Then and now: Consumption and dependence in e-cigarette users who formerly smoked cigarettes

Addictive Behaviors 76 (2018) 113–121

Contents lists available at ScienceDirect

Addictive Behaviors journal homepage: www.elsevier.com/locate/addictbeh

Then and now: Consumption and dependence in e-cigarette users who formerly smoked cigarettes

MARK

Matthew Browne⁎,1, Daniel G. Todd1 Central Queensland University, Australia

H I G H L I G H T S and female e-cigarette users prefer high-nicotine, low-power delivery. • Older nicotine concentration tends to be higher for those who vaped for longer. • E-juice of consumption and dependence is markedly less for vaping than smoking. • Coupling • The Fagerström test does not capture motivations for e-cigarette use.

A R T I C L E I N F O

A B S T R A C T

Keywords: Vaping E-cigarettes Nicotine Dependence Consumption

Electronic cigarette use, or vaping, continues to be a focus for regulators and policy makers in public health, particularly since it can compete with or be a substitute for smoking. This study investigated characteristics of nicotine dependence and consumption in a sample of vapers who formerly smoked cigarettes. We recruited 436 (80% male) vapers from several internet discussion forums; 95% of whom previously smoked, but ceased after commencing vaping. These participants completed a retrospective version of the Fagerström Test for Nicotine Dependence (FTND-R), as well as a version modified to suit current vaping (FTND-V), along with measures of consumption. Nicotine dependence appears to reduce markedly when smokers transition to vaping. However, ‘decoupling’ is observed in the relationship between consumption and dependence in vaping, and the FTND-V showed inadequate psychometric properties. Older and female vapers tend to employ a low-power, higher nicotine-concentration style of vaping. Overall, nicotine concentration tended to increase over time, although this effect was moderated by users' intentions to reduce their intake. Indicators of smoking addiction do not appear to be applicable to vaping, with respect to both internal consistency and relationship to consumption. This suggests that motivations for vaping are less dominated by nicotine delivery (negative reinforcement), and may be driven more by positive reinforcement factors. Nevertheless, e-liquid nicotine concentration was associated, albeit weakly, with dependence among e-cigarette users. Finally, vapers are heterogeneous group with respect to style of consumption, with a high-power/lower nicotine set-up more common among younger men.

1. Introduction Although originally designed to emulate the experience of smoking, ‘vaping’ electronic cigarettes (e-cigarettes) is qualitatively different in a number of respects. Most obviously, vaping involves the inhalation of a vapour fog, rather than smoke; with scope for a very broad range of flavouring options (Measham, O'Brien, & Turnbull, 2016). Whilst tobacco smoke is intrinsically an irritant, the ‘throat-hit’ of e-liquid can be adjusted by varying the proportion of vegetable glycerine and propylene glycol in the e-liquid (Etter, 2016). This customisation can be combined with a choice of nicotine concentration, and device power ⁎

1

settings; yielding a variety of potential user experiences. E-cigarettes yield no ash, and very little residual odour – potentially resulting in a more pleasant subjective experience, especially for novice users or those affected by second-hand vapour (Dawkins & Corcoran, 2013). Additionally, vaping devices can be triggered conveniently at will – accommodating a range of usage patterns. Vaping is generally very positively perceived by users; as a healthy alternative to the increasing burdens of smoking, and as being more acceptable for consumption in the home or in social situations (Keane, Weier, Fraser, & Gartner, 2016) the public health literature is somewhat polarised on the question of whether vaping is a beneficial or dangerous

Corresponding author at: Bundaberg Campus, B8 G.47 University Drive, Branyan, QLD 4670, Australia. E-mail address: [email protected] (M. Browne). Current address: School of Health, Medical and Applied Sciences, Central Queensland University, Australia.

http://dx.doi.org/10.1016/j.addbeh.2017.07.034 Received 4 May 2017; Received in revised form 14 July 2017; Accepted 26 July 2017 Available online 28 July 2017 0306-4603/ © 2017 Elsevier Ltd. All rights reserved.

Addictive Behaviors 76 (2018) 113–121

M. Browne, D.G. Todd

social phenomena. It has been described as having great potential for public health benefit – these writers focus on the diversion of individuals who would otherwise smoke cigarettes, and point to the lack of evidence for any significant health impacts from vaping (Hajek, 2014). Others, including the World Health Organisation, label vaping as a threat to public health, citing similarly the lack of evidence regarding safety of long-term inhalation of various chemical flavourings, and the potential for uptake among non-smokers and the young (Sim & Mackie, 2014) Central to this controversy is the open question of whether or not vaping acts to increase or reduce nicotine addiction (Palazzolo, 2013) Despite conflicting results in the literature, research suggests that vaping is reasonably effective at delivering nicotine for at least some users (Dawkins & Corcoran, 2013; Etter & Bullen, 2011a,b; Nides, Leischow, Bhatter, & Simmons, 2014; Vansickel & Eissenberg, 2013). Increased nicotine uptake might be accomplished by the use of higher power devices and/or higher concentrations of nicotine in e-liquid (Farsalinos, Romagna, Tsiapras, Kyrzopoulos, & Voudris, 2013); and one study has shown that acute administration of e-cigarettes to smokers increases blood plasma nicotine levels, and decreases self-reported cigarette cravings (Dawkins & Corcoran, 2013; Vansickel & Eissenberg, 2013). However, vaping delivers nicotine less efficiently and more slowly than smoking (Eissenberg, 2010; Vansickel, Weaver, & Eissenberg, 2012). Vaping may act as an effective substitute for smoking for at least some smokers (Barbeau, Burda, & Siegel, 2013; Polosa, Caponnetto, Cibella, & Le-Houezec, 2015; Polosa et al., 2011), with most users reporting successful cigarette cessation (Goniewicz, Lingas, & Hajek, 2013). The above points lead to ambiguity as to whether vaping has similar or less potential for dependence than smoking – or indeed whether motivations for; and patterns of, consumption are simply qualitatively different than smoking. Hitherto, relatively little research has attempted to measure consumption and dependence among vapers; or to understand changes in dependence after transitioning from smoking, and over extended use. The lack of an established measure of vaping dependence – that is comparable to that associated with cigarettes – presents a significant obstacle to resolving these questions.

stronger positive reinforcement, whilst smoking was associated with greater negative reinforcement.

1.1. Measuring vaping dependence

1.3. Vaping consumption and demographic correlates

Measurement of vaping dependence has generally been done through modifying existing measures for smoking. The Fagerström Test for Nicotine Dependence (Heatherton, Kozlowski, Frecker, & Fagerström, 1991) (FTND) is probably the most widely used measure for smoking addiction. The FTND is subject to psychometric limitations, including suboptimal reliability, and some evidence for lack of unidimensionality (Breteler, Hilberink, Zeeman, & Lammers, 2004; Haddock, Lando, Klesges, Talcott, & Renaud, 1999). Nevertheless, use of the FTND is often recommended on the basis of substantial prior research, its ability to predict relapse, and its brevity (Piper, McCarthy, & Baker, 2006). Two studies have reported using modifications of the FTND for vaping (Etter, 2015; Etter & Eissenberg, 2015). However, neither the specific item modifications nor the psychometric properties of the modified scale are described. One study (2013) employed modified versions of certain items of the FTND (e.g. ‘time to first vape’), but only conducted single-item comparisons. At the time of data collection for the present study, no validated measure of vaping dependence had been reported., Foulds et al. (2015) recently describe a new 10 item Penn State [Electronic] Cigarette Dependence Index (PS ECDI), including both vaping and smoking variants. The PS ECDI combines probes from the FTND along with items from other sources (Bover, Foulds, Steinberg, Richardson, & Marcella, 2008; Fidler, Shahab, & West, 2011). However, measures of reliability were not reported. Morean and L'Insalata (2017) present a detailed psychometric evaluation of a fourfactor vaping questionnaire designed to measure expectancies around vaping. They found that among dual users, vaping was associated with

Vaping consumption can be measured in several ways, including numbers of puffs or volume of e-juice consumed per day and nicotine strength of e-juice (Dawkins et al., 2013). Differences in patterns of consumption have also been noted in qualitative studies; for instance ‘cloud chasing’ where user employ high powered devices and low or zero nicotine e-juice to produce large clouds of vapour. For younger people, such demonstrative activities, as well as novel flavour combinations, may be a more important motivator than smoking cessation and nicotine consumption (Measham et al., 2016). On the other hand, Goniewicz et al. (2013) concluded that the amount of nicotine delivered was a key factor that determined patterns of e-cigarette use. Further, they found that most used their device within 30 min of waking up, and that consumption followed patterns of prior cigarette smoking. E-cigarette use appears to be more prevalent among younger males (Kenne et al. 2016), and has been correlated with heavy drinking among college students (Littlefield et al. 2015).

1.2. Vaping and nicotine dependence There is indirect evidence that nicotine dependence plays a role in the initiation and continuation of e-cigarette use. Dawkins, Turner, Roberts, and Soar (2013) found that only 1% of vapers employ zeronicotine e-liquid, and 83% describe themselves as ex-smokers. E-liquid containing nicotine, and higher powered device settings, tend to attenuate withdrawal symptoms among smokers to a greater degree (Caponnetto et al., 2013; Etter, 2015; Etter & Bullen, 2011a, 2011b). This is congruent with reports that latest generation, high-powered devices are more effective at increasing plasma nicotine levels (Farsalinos et al., 2014; Vansickel & Eissenberg, 2013). Nevertheless, nicotine dependence may reduce in individuals who transition from cigarettes to vaping. Dawkins et al. (2013) found that time from waking before first use was significantly longer than when smoking. Foulds et al. (2015) found a large overall reduction in all dependence indicators when comparing retrospective smoking to current vaping. Results from other studies also imply lower levels of dependence among ex-smoking vapers, apparently due to less effective nicotine delivery (Farsalinos et al., 2013, 2015; Vansickel et al., 2012). Most vapers rate their dependence as weaker than their prior dependence on cigarettes, but those who vape with nicotine tend to have a higher degree of dependence, and are less likely to intend to stop vaping (Etter & Eissenberg, 2015). Much less is known about changes in dependence among vapers over time. Dawkins et al. (2013) found that only about one third of vapers were attempting to cut down their e-cigarette use. However, one study on dual-users found a 31% abstinence rate at 6 months (Siegel, Tanwar, & Wood, 2011). A trend towards reduction in e-liquid nicotine strength over time has also been reported (Polosa et al., 2015). In their narrative review, Rahman, Hann, Wilson, and Worrall-Carter (2014) conclude that vaping can either perpetuate or attenuate nicotine addiction, depending on whether the user is motivated to quit.

1.4. Summary and aims A consensus appears to be emerging that e-cigarette use is associated with lower levels of dependence than smoking (Etter & Eissenberg, 2015; Farsalinos et al., 2013; Foulds et al., 2015; Goniewicz et al., 2013). However, this evidence is based on either single report items, or vaping dependence measures without psychometric validation – a recognised barrier to progress (Etter, 2015; Etter & Eissenberg, 2015). It is also not clear whether dependence among vapers might increase or decrease over time, and whether or not 114

Addictive Behaviors 76 (2018) 113–121

M. Browne, D.G. Todd

language vaping forums (47.9%) and other online social vaping forums (Facebook 34.4%, Twitter, 12.4%, Other 5.3%). The survey was delivered online via the SurveyMonkey platform, which included an initial information screen to provide informed consent.

this is mediated by user intentions. There is some evidence that device power and nicotine e-liquid concentration may be associated with increased dependence, which needs to be confirmed. Also, prior work has not accounted for dependencies between measures of vaping consumption. Anecdotal evidence suggest that there is significant variability in vaping ‘style’, on a spectrum between high volume/low nicotine concentration (e.g. ‘cloud chasers’, employing direct lung inhalation) versus more traditional low volume/high nicotine consumption (associated with mouth to lung inhalation). Our aim is to better understand e-cigarette dependence and consumption among ex-smoking vapers in a cross sectional survey. Based on prior studies, we expected to observe:

2.4. Analysis

compared to smoking.

All analysis was done in the R statistical programming environment (Team, 2013). Confirmatory factor analyses (CFA) were done using the lavaan package (Rosseel, 2012), specifying ordered logistic links for all binary and ordinal measures that did not have a normal distribution. Classical coefficient alpha was used to describe scale reliability, along with coefficient omega, which also captures scale unidimensionality (Zinbarg & Alden, 2015). Standardised Root Mean Square Residual (SRMR) was used to compare CFA models. Spearman correlations were calculated due to non-normality of most variables. Wilcoxon nonparametric t-tests were used to compare FTND repeated-measures across the two forms.

We also aim to explore:

3. Results

• That the large majority of vapers surveyed will be ex-smokers and a relatively small percentage will be dual-users. • A reduction in nicotine dependence in vaping compared to smoking. • A reduced relationship between dependence and consumption • The psychometric properties of a dependence measure (the FTND) applied to vaping. • Correlates of vaping consumption and dependence: demographics,

The average participant smoked 26.3 CPD prior to vaping, and 0.28 currently, used a nicotine concentration in their e-juice of 7–8 mg/ml, and vaped around 5 ml of e-juice per day. Reasons (non-exclusive) given for vaping included: health benefits (74.1%), other NRTs ineffective (45.2%), more enjoyable (35.1%) and for pleasure (22.7%), less offensive to others (31.8%), and being easier than quitting nicotine completely (35.1%). For brevity, hereafter we employ abbreviations for FTND items, e.g. ‘wake’ – ‘How soon after you wake up do you have your first vape?’. See Table 1 for a key to abbreviations.

duration of use, and intention to quit.

2. Material and methods 2.1. Participants Four hundred and thirty-six (350, 80% male) current vapers were recruited, aged between 17 and 88, with a mean of 41.4 (SD = 13.1). Most had completed some higher education or vocational training (77.3%), and a large majority was engaged in some form of paid work or study (87.4%). Only 4 participants (0.09%) reported not smoking prior to commencing vaping. Twenty-two (5.0%) were ‘dual users’, reporting some degree of present cigarette use concurrently with vaping. The mean age of smoking commencement was 15.7 years (SD = 3.4). Participants had been vaping for an average of 2.1 years (SD = 1.5) and had been smoking for 26.0 years (SD = 13.0). They resided in the United Kingdom (24.1%), Australia (19.7%), Finland (18.3%), Ireland (10.6%) and the United States (4.8%), with the remaining 22.5% located primarily in other Western countries.

3.1. Comparison of FTND addiction probes Fig. 1 compares sample mean responses for FTND dependence probes for retrospective smoking (FTND-R) and current vaping (FTNDV) forms. Wilcoxon non-parametric t-tests confirmed that mean responses on all FTND-V probes were significantly less than their FTND-R counterparts (p < 0.001), with the largest effect size observed for ‘first’ (Ref. Table 1). Additionally, there were significant positive correlations (p < 0.001) between participants' responses on all corresponding items on the FTND-R and the FTND-V: Spearman's rho/ Pearons phi coefficients ranged between 0.16 (hate) and 0.31 (first). 3.2. Reliability of the FTND-R and FTND-V

2.2. Assessments and measures The FTND-R yielded near-to-adequate fit reliability characteristics (alpha = 0.69, omega = 0.68, SRMR = 0.054), whilst the FTND-V had a lower – and inadequate – degree of internal consistency (alpha = 0.54, omega = 0.43, SRMR = 0.089).

We employed two variants of the FTND; for either retrospective smoking (FTND-R) or current vaping (FTND-V). Retrospective assessment of nicotine dependence using the FTND has been shown to have acceptable test-retest reliability (Hudmon, Pomerleau, Brigham, Javitz, & Swan, 2005). For the FTND-R, participants were directed to ‘think back to the time just before they started vaping’. Table 1 compares the customised FTND items used. Given that vaping involves qualitatively different indices of consumption, these were treated separately in analyses. Hereafter, FTND-V and FTND-R will refer to the common behavioural indicators (e.g. excluding cigarettes per day; CPD). We employed a single item, reduce, to assess intentions regarding nicotine dependence – After you started vaping, did you try to reduce your nicotine intake? {0 = No, 1 = Yes}. Other single item probes, e.g. reasons for starting vaping, are mentioned in the results section.

3.3. Bivariate and multivariate effects We calculated scale sums for the FTND-R and the FTND-V behavioural indicators. Table 2 presents bivariate correlations between demographics, measures of dependence and substance consumption for prior smoking and current vaping. 3.3.1. Coupling of consumption and dependence measures CPD was strongly related to the FTND-R scale sum (r = 0.58). However, the FTND-V had a weak correlation with e-liquid nicotine concentration (r = 0.14), and was not significantly associated with other indices of vaping consumption. Structural relationships between e-liquid (volume consumption), (coil) resistance and watts (typically vaped at) were apparent. For example, those using a lower resistance coil vaped at higher watts (r = 0.78), whilst higher nicotine concentrations were associated with lower-power vaping, and less e-liquid consumption.

2.3. Procedure Ethical approval was provided by the institutional Human Research Ethics Committee, #H15/05-112. Adults who were regular current vapers were recruited initially via posting invitations to online English 115

116

Do you smoke more during the first hours after waking than during the rest of the day? Do you smoke even when you are ill enough to be in bed most of the day? How many cigarettes per day do you smoke?

First

What resistance coil do you usually use? What wattage do you usually vape at?

Watts

How much e-liquid per day do you vape (on average)? What nicotine concentration do you usually use in your e-liquid?

Do you vape more during the first hours after waking than during the rest of the day? Do you vape even when you are ill enough to be in bed most of the day?

[1] The first one in the morning [0] Any other [1] In the morning, after I wake up [0] Any other time [1] Yes [0] No

When would you find it most difficult not to vape?

[0] 10 or less [1] 11–20 [2] 21–30 [3] 31 or more [0] 1–3 ml [1] 3–5 ml [2] 5–8 ml [3] 8 ml or more [0] Zero nicotine [1] 1–5 mg/ml [2] 6–7 mg/ml [3] 8–10 mg/ml [4] 10–14 mg/ml [5] 15–24 mg/ml [6] 25 mg/ml or greater b [4] < 0.2 Ω [3] 0.2–0.4 Ω [2] 0.4–1.0 Ω [1] 1.0–1.5 Ω [0] Above 1.5 Ω [NA] I don't know [0] 10 W or less [1] 10-20 W [2] 20-30 W [3] 30-50 W [4] Above 50 W [NA] I don't know

[1] Yes [0] No

a

[3] Within 5 min [2] Within 6–30 min [1] Within 31–60 min [0] After 60 min [1] Yes [0] No

Response [scoring]

How soon after you wake up do you have your first vape? Do you find it difficult to refrain from vaping in places where it is forbidden (e.g., in an airport, at the library, at the cinema, etc)?

FTND-V

Resist

Nicotine

E-liquid

CPD

Ill

Notes: Underlined text denotes differences between original and altered forms of the FTND. a Altered response for FTND-V ‘hate’. b Response > 3 on ‘resist’ imply ‘sub-ohming’, a high power device style associated with high watts and high vapour production.

Substance consumption

Did you smoke more during the first hours after waking than during the rest of the day? Did you smoke even when you were ill enough to be in bed most of the day? How many cigarettes per day did you smoke?

Which cigarette would you hate most to give up?

Refrain

How soon after you wake up do you smoke your first cigarette? Do you find it difficult to refrain from smoking in places where it is forbidden (e.g., in church, at the library, in cinema, etc)?

FTND-R

Hate

Wake

Behavioural indicator of addiction

FTND How soon after you woke up would you smoke your first cigarette? Did you find it difficult to refrain from smoking in places where it was forbidden (e.g., in church, at the library, in cinema, etc)? Which cigarette would you have hated most to give up?

Abbreviation

Category

Table 1 Comparison of the FTND and corresponding items in retrospective and vaping forms.

M. Browne, D.G. Todd

Addictive Behaviors 76 (2018) 113–121

Addictive Behaviors 76 (2018) 113–121

M. Browne, D.G. Todd

(a)

Fig. 1. Item-by-item comparison of FTND dependence probes between retrospective and current vaping forms.

(b) 1.0

3.0

2.5 0.8

0.6

FTND Form

Proportion

Likert Response

2.0

1.5

Retrospective Smoking Current Vaping 0.4

1.0

0.2 0.5

wake

first

hate

ill

refrain

FTND Probe

3.3.4. Multivariate regression of vaping consumption We treated all four vaping consumption characteristics as a joint multivariate (MV) response and considered four stepwise MV regression models with progressive additional predictors (A) demographics only, (B) + prior smoking history, (C) + years vaped, (D) + years vaped ∗ intention to reduce their nicotine intake (interaction). All successive model comparisons were significant at the p < 0.001 level, although the stronger increases in model fit were observed for models (A vs null) and (D vs C). We then conducted univariate analyses on each the four vaping consumption variables separately. Table 3 compares standardised beta coefficients for main effects only (columns 1–4) and interaction models (columns 5–8). Increasing age was associated with a higher nicotine, lower power vaping style. Females tended to vape at a lower power, with only a slightly increased nicotine concentration. Of the prior smoking characteristics, only CPD appeared to be relevant: being related to a higher volume e-liquid use, but not higher power settings. Overall, and controlling for age, participants who had been vaping for longer tended to report the use of a higher nicotine concentration. However, by comparing models (3) and (7), it can be seen that this effect was different for the 70% of participants who reported

3.3.2. Attempts to reduce nicotine dependence Most (70.4%) participants reported that they had tried to reduce their nicotine dependence whilst vaping, and this was negatively related to nicotine concentration employed (r = − 0.40).

3.3.3. Demographic effects and vaping style Older vapers tended to vape at higher nicotine concentrations (r = 0.42), using lower power device settings (watts r = − 0.50, resist r = −0.39). However, there was only a small negative relationship between age and e-liquid consumption (r = − 0.16). Women showed a similar vaping profile, typically using a slightly higher nicotine concentration (r = 0.17), and using less e-liquid (r = −0.27), and lower power device settings (watts r = − 0.34, resist r = − 0.30). At the time they transitioned to vaping, older individuals tended to consume more CPD (r = 0.33) and score more highly on the FTND-R (r = 0.22). Participants who had been vaping longer also tended to employ higher nicotine (r = 0.28) and lower power device settings (watts r = − 0.16, resist r = − 0.15). However, this relationship may be partially due to confounding with age, as older participants tend to have been vaping longer (r = 0.32). Table 2 Spearman correlations between prior smoking and current vaping characteristics.

Female Age Smoke age CPD FTND-Ra FTND-Va Nicotine E-liquid Resist Watts Reduce Years vaped

Female

Age

– 0.20⁎ 0.21⁎⁎

–

Smoke age

0.33⁎⁎ 0.22⁎⁎ ⁎⁎

0.17 − 0.27⁎⁎ − 0.30⁎⁎ − 0.34⁎⁎

⁎⁎

0.42 − 0.16⁎⁎ − 0.39⁎⁎ − 0.50⁎⁎ − 0.14⁎ 0.32⁎⁎

CPD

FTND-Ra

FTND-Va

– 0.21⁎⁎ 0.11

– 0.14⁎

Nicotine

E-liquid

Resist

– 0.39⁎⁎ 0.49⁎⁎

– 0.78⁎⁎

Watts

Reduce

– 0.10 − 0.15⁎

– −0.15⁎

– − 0.11

– 0.58⁎⁎ ⁎

0.15 0.16⁎⁎ − 0.13⁎

−0.10

0.16⁎⁎

0.15⁎

0.10

Note: N = 436. Correlations below 0.10 suppressed. All listed correlations significant at 0.05 level. ⁎ p < 0.01. ⁎⁎ p < 0.001. a Sum of behavioural indicators, excludes CPD or vaping consumption measures.

117

– − 0.31⁎⁎ − 0.40⁎⁎ − 0.44⁎⁎ − 0.40⁎⁎ 0.28⁎⁎

− 0.16⁎⁎

Addictive Behaviors 76 (2018) 113–121

M. Browne, D.G. Todd

Table 3 Standardised beta coefficients and model summaries for vaping consumption: comparison of interaction versus main effects model for years vaping.

IV

Main effect models

Interaction models

DV

DV

E-liquid

Watts

(1)

(2) ⁎⁎⁎

Age Female FTND-R CPD Years vaped

⁎⁎⁎

Nicotine

Resist

(3)

(4) ⁎⁎⁎

E-liquid (5) ⁎⁎⁎

− 0.186 (0.050) − 0.233⁎⁎⁎ (0.046) 0.013 (0.055) 0.234⁎⁎⁎ (0.057) − 0.060 (0.047)

− 0.443 (0.045) − 0.237⁎⁎⁎ (0.041) 0.001 (0.050) 0.028 (0.052) − 0.030 (0.042)

0.358 (0.048) 0.079⁎ (0.044) −0.0001 (0.053) 0.001 (0.055) 0.184⁎⁎⁎ (0.045)

−0.317 (0.048) −0.231⁎⁎⁎ (0.044) −0.027 (0.054) 0.057 (0.055) −0.071 (0.045)

0.138 0.126 11.439⁎⁎⁎ (df = 6; 429)

0.299 0.289 30.511⁎⁎⁎ (df = 6; 429)

0.216 0.205 19.659⁎⁎⁎ (df = 6; 429)

0.194 0.183 17.247⁎⁎⁎ (df = 6; 429)

Reduce Years vaped × Reduce R2 Adj. R2 F

Watts (6) ⁎⁎⁎

− 0.179 (0.050) − 0.233⁎⁎⁎ (0.046) 0.014 (0.056) 0.236⁎⁎⁎ (0.057) − 0.053 (0.047) 0.006 (0.046) 0.081⁎ (0.044) 0.145 0.129 9.042⁎⁎⁎ (df = 8; 427)

⁎⁎⁎

− 0.431 (0.045) − 0.237⁎⁎⁎ (0.041) 0.0002 (0.050) 0.030 (0.051) − 0.020 (0.042) 0.029 (0.042) 0.088⁎⁎ (0.040) 0.308 0.295 23.807⁎⁎⁎ (df = 8; 427)

Nicotine

Resist

(7)

(8) ⁎⁎⁎

0.300 (0.044) 0.089⁎⁎ (0.040) 0.027 (0.049) 0.004 (0.050) 0.129⁎⁎⁎ (0.041) − 0.316⁎⁎⁎ (0.040) − 0.149⁎⁎⁎ (0.039) 0.344 0.332 28.040⁎⁎⁎ (df = 8; 427)

− 0.309⁎⁎⁎ (0.049) − 0.231⁎⁎⁎ (0.044) − 0.028 (0.054) 0.058 (0.055) − 0.064 (0.046) 0.019 (0.045) 0.059 (0.043) 0.199 0.184 13.228⁎⁎⁎ (df = 8; 427)

Note: Significant (p < 0.05) coefficients in bold font. ⁎ p < 0.1. ⁎⁎ p < 0.05. ⁎⁎⁎ p < 0.01.

predictors as that reported in section 3.3.4 above. Only the inclusion of prior smoking dependence, yielded a significant improvement in model fit, FTND-R standardised β = 0.403 , p < 0.001, although this model explained only 5.1% of variance in FTND-V. In a separate regression, we considered the joint ability of vaping consumption measures (e-liquid, watts, nicotine, resist) to explain variance in FTND-V. Although eliquid (β = 0.247 , p = 0.003) and nicotine (β = 0.250 , p = 0.002) were significant predictors of dependence, only 4.4% of variance was explained. Note that the unreliability of the FTND-V – indicating a high level of error in measuring the underlying construct – would contribute to these marginal associations with vaping consumption and other factors.

Nicotine concentration in e-liquid

25 mg/ml+

15-18 mg/ml

reduce No Yes 10-14 mg/ml

4. Discussion 8-10 mg/ml

We shall organise our discussion with respect to the study hypotheses and aims.

6-7 mg/ml

4.1. Retrospective and current cigarette use by vapers

0

2

4

6

The vast majority of our sample of vapers was ex-smokers, who generally used vaping as a safer alternative to cigarettes. This is in line with previous findings: Hajek, Etter, Benowitz, Eissenberg, and McRobbie (2014) summarise recent studies that showed that the proportion of never-smokers who have ‘ever used’ an e-cigarette ranged from 0.1 to 3.8% (median 0.5%), and use within the last 30 days between 0 and 2.2% (median 0.3%). Similarly, in an internet survey of vaping forums (N = 3587), Etter and Bullen (2011a,b) found only 0.3% nominating as never-smokers. Furthermore, a large representative study in the UK found that of never-smokers (N = 5866), 0.4% had tried e-cigarettes previously and 0.1% were current users (Dockrell, Morrison, Bauld, & McNeill, 2013). However, studies focused on adolescent vapers suggests that rates of dual-use and never-smoked may be higher in certain subpopulations (Dutra & Glantz, 2014). Average CPD among ex-smoking vapers in our sample was retrospectively reported to have dropped dramatically from 15.72 to 0.28

8

Years vaping

Fig. 2. Linear fit lines between years vaping and nicotine concentration, by intention to reduce nicotine dependence. Shaded areas indicate 95% confidence interval.

attempting to reduce their nicotine dependence. Fig. 2 illustrates the moderating effect of intention on the relationship between years spent vaping and nicotine concentration. The best performing model was (7), with 34.4% of variance in nicotine concentration explained. 3.3.5. Regression on vaping dependence Given the low reliability of the FTND-V, it is not strictly advisable to interpret an analysis of the scale sum. However, with this strong caveat in mind, it is useful to compare the predictability of the FTND-V with the FTND-R. We conducted a stepwise regression with the same 118

Addictive Behaviors 76 (2018) 113–121

M. Browne, D.G. Todd

dependence/motivation, rather than consumption. This is congruent with the idea that vaping consumption is driven by reward-approach motivation, rather than dependence or negative reinforcement (Morean & L'Insalata, 2017). Nevertheless, though the effect size was small, in agreement with prior results (2015), nicotine concentration and e-liquid consumption were uniquely associated with increased dependence. This suggests that negative reinforcement and nicotine dependence may play a role, albeit minor, in motivating e-cigarette use.

after commencing vaping, with the vast majority reporting zero current cigarette use. However, our sample probably over-represents vaping ‘enthusiasts’, and under-represents casual or semi-committed vapers, and some might wish to minimise their cigarette consumption so as to present vaping in a more favourable light. Nevertheless, this marked decrease in cigarette consumption among vapers is congruent with prior findings (2013). 4.2. Vaping dependence compared to smoking

4.5. Variability in vaping style In an item by item comparison of retrospective and current FTND probes, we found a strong and consistent reduction in dependence after transitioning to vaping. This is consistent with findings of Farsalinos et al. (2013) and Foulds et al. (2015), who employed similar items within the 10-item PS CDI. Not surprisingly, those participants with a stronger retrospective dependence on cigarettes, also appeared to be more dependent on vaping. However, inference based on this result depends on the assumption of an ‘apples to apples’ comparison of the FTND, which is discussed next.

To our knowledge this is the first quantitative study to measure variability in the style of vaping consumption, and to link this to demographic characteristics. The strong negative relationship between quantity (e-liquid, resist, watts) and concentration (nicotine), suggests that it is important to distinguish vaping ‘style’, i.e. higher/lower nicotine combined with lower/higher device settings; from nicotine consumption, which is potentially positively indicated by a combination of both quantity of vapour and concentration. Women and older vapers tend to employ a lower power, higher nicotine consumption profile. Given that older vapers tend to have been vaping longer, it is tempting to conclude that their use of the lower power style is due to having adopted vaping earlier, and thereby being accustomed to less powerful, earlier generation devices. However, since we controlled for years vaped, there appears to be a distinct unique effect for age. This may be because more older, more experienced smokers are more accustomed to a mouth-to-lung inhalation style associated with lower vapour production. This is congruent with our impressions of the online vaping community, that appears to comprises of at least two distinct groups. Firstly, a younger, mostly male, demographic that is attracted to a high power vaping style (popular terms include ‘cloud chasing’, ‘sub-ohming’) using more powerful devices (‘box-mods’) and direct lung inhalation. The second group includes older men and women who employ smaller, more cigarette-similar devices, with lower power settings.

4.3. Implications of the psychometric inadequacy of the FTND-V This is the first study to check the psychometric validity of a FTNDsimilar dependence measure applied to vaping. Our analyses suggest inadequate psychometric properties of the FTND-R, which limits the strength of inference regarding reduction in dependence from smoking to vaping, and potentially also prior results derived from the FTND or related smoking dependence measures (Etter, 2015; Etter & Eissenberg, 2015; Foulds et al., 2015). This is a particularly salient point, given that reliability analysis was not reported for the PS ECDI (Foulds et al., 2015), which employs FTND variants. Whilst ‘cigarette like’ dependence appears to be considerably lower among vapers, it is also possible that vaping reflects different patterns of motivation that is inadequately captured by cigarette-similar probes. We speculate that the lack of reliability of the FTND-V may be due to unsuitability of item content for vaping, which would explain both the low mean scores, and the lack of inter-correlations between items. To illustrate, whilst smoking while sick is arguably inherently somewhat aversive – and therefore a useful threshold for detecting dependent individuals, vaping while sick may be inherently less unpleasant – and therefore presents no such hurdle. Similar arguments can be made regarding vaping versus smoking inside or bed, and in close proximity to others. Morean and L'Insalata's (2017) finding of different expectancies for vaping (positive reinforcement) and smoking (negative reinforcement) may also provide an explanation for the poor psychometric performance of the FTND-V. That is, vaping may be driven by positive motivations, such as enjoyable taste associated with flavourings. Thus, behaviours captured by the FTND-V may be a poor fit to vaping motivations because they tend to measure negative, rather than positive reinforcement.

4.6. Dependence and length of time vaping Foulds et al. (2015) found that dependence tended to increase with the number of years vaped. In our study, dependence did not significantly increase with number of years vaped, although the observed effect was in the positive direction. However, we did find that nicotine concentration tended to increase overall over time, although this was heavily moderated by whether the user intended to reduce their nicotine intake. This result is contrary to results reported by Polosa et al. (2011), but consistent with conclusions made in a review article by Rahman et al. (2014, p. 7), that e-cigarettes may ‘either perpetuate or attenuate nicotine addiction, depending on whether users are motivated to quit or just use them recreationally’. Given that around ¾ of vapers consider vaping to be healthier than cigarettes, many individuals may consider there to be little reason to reduce their dependence on nicotine.

4.4. Decoupling of dependence and consumption for vaping compared to smoking

4.7. Limitations Consistent with previous research, we observed a marked decoupling of the relationship between indicators of dependence and consumption-measures for cigarettes and vaping. Whilst CPD was strongly related to the FTND-R (r = 0.58/R2 = 0.34), there was only a slight relationship between indices of vaping consumption and the FTND-V (multiple R2 = 0.04). Foulds et al. (2015) suggests that this may be due to wider variation in nicotine absorption from different e-cigs; i.e., attributing decoupling to poor measurement of vaping consumption. However, our analyses incorporated multiple indices of vaping and nicotine consumption that one would expect to account for much of this variation. Given the poor reliability of the FTND-V, it seems more likely that apparent decoupling may be due to inadequate measurement of

The study is subject to limitations inherent to a retrospective design. We do expect that for ex-smokers, their prior smoking behaviour is quite salient, and should be recalled accurately. Nevertheless, the topic would benefit from more prospective, longitudinal studies. Our retrospective smoking probes were framed relative with respect to the time just before commencing vaping. However, the potential impact of duration of recall on the FTND-V and prior CPD appears to be small (ref. correlations with years vaped in Table 2). Our sample was one of convenience, and was drawn from online forums – thus probably over-representing vaping enthusiasts who would be more likely to have successfully transitioned from smoking to 119

Addictive Behaviors 76 (2018) 113–121

M. Browne, D.G. Todd

Caponnetto, P., Campagna, D., Cibella, F., Morjaria, J. B., Caruso, M., Russo, C., & Polosa, R. (2013). EffiCiency and Safety of an eLectronic cigAreTte (ECLAT) as tobacco cigarettes substitute: A prospective 12-month randomized control design study. PloS One, 8(6), e66317. Dawkins, L., & Corcoran, O. (2013). Acute electronic cigarette use: Nicotine delivery and subjective effects in regular users. Psychopharmacology, 231(2), 401–407. Dawkins, L., Turner, J., Roberts, A., & Soar, K. (2013). “Vaping” profiles and preferences: An online survey of electronic cigarette users. Addiction, 108(6), 1115–1125. Dockrell, M., Morrison, R., Bauld, L., & McNeill, A. (2013). E-cigarettes: Prevalence and attitudes in Great Britain. Nicotine & Tobacco Research, 15(10), 1737–1744. Dutra, L. M., & Glantz, S. A. (2014). High international electronic cigarette use among never smoker adolescents. The Journal of Adolescent Health, 55(5), 595–597. Eissenberg, T. (2010). Electronic nicotine delivery devices: Ineffective nicotine delivery and craving suppression after acute administration. Tobacco Control, 19(1), 87–88. Etter, J.-F. (2015). Explaining the effects of electronic cigarettes on craving for tobacco in recent quitters. Drug and Alcohol Dependence, 148, 102–108. Etter, J.-F. (2016). Throat hit in users of the electronic cigarette: An exploratory study. Psychology of Addictive Behaviors, 30(1), 93–100. Etter, J.-F., & Bullen, C. (2011a). Electronic cigarette: Users profile, utilization, satisfaction and perceived efficacy. Addiction, 106(11), 2017–2028. Etter, J.-F., & Bullen, C. (2011b). Saliva cotinine levels in users of electronic cigarettes. The European Respiratory Journal, 38(5), 1219–1220. Etter, J.-F., & Eissenberg, T. (2015). Dependence levels in users of electronic cigarettes, nicotine gums and tobacco cigarettes. Drug and Alcohol Dependence, 147, 68–75. Farsalinos, K. E., Spyrou, A., Stefopoulos, C., Tsimopoulou, K., Kourkoveli, P., Tsiapras, D., ... Voudris, V. (2015). Nicotine absorption from electronic cigarette use: Comparison between experienced consumers (vapers) and naïve users (smokers). Scientific Reports, 5, 11269. Farsalinos, K. E., Spyrou, A., Tsimopoulou, K., Stefopoulos, C., Romagna, G., & Voudris, V. (2014). Nicotine absorption from electronic cigarette use: Comparison between first and new-generation devices. Scientific Reports, 4, 4133. Farsalinos, Romagna, G., Tsiapras, D., Kyrzopoulos, S., & Voudris, V. (2013). Evaluating nicotine levels selection and patterns of electronic cigarette use in a group of “vapers” who had achieved complete substitution of smoking. Substance Abuse, 139. Fidler, J. A., Shahab, L., & West, R. (2011). Strength of urges to smoke as a measure of severity of cigarette dependence: Comparison with the Fagerström Test for Nicotine Dependence and its components. Addiction, 106(3), 631–638. Foulds, J., Veldheer, S., Yingst, J., Hrabovsky, S., Wilson, S. J., Nichols, T. T., & Eissenberg, T. (2015). Development of a questionnaire for assessing dependence on electronic cigarettes among a large sample of ex-smoking E-cigarette users. Nicotine & Tobacco Research, 17(2), 186–192. Goniewicz, M. L., Lingas, E. O., & Hajek, P. (2013). Patterns of electronic cigarette use and user beliefs about their safety and benefits: An internet survey. Drug and Alcohol Review, 32(2), 133–140. Haddock, C. K., Lando, H., Klesges, R. C., Talcott, G. W., & Renaud, E. A. (1999). A study of the psychometric and predictive properties of the Fagerström Test for Nicotine Dependence in a population of young smokers. Nicotine & Tobacco Research, 1(1), 59–66. Hajek, P. (2014). Electronic cigarettes have a potential for huge public health benefit. BMC Medicine, 12, 225. Hajek, P., Etter, J.-F., Benowitz, N., Eissenberg, T., & McRobbie, H. (2014). Electronic cigarettes: Review of use, content, safety, effects on smokers and potential for harm and benefit. Addiction, 109(11), 1801–1810. Harrold, T. C., Maag, A. K., Thackway, S., Mitchell, J., & Taylor, L. K. (2015). Prevalence of e-cigarette users in New South Wales. The Medical Journal of Australia, 203(8), 326. Heatherton, T. F., Kozlowski, L. T., Frecker, R. C., & Fagerström, K. O. (1991). The Fagerström Test for Nicotine Dependence: A revision of the Fagerström Tolerance Questionnaire. British Journal of Addiction, 86(9), 1119–1127. Hudmon, K. S., Pomerleau, C. S., Brigham, J., Javitz, H., & Swan, G. E. (2005). Validity of retrospective assessments of nicotine dependence: A preliminary report. Addictive Behaviors, 30(3), 613–617. Keane, H., Weier, M., Fraser, D., & Gartner, C. (2016). “Anytime, anywhere”: Vaping as social practice. Critical Public Health, 27(4), 465–476. Kenne, D. R., Mix, D., Banks, M., & Fischbein, R. (2016). Electronic cigarette initiation and correlates of use among never, former, and current tobacco cigarette smoking college students. Journal of Substance Use, 21(5), 491–494. Littlefield, A. K., Gottlieb, J. C., Cohen, L. M., & Trotter, D. R. M. (2015). Electronic cigarette use among college students: Links to gender, race/ethnicity, smoking, and heavy drinking. Journal of American College Health, 63(8), 523–529. Measham, F., O'Brien, K., & Turnbull, G. (2016). “Skittles & Red Bull is my favourite flavour”: E-cigarettes, smoking, vaping and the changing landscape of nicotine consumption amongst British teenagers – Implications for the normalisation debate. Drugs: Education, Prevention and Policy, 23(3), 224–237. Morean, M. E., & L'Insalata, A. (2017). The Short Form Vaping Consequences Questionnaire: Psychometric properties of a measure of vaping expectancies for use with adult E-cigarette users. Nicotine & Tobacco Research, 19(2), 215–221. Nides, M. A., Leischow, S. J., Bhatter, M., & Simmons, M. (2014). Nicotine blood levels and short-term smoking reduction with an electronic nicotine delivery system. American Journal of Health Behavior, 38(2), 265–274. Palazzolo, D. L. (2013). Electronic cigarettes and vaping: a new challenge in clinical medicine and public health. A literature review. Frontiers in Public Health, 1, 56. Piper, M. E., McCarthy, D. E., & Baker, T. B. (2006). Assessing tobacco dependence: a guide to measure evaluation and selection. Nicotine & Tobacco Research, 8(3), 339–351. Polosa, R., Caponnetto, P., Cibella, F., & Le-Houezec, J. (2015). Quit and smoking reduction rates in vape shop consumers: A prospective 12-month survey. International

vaping. Therefore, we do not draw inference regarding the efficacy of ecigarettes as a smoking cessation tool from this sample. Males (80% of the sample) were also overrepresented compared to population-representative statistics (see e.g. Harrold, Maag, Thackway, Mitchell, & Taylor, 2015). In contrast to the FTND-R, the FTND-V was found to be unreliable, which limits the inference we can make regarding the dependence construct, nominally measured by the FTND-V. Thus, although we replicated prior findings that dependence is nominally lower for vaping than smoking, the results admit the more limited conclusion that vaping dependence/motivation is simply different from that of smoking. Keeping in mind the quite large item-by-item decreases in each FTNDR/V probe, our view is that a reasonable conclusion is that cigarettesimilar dependence is certainly lower for e-cigarette users – at least on the concrete behaviours captured by the FTND. Ideally, the field would benefit from a dependence measure that was demonstrably valid for both vaping and smoking – enabling a robust comparison of dependence between the two behaviours. However, if vaping behaviour is motivated by different motivations: positive reinforcement rather than negative reinforcement, as suggested above, such a measure may not even be feasible. 5. Conclusions Our results were largely consistent with expectations. First, we found that the large majority of vapers were ex-smokers who had either ceased or dramatically reduced their cigarette consumption. Second, there was a marked decrease in dependence among vapers compared to their retrospective prior cigarette dependence. Finally, we also observed decoupling: a large attenuation of the relationship between dependence and consumption for vapers as compared to their retrospective prior smoking. We incorporated multiple measures of vaping consumption, which showed high variability with respect to a vapour volume/(negative) nicotine concentration continuum, with female and older vapers tending to vape at lower volumes combined with higher nicotine concentrations. However, the lack of reliability and unidimensionality of the FTND-V raise concerns about the adequacy of cigarette-analogous dependence measures for vaping, and whether ‘apples to apples’ comparisons with smoking are strictly valid. Finally, we observed no relationship between dependence or e-liquid volume consumption and duration of vaping. There was a tendency for those who have been vaping longer to employ increased nicotine concentration, but this was moderated by vapers' intentions to reduce their intake. Future research should focus on better measurement of consumption patterns and dependence indices for vaping, and employ these measures in prospective longitudinal designs. Funding sources Research was supported by block funding to the School of Health, Medical and Applied Sciences, Central Queensland University. Declaration of interests No conflicts of interest to declare. References Barbeau, A. M., Burda, J., & Siegel, M. (2013). Perceived efficacy of e-cigarettes versus nicotine replacement therapy among successful e-cigarette users: A qualitative approach. Addiction Science & Clinical Practice, 8, 5. Bover, M. T., Foulds, J., Steinberg, M. B., Richardson, D., & Marcella, S. W. (2008). Waking at night to smoke as a marker for tobacco dependence: Patient characteristics and relationship to treatment outcome. International Journal of Clinical Practice, 62(2), 182–190. Breteler, M. H. M., Hilberink, S. R., Zeeman, G., & Lammers, S. M. M. (2004). Compulsive smoking: The development of a Rasch homogeneous scale of nicotine dependence. Addictive Behaviors, 29(1), 199–205.

120

Addictive Behaviors 76 (2018) 113–121

M. Browne, D.G. Todd

Sim, F., & Mackie, P. (2014). Vape and vitriol in public health. Public Health, 128(10), 869–871. Team, R. C. (2013). R Core Team. 2013. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. ISBN 3-90005107-0 http://www.R-project.org. Vansickel, A. R., & Eissenberg, T. (2013). Electronic cigarettes: effective nicotine delivery after acute administration. Nicotine & Tobacco Research, 15(1), 267–270. Vansickel, A. R., Weaver, M. F., & Eissenberg, T. (2012). Clinical laboratory assessment of the abuse liability of an electronic cigarette. Addiction, 107(8), 1493–1500. Zinbarg, R. E., & Alden, A. R. (2015). Coefficient alpha and coefficient omega hierarchical. The Encyclopedia of Clinical Psychology (pp. 1–6). .

Journal of Environmental Research and Public Health, 12(4), 3428–3438. Polosa, R., Caponnetto, P., Morjaria, J. B., Papale, G., Campagna, D., & Russo, C. (2011). Effect of an electronic nicotine delivery device (e-Cigarette) on smoking reduction and cessation: A prospective 6-month pilot study. BMC Public Health, 11, 786. Rahman, M. A., Hann, N., Wilson, A., & Worrall-Carter, L. (2014). Electronic cigarettes: Patterns of use, health effects, use in smoking cessation and regulatory issues. Tobacco Induced Diseases, 12(1), 21. Rosseel, Y. (2012). Lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48(2), http://dx.doi.org/10.18637/jss.v048.i02. Siegel, M. B., Tanwar, K. L., & Wood, K. S. (2011). Electronic cigarettes as a smokingcessation: Tool results from an online survey. American Journal of Preventive Medicine, 40(4), 472–475.

121