smoking history, whereas another study [53] did not account for prior vaping experience.

Sweet flavors may also be relevant in smoking cessation strategies. Because sensory cues associated with craving can act as substitutes for smoking-related cues, sweet flavors may alleviate craving in abstinent smokers [86]. Glucose supplementation during nicotine withdrawal has been reported to reduce craving and withdrawal symptoms and to improve cessation rates when combined with nicotine replacement therapy [91–94], although not all studies have replicated these effects [95].

Limitations

The studies included in this review exhibit several important limitations. Many relied on convenience samples or heterogeneous populations comprising conventional tobacco smokers, former smokers, ECIG-only smokers, experienced ECIG users, and dual users. This is an important issue because the rewarding effects of flavored nicotine-containing ECIGs differ between experienced and inexperienced vapers [23, 29], potentially confounding the results. Moreover, sensory perceptions of nicotine bitterness [79] and tobacco flavor differ between experienced smokers and non-smokers due to chronic exposure. A prior history of smoking may also affect the findings, as ECIG users with dual nicotine use and an earlier initiation of nicotine consumption have been shown to exhibit significantly higher dependence scores than exclusive users and individuals with later smoking onset [96]. Notably, older adult participants generally demonstrated a higher prevalence of past smoking and smoking cessation attempts compared with younger participants (Soneji et al. [82]). It is therefore plausible that tobacco-flavored ECIG users may exhibit greater nicotine dependence as a consequence of prior tobacco use before initiating ECIG use. Age itself may also act as confounding factor, as preference for sweet flavors is stronger during adolescence and tends to decline with age [20, 21]. It should further be noted that commercial sweet/fruity e-liquids may contain minty flavors or non-menthol synthetic cooling agents (not declared) which may have introduced bias in some of the results. Ethical considerations have largely restricted clinical studies to adult populations, despite adolescents representing a particularly vulnerable high-risk group. Finally, most experimental studies were limited by relatively small sample sizes (typically 10–30 participants), thereby reducing statistical power.

Conclusions

Previous studies have demonstrated that sweet/fruity flavors substantially enhance the popularity and attractiveness of ECIGs. The experimental evidence summarized in this review indicates that such flavors promote ECIG use and alter vaping topography in ways that may increase abuse liability. However, it remains unclear whether these flavor-induced changes in puff topography consistently translate into greater nicotine uptake and higher levels of nicotine dependence. Sweet and fruity flavors may also function as a secondary reinforcer, with several studies showing increased reinforcement and craving, although findings remain inconsistent.

Taken together, the available evidence suggests that sweet/fruity flavors may increase the dependence potential and abuse liability of ECIGs. Nevertheless, inconsistencies in the findings and important methodological limitations underscore the need for additional high-quality research, particularly among adolescents who appear to be especially vulnerable. In light of emerging toxicological evidence regarding vaping, the potential contribution of flavors – particularly sweet/fruity flavors – to ECIG dependence should be carefully considered in regulatory decisions concerning flavor restrictions in ECIG products.

Statement of Ethics

A statement of ethics is not applicable because this study is based exclusively on published literature.

Conflict of Interest Statement

The authors have no conflicts of interest to disclose.

Funding Sources

No funding was received.

Author Contributions

The following contributions were provided: conceptualization, Jan van Amsterdam; methodology, investigation, writing – original draft, review, and editing, and preparation, Jan van Amsterdam and Wim van den Brink; supervision, Wim van den Brink.

Funding Statement

No funding was received.

Data Availability Statement

This study is based exclusively on published literature which is included in this article and its online supplementary material files.