Patterns of awareness and use of electronic cigarettes in Mexico, a middle-income country that bans them: Results from a 2016 national survey

INTRODUCTION

Use of electronic cigarettes (e-cigarettes) is rapidly increasing worldwide, although previous studies of their uptake and use are primarily limited to high-income countries (HICs).^1–3 E-cigarettes are battery-operated devices that produce an aerosol that can contain different levels of nicotine, which is a highly addictive substance and can affect the neurological development of adolescents.^4–7 In addition, e-cigarette liquids contain flavorings that are potentially hazardous, especially for the respiratory system.⁵ Marketing of e-cigarettes often emphasized their being a safe alternative to conventional cigarettes and as an effective method for smoking cessation.^8–11 Nevertheless, studies that have evaluated the effectiveness of e-cigarettes for smoking cessation have produced mixed results.^5,12 Moreover, longitudinal studies in the US, Canada and the UK indicate that e-cigarette use among non-smoker adolescents substantially increases the risk of initiating cigarette use,^13–18 raising concerns that e-cigarettes may promote cigarette smoking among youth who would not have otherwise become smokers. Indeed, a recent longitudinal study of Mexican adolescents found similar results;¹⁹ however, more research is sorely needed to better understand the patterns of use and potential public health impact of e-cigarettes in low- and middle-income countries (LMICs).

Prior research indicates that high-income countries (HICs) with e-cigarette bans have a lower prevalence of awareness (e.g., Canada=40%, Australia=20% in 2011) than in similar countries that allow marketing and sales of e-cigarettes (e.g., USA=73%, UK=54%).^1,20 A similar pattern has been found in upper middle income countries (UMIC), where awareness was found to be lower when e-cigarettes are banned (Mexico=34% in 2012; Brazil=35% in 2013)²¹ compared to when they are not banned (Costa Rica=52.5% in 2015;²² Malaysia=62.0% in 2013²¹). Nevertheless, these levels were significantly higher than in the low-income country of Indonesia (10.9% in 2011).²³ While it is difficult to separate out the different temporal and contextual factors that account for differences across countries, the population-level of awareness may depend both on policy context and the level of economic development.

Correlates of e-cigarette awareness and use have been studied almost exclusively in HICs. A systematic review of 26 studies in HICs indicates that e-cigarettes awareness and trial is associated with higher levels of education,¹ with current use additionally associated with younger age and being male.^{1,20, 24} Across 27 European Union countries, those who lived in urban areas had higher e-cigarette awareness and trial compared to rural inhabitants.³ In general, being a current smoker has been associated with awareness and use of e-cigarettes,³ with even higher levels found among smokers who have attempted to quit in the last year.^2,25 In addition, trial of e-cigarettes appears higher in former smokers than never smokers.^26,27

The few studies to examine these patterns in UMICs have found similar results. A cross-sectional study among university students (Age 19–24) in the UMIC of Romania in 2015 found that e-cigarette trial was associated with being male, a smoker, having friends who experimented with e-cigarettes, and having beliefs that e-cigarettes could help them quit smoking.²⁸ Studies of early adolescents in the UMICs of Argentina²⁹and Mexico³⁰ found that e-cigarette use was associated with cigarette smoking, as well as an array of traditional risk factors for cigarette smoking (e.g., sensation seeking, having friends who smoke). Among the general population in Indonesia a higher prevalence of e-cigarette awareness and use were found among the subgroup with higher wealth in comparison with group with very low wealth (e-cigarette awareness 24.7% vs 1.6% and e-cigarette current use 0.7% vs 0.0% respectively).²³

In Mexico, however, some risk factors that appear unique to e-cigarette use have been identified (i.e., internet ad exposure, “technophilia” or the appeal of electronic devices),³⁰ supporting other research that has found e-cigarettes attract a somewhat different, lower risk population segment than cigarettes.³¹ For example, a longitudinal study of Mexican students found that those who had tried only e-cigarettes were more likely than those who had tried no tobacco product to initiate cigarette smoking,¹⁹ although the risk of initiation was somewhat lower than that found in studies from HICs.³²

To the best of our knowledge, there are no studies that have examined sociodemographic or smoking-related correlates of e-cigarette awareness, trial and use in nationally representative samples from LMICs. Such studies are important because of the potential for e-cigarettes to undermine the significant strengthening of tobacco control policies that the World Health Organization’ Framework Convention on Tobacco Control has promoted around the world over the last decade.⁵ Furthermore, it is important to understand whether strong e-cigarette policies appear effective in staving off their use or have resulted in different patterns of uptake than in HICs, whether from countries with relatively stronger or weaker policies. Such studies can also help identify potential issues with implementing strong e-cigarette policies in LMICs, so that policies that address e-cigarettes and other novel tobacco products are effective.

Study context

The importation, distribution, marketing and sales of e-cigarette are banned in Mexico,³³ as in many countries in Latin America.³⁴ Nevertheless, a 2012 population-based survey of adult smokers in Mexico found that 34% had heard of e-cigarettes and 4% had tried them,²¹ which is similar to other UMICs with strong e-cigarette regulations (e.g., Brazil),²¹ but substantially lower than in HICs. Three years later, the 2015 Global Adult Tobacco Survey (GATS), a nationally representative survey of Mexicans 15 years old and older, found the general population had similar levels of awareness (34%) and trial (5%), with 0.6% of the population reporting current e-cigarette use.³⁵ Estimates were higher in a 2015 representative survey of first-year middle school students (average age=12.5 years) in the three largest cities in Mexico, in which half were aware of e-cigarettes (51%) and 10% had tried them.³⁰ The aim of the current study was to determine the correlates of e-cigarette awareness, trial and use in a large, nationally representative survey of Mexico conducted in 2016. Based on the literature reviewed above, we hypothesized that e-cigarette awareness and use will be higher among current smokers than non-smokers. We also hypothesized that current smokers who tried to quit in the last year will be more likely to currently use e-cigarettes compared to current smokers who had not tried to quit. Finally, we expected that former smokers would be more likely to be aware of and use e-cigarettes compared to never smokers.

METHODS

Sample

Data from this study came from the National Survey of Drugs, Alcohol and Tobacco Use (ENCODAT), which were collected between June and November 2016. A multi-stage, cluster sample design was used to produce representative data at the national and state levels (32 states in Mexico). Sampling strata were constructed for urban and rural census tracts in each state. First, within each stratum, census tracts (Basic Geostatsitical Area or AGEB) were selected with probability proportional to the number of inhabitants. Second, for each AGEB selected, 6 blocks were randomly chosen, again with probability proportional to population size. Third, for each block selected, groupings with less than 80 households were created, then a grouping was randomly selected from which six households were selected using systematic sampling. Within each selected household, one 18 to 65 year old household member was randomly selected, as was one 12 to 17 year old (if there were any).³⁶ With an overall response rate of 74%, a total of 56,877 people (12 to 65 years old) were surveyed face-to-face. All adult participants signed written informed consent, and 12- to 17-year olds provided written assent and their parents provided written consent. Questions on tobacco were drawn from prior national and international surveys.^35,36,37 The protocol was approved by the IRB at the Mexican National Institute of Public Health.

Measurement

Awareness, trial, and current use of e-cigarettes:

E-cigarette awareness was evaluated by asking: “Before today, have you ever heard of e-cigarettes?” (yes/no). Respondents who indicated awareness were asked: “Have you ever, even once, used an e-cigarette?” Those responding, “yes” were categorize as trial users of e-cigarettes. Trial users were also asked: “Do you currently use e-cigarettes on a daily basis, less than daily, or not at all?” Participants who responding “daily or less than daily” were considered to be current users of e-cigarettes. Those who indicated that they currently used e-cigarettes were also asked: “The last time you bought e-cigarettes, where did you buy them?” Response options included: convenience store or supermarket; street vendor; outside the country; internet; other people; never bought; they were a gift.

Independent Variables:

Sociodemographic variables included: age, broken down into different age ranges for adolescents (i.e., 12, 13–15, 16–17) and adults (i.e., 18–24, 25–44 and 45–65 years); sex (male, female); place of residence (urban/rural); highest educational attainment for adult participants and, among adolescents, for the head of household (completed or incomplete primary school; completed secondary school; completed high school; completed university or more). A wealth index was constructed using polychoric principal component analysis of information on household assets (i.e., ownership of the house; house having an independent room for cooking; household possession of landline telephone, TV cable service, computer, DVD, microwave, internet, car), with scores divided into quintiles.^38–41 Conventional cigarette smoking status included four categories for adolescents: current smoker (i.e., reported smoking in the last 30 days); former smoker (i.e., include those who did not smoke in the last month, but smoked regularly in the past); experimental smoker (i.e., reported trial of smoking, but had not smoked tobacco regularly in the past or currently); and never smoker. Adults were classified into five categories: never smokers; former smokers; non-daily smoker; daily smokers who smoke less than five cigarettes; daily smokers who smoke more than five cigarettes. The three categories of current smokers are based on prior research that recognizes the high percentage of nondaily smokers and uses the relatively low median number of cigarettes per day in Mexico to divide daily smokers into roughly equal groups.⁴² Current smokers were also asked if they had attempted to quit in the prior 12 months.

Analysis

Descriptive statistics and all regression models were estimated after adjusting for the sampling design. To ensure the adequacy of estimates, the relative standard error was estimated, which indicated that the estimates shown here were stable (i.e., less than 10% for all variables). Bivariate and adjusted logistic regression models were then estimated, regressing each key outcome (e-cigarette awareness, trial, current use) on study variables, with models stratified by age and smoking status (i.e., adolescents, adult non-smokers, adult smokers). The models were adjusted for the state of residence. In addition, sensitivity analyses were estimated using multinomial logistic regression, where the dependent variable involved three mutually exclusive categories (never used e-cigarettes; tried e-cigarettes, but no current use; current use of e-cigarettes). Multiple Imputation by Chained Equations were conducted to address any potential biases introduced by the exclusion of observations with missing data (3% of adolescents; 0.3% of adult nonsmokers; 0.6% of adult smokers). This method makes multiple predictions for each value, that allows to get accurate standard errors.⁴³ Statistical analyses, were performed using STATA 15.

RESULTS

In the total sample (table 1) of 56,877, 17% were adolescents (n=12,436), 66% were nonsmoker adults (n=36,966), and 17% were adults who currently smoked conventional cigarettes (n=7,347). E-cigarette awareness was reported by 45% of adolescents, 34% of adult nonsmokers, and 54% of adult smokers. Prevalence of e-cigarette trial (Appendix-Figure 1) was 7% for adolescents, 3% for adult nonsmokers and 18% for adult smokers. Our study also found that 1.1% of adolescents, 0.3% of adult nonsmokers and 5% of the adult smokers currently used e-cigarettes.

Table 1.

Characteristics of 12 to 65 year old subsamples of adolescents and adult smokers and nonsmokers, México 2016.

Demographic characteristics	Adolescents (n=12,436)	Adult Nonsmokers* (n=36,966)	Adult smokers** (n=7,347)
Sex
Men	51%	41%	75%
Female	49%	59%	25%
Age groups
12 years	15%
13–15 years	53%
16–17 years	32%
18–24 years		21%	25%
25–44 years		46%	48%
45–65 years		33%	28%
Residence
Rural	28%	23%	14%
Urban	72%	77%	86%
Educationa
Primary or less	38%	32%	27%
Secondary	35%	32%	35%
Technical	19%	25%	27%
University or more	7%	11%	11%
Wealth indexb
First quintile	20%	17%	14%
Second quintile	24%	20%	20%
Third quintile	19%	19%	18%
Forth quintile	20%	21%	21%
Fifth quintile	17%	22%	28%
Smoking status
Never smoker	77%	78%
Experimental smoker (adolescents)	13%
Former smoker	5%	22%
Current smoker (adolescents)	5%
Occasional smoker (adults)			58%
Daily smoker, <5 cigs/day (adults)			16%
Daily smoker, 5+ cigs/day (adults)			25%
Quit attempt last 12 month
No			53%
Yes			47%
Awareness of e-cigarettes	45%	34%	54%
Trial of e-cigarettes	7%	3%	18%
Current use of e-cigarettes	1%	0.3%	5%

^a. Education: For adolescents, education of the head of household was used, with the same categories used as for adults: Primary or less (never attended school, incomplete primary, full primary or incomplete secondary); Secondary (Full secondary or incomplete high school); Technical (Complete Technical/preparatory or equivalent or incomplete superior studies); University or more (Complete University studies or postgraduate studies. For adults, we used the same categories at individual level.)

^b. Wealth index: was represents household assets, with Principal Component Analysis scores used to divide households into quintiles.

^*Non-smoking adults include both ex-smokers and never smokers.

^**Adult smokers include those who currently smoke cigarette

E-cigarette awareness

Models that estimated the correlates of e-cigarette awareness (Table 2) indicated that female adolescents and female adult non-smokers were less likely to have heard about e-cigarettes than their male counterparts. E-cigarette awareness was higher among older adolescents (e.g., 16–17 years vs. 12 years: (AOR = 2.60 95% CI 2.08–3.25), although the opposite pattern was found for both adult samples. Higher e-cigarette awareness was associated with living in urban compared to rural areas, higher education, and higher wealth across all three subsamples.

TABLE 2.

CORRELATES OF E-CIGARETTE AWARENESS, MEXICO 2016.

Demographic characteristics	Adolescents (n=12,065)			Adult nonsmokers (n=36,868)			Adult smokers (n=7,306)
	%	OR 95% CI	AOR 95% CI	%	OR 95% CI	AOR 95% CI	%	OR 95% CI	AOR 95% CI
Sex
Men	49	1.00	1.00	40	1.00	1.00	53	1.00	1.00
Female	42	0.74 [0.66–0.83]	0.74 [0.65–0.84]	30	0.62 [0.57–0.67]	0.73 [0.66–0.81]	58	1.25 [1.06–1.46]	1.10 [0.91–1.33]
Age groups
12 years	29	1.00	1.00
13–15 years	45	1.99 [1.67–2.37]	2.08 [1.70–2.56]
16–17 years	54	2.84 [2.34–3.44]	2.60 [2.08–3.25]
18–24 years				49	1.00	1.00	70	1.00	1.00
25–44 years				34	0.53 [0.48–0.59]	0.57 [0.50-.0.64]	54	0.50 [0.39–0.62]	0.47 [0.37–0.60]
45–65 years				25	0.35 [0.31–0.39]	0.38 [0.33–0.43]	41	0.30 [0.23–0.38]	0.25 [0.19–0.33]
Residence
Rural	26	1.00	1.00	17	1.00	1.00	35	1.00	1.00
Urban	53	3.14 [2.69–3.65]	2.03 [1.71–2.40]	39	3.13 [2.79–3.51]	1.74 [1.55–1.96]	57	2.48 [1.91–3.24]	1.60 [1.15–2.23]
Educationa
Primary or less	35	1.00	1.00	15	1.00	1.00	34	1.00	1.00
Secondary	45	1.51 [1.31–1.74]	1.18 [1.02–1.37]	32	2.6 [2.3–2.93]	1.79 [1.57–2.02]	53	2.16 [1.78–2.62]	1.65 [1.34–2.04]
Technical	57	2.40 [2.02–2.84]	1.56 [1.29–1.89]	49	5.29 [4.61–6.07]	3.04 [2.61–3.55]	68	4.02 [3.13–5.18]	2.80 [2.14–3.67]
University or more	65	3.36 [2.63–4.29]	1.96 [1.48–2.59]	56	7.01 [5.98–8.22]	4.17 [3.50–4.96]	72	4.74 [3.39–6.63]	3.48 [2.38–5.09]
Wealth indexb
First quintile	30	1.00	1.00	17	1.00	1.00	35	1.00	1.00
Second quintile	33	1.16 [0.96–1.41]	1.01 [0.82–1.24]	22	1.43 [1.24–1.64]	1.21 [1.04–1.39]	41	1.29 [0.98–1.68]	1.15 [0.85–1.54]
Third quintile	44	1.82 [1.5–2.19]	1.39 [1.13–1.70]	32	2.37 [2.03–2.77]	1.68[1.42–2.00]	55	2.22 [1.68–2.94]	1.80 [1.34–2.42]
Forth quintile	58	3.26 [2.70–3.93]	2.00 [1.64–2.45]	40	3.31 [2.88–3.80]	1.91 [1.65–2.22]	60	2.76 [2.14–3.57]	1.99 [1.49–2.64]
Fifth quintile	65	4.33 [3.52–5.33]	2.04 [1.62–2.56]	53	5.59 [4.87–6.43]	2.31 [1.98–2.69]	69	4.11 [3.12–5.44]	2.68 [1.98–3.61]
Smoking Status
Never smoker	41	1.00	1.00	31	1.00	1.00
Experimental smoker	57	1.87 [1.56,2,25]	1.76 [1.44,2,16]
Former smoker	70	3.37 [2.44,4.65]	3.17 [2.25,4.46]	46	1.91 [1.74–2.09]	2.03 [1.82–2.27]
Current smoker (adolescents)	59	2.09 [1.62,2.71]	1.83 [1.36,2.46]
Nondaily smoker (adults)							51	1.00	1.00
Daily smoker, <5 cigs (adults)							58	1.32 [1.06–1.64}	1.44 [1.12–1.85]
Daily smoker, 5+ cigs (adults)							59	1.35 [1.11–1.63]	1.72 [1.38–2.15]
Quit attempt last 12 months
No							52	1.00	1.00
Yes							57	1.25 [1.05–1.48]	1.37 [1.15–1.65]

^a. Education: For adolescents, education of the head of household was used, with the same categories used as for adults

^b. Wealth index: was represents household assets, with Principal Component Analysis scores used to divide households into quintiles.

Compared to their never smoker counterparts, adolescent current smokers (AOR = 1.83, 95% CI 1.36 – 2.46), adolescent former smokers (AOR=3.17, 95% CI 2.25 – 4.46), adolescent experimental smokers (AOR=1.76, [1.44–2.16], and adult former smokers (AOR = 2.03, 95% CI 1.82 – 2.27) were more like to have heard about e-cigarettes. Among adult smokers, daily smokers were more likely to have heard about e-cigarettes than nondaily smokers (i.e., daily smoker <5 cigs/day vs nondaily smokers: AOR = 1.44, 95% CI 1.12–1.85; daily smoker 5+ cigs/day vs nondaily smokers: AOR = 1.72, 95% CI 1.38 – 2.15). Finally, adult current smokers who had tried to quit in the last 12 months were more likely to have heard about e-cigarettes than those who had not tried to quit (AOR= 1.37, 95% CI 1.15–1.65).

Trial of e-cigarettes

Results regarding trial of e-cigarettes (Table 3) indicated that female adolescents and female adult non-smokers were less likely to have tried e-cigarettes than males. E-cigarette trial was positively associated with age among adolescents (e.g., 16–17 years vs. 12 years: AOR = 5.47, 95% CI 2.94 – 10.19), but negatively associated with age among adults. Living in urban areas was positively correlated with e-cigarette trial, but only among adolescents. Higher education was positively associated with e-cigarette trial only among adult smokers and nonsmokers. Higher wealth was associated with greater likelihood of e-cigarette trial in all three groups, although adjusted models showed a trend towards a dose-response trend relationship only among adult smokers.

TABLE 3.

CORRELATES OF E-CIGARETTE TRIAL, MEXICO 2016.

Demographic characteristics	Adolescents (n=12,081)			Adult nonsmokers (n=36,874)			Adult smokers (n=7,312)
	%	OR 95% CI	AOR 95% CI	%	OR 95% CI	AOR 95% CI	%	OR 95% CI	AOR 95% CI
Sex
Men	8	1.00	1.00	4	1.00	1.00	17	1.00	1.00
Female	5	0.57 [0.43–0.77]	0.70 [0.51–0.95]	2	0.42 [0.32–0.55]	0.61 [0.45–0.83]	21	1.29 [1.03–1.62]	1.26 [0.97–1.64]
Age groups
12 years	1	1.00	1.00
13–15 years	6	5.41 [3.18–9.22]	4.48 [2.57–7.78]
16–17 years	11	10.86 [6.38–18.44]	5.47 [2.94–10.19]
18–24 years				7	1.00	1.00	34	1.00	1.00
25–44 years				2	0.27 [0.21–0.34]	0.26 [0.20–0.33]	16	0.36 [0.28–0.45]	0.31 [0.24–0.40]
45–65 years				1	0.10 [0.06–0.17]	0.09 [0.05–0.16]	8	0.17 [0.12–0.24]	0.14 [0.09–0.19]
Residence
Rural	3	1.00	1.00	1	1.00	1.00	10	1.00	1.00
Urban	8	2.67 [1.9–3.77]	2.10 [1.44–3.06]	3	2.73 [1.82–4.09]	1.19 [0.79–1.77]	20	2.18 [1.43–3.32]	1.32 [0.84–2.08]
Educationa
Primary or less	6	1.00	1.00	0.4	1.00	1.00	7	1.00	1.00
Secondary	7	1.19 [0.85–1.68]	0.99 [0.68–1.44]	2	5.53 [3.51–8.71]	2.71 [1.68–4.39]	17	2.58 [1.93–3.46]	1.81 [1.33–2.45]
Technical	9	1.62 [1.15–2.28]	1.23 [0.82–1.84]	5	11.52 [7.37–18.02]	3.97 [2.44–6.49]	27	4.78 [3.5–6.54]	3.04 [2.1–4.29]
University or more	7	1.27 [0.86–1.87]	0.92 [0.54–1.59]	5	10.51 [6.21–17.78]	4.97 [2.82–8.76]	28	5.1 [3.55–7.32]	4.34 [2.93–6.44]
Wealth indexb
First quintile	3	1.00	1.00	1	1.00	1.00	9	1.00	1.00
Second quintile	5	1.40 [0.93–2.10]	1.40 [0.87–2.27]	1	1.29 [0.74–2.23]	1.08 [0.62–1.87]	11	1.29 [0.87–1.92]	1.19 [0.79–1.79]
Third quintile	7	2.31 [1.54–3.46]	2.28 [1.40–3.72]	2	1.87 [1.07–3.24]	1.27 [0.69–2.33]	17	2.03 [1.34–3.06]	1.67 [1.09–2.58]
Forth quintile	9	2.84 [1.76–4.57]	2.37 [1.33–4.21]	3	2.85 [1.73–4.7]	1.41 [0.81–2.44]	22	2.88 [1.98–4.2]	2.04 [1.36–3.08]
Fifth quintile	10	3.22 [2.14–4.84]	2.23 [1.30–3.84]	5	4.79 [2.93–7.84]	1.94 [1.10–3.41]	26	3.61 [2.50–5.21]	2.34 [1.55–3.53]
Smoking Status
Never smoker	4	1.00	1.00	2	1.00	1.00
Experimental smoker	15	7.56 [5.13,11.1]	6.67 [4.33,10.27]
Former smoker	28	16.9 [9.3–30.7]	15.0 [7.81,28.80]	6	4.23 [3.31–5.40]	5.17 [3.96–6.74]
Current smoker (adolescents)	30	18.6 [12.7–27.2]	16.23 [10.03,26.27]
Nondaily smoker (adults)							16	1.00	1.00
Daily smoker, <5 cigs (adults)							20	1.23 [0.92–1.67]	1.43 [1.04–1.97]
Daily smoker, 5+ cigs (adults)							22	1.39 [1.1–1.77]	2.07 [1.56–2.73]
Quit attempt, last 12 months
No							17	1.00	1.00
Yes							19	1.14 [0.92–1.42]	1.10 [0.86–1.40]

^a. Education: For adolescents, education of the head of household was used, with the same categories used as for adults

^b. Wealth index: was represents household assets, with Principal Component Analysis scores used to divide households into quintiles.

Compared to adolescent never smokers, current smokers (AOR = 16.23, 95% CI 10.03–26.27), former smokers (AOR = 15.0, 95% CI 7.81–28.80), and experimental smokers (AOR=6.67, 95% CI 4.33–10.27] were more likely to have tried e-cigarettes. For nonsmoker adults, e-cigarette trial was more likely among former smokers than never smokers (OR 5.17, 95% CI 3.96 – 6.74); while among adult smokers the likelihood of trial increased with greater frequency of conventional cigarette use (daily smoker <5 cigs/day vs nondaily smokers: AOR = 1.43, 95% CI 1.04 – 1.97; daily smoker +5 cigs/day vs nondaily smokers: AOR = 2.07, 95% CI 1.56 – 2.73).

Current e-cigarette use

As shown in table 4, female adolescents and female non-smoker adults were less likely to currently use e-cigarettes than their male counterparts (AOR = 0.51, 95% CI 0.30 – 0.85; AOR = 0.32, 95% CI 0.17 – 0.61, respectively). By contrast, female adult smokers were more likely to use e-cigarettes than adult male smokers (AOR = 1.80, 95% CI 1.08 – 3.02). Current use of e-cigarettes was positively associated with age for adolescents (e.g., 14–15 years vs 12 years: AOR = 4.38, 95% CI 1.45 – 13.26) but inversely associated with age among both adult groups (i.e., smokers, nonsmokers).

TABLE 4.

CORRELATES OF CURRENT E-CIGARETTE USE, MEXICO 2016.

Demographic characteristics	Adolescents (n=12,068)			Adult nonsmokers (n=36,874)			Adult smokers (n=7,312)
	%	OR 95% CI	AOR 95% CI	%	OR 95% CI	AOR 95% CI	%	OR 95% CI	AOR 95% CI
Sex
Men	2	1.00	1.00	0.6	1.00	1.00	4	1.00	1.00
Female	0.6	0.40 [0.25–0.64]	0.51 [0.30–0.85]	0.1	0.19 [0.10–0.35]	0.32 [0.17–0.61]	6	1.61 [0.99–2.61]	1.80 [1.08–3.02]
Age groups
12 years	0.1	1.00	1.00
13–15 years	1	8.55 [2.76–24.47]	4.38 [1.45–13.26]
16–17 years	1	10.53 [3.49–31.83]	2.37 [0.75–7.43]
18–24 years				0.8	1.00	1.00	9	1.00	1.00
25–44 years				0.3	0.35 [0.18–0.66]	0.38 [0.19–0.76]	4	0.39 [0.29–0.54]	0.35 [0.24–0.52]
45–65 years				0.1	0.09 [0.03–0.34}	0.09 [0.03–0.34]	2	0.19 [0.10–0.35]	0.15 [0.08–0.27]
Residence
Rural	0.6	1.00	1.00	0.2	1.00	1.00	3	1.00	1.00
Urban	1	2.16 [1.20–3.92]	1.81 [0.96–3.40]	0.4	2.29 [0.79–6.66]	1.02 [0.32–3.24]	5	1.82 [0.99–3.36]	1.43 [0.76–2.67]
Educationa
Primary or less	1	1.00	1.00	0.1	1.00	1.00	2	1.00	1.00
Secondary	1	0.62 [0.33–1.16]	0.58 [0.28–1.19]	0.1	2.77 [1.03–7.44}	1.34 [0.47–3.77]	4	1.84 [1.1–3.09]	1.16 [0.66–2.04]
Technical	2	1.57 [0.77–3.20]	1.24 [0.53–2.89]	0.7	13.37 [5.63–31.73]	5.88 [1.89–18.31]	7	3.79 [2.15–6.71]	1.96 [1.10–3.50]
University or more	1	0.64 [0.24–1.65]	0.51 [0.15–1.74]	0.7	13.42 [5.14–35.01]	8.48 [2.25–31.93]	6	3.02 [1.69–5.41]	1.87 [0.96–3.64]
Wealth indexb
First quintile	0.3	1.00	1.00	0.2	1.00	1.00	2	1.00	1.00
Second quintile	1	3.44 [1.52–7.77]	3.62 [1.49-8-75]	0.1	0.46 [0.10–2.07]	0.37 [0.09–1.60]	2	0.95 [0.47–1.91]	0.86 [0.41–1.79]
Third quintile	2	5.46 [2.27–13.17]	6.69 [2.56–17.51]	0.3	1.55 [3.25–7.37]	1.09 [0.24–4.99]	5	2.37 [1.18–4.75]	2.35 [1.14–4.85]
Forth quintile	1	5.14 [2.05–12.85]	5.06 [1.87–13.69]	0.6	3.00 [0.74–12.29]	1.53 [0.36–6.39]	5	2.07 [1.08–3.93]	1.69 [0.82–3.49]
Fifth quintile	2	6.46 [2.99–13.95]	5.20 [1.91–14.15]	0.5	2.48 [0.60–10.26]	0.81 [0.15–4.26]	7	3.03 [1.65–5.58]	2.54 [1.29–5.01]
Smoking Status (Adolescents)
Never smoker	0.3	1.00	1.00	0.1	1.00	1.00
Experimental smoker	2	5.89 [2.61–13.3]	5.13 [2.40–10.99]
Former smoker	5	18.10 [6.90–47.45]	17.91 [6.88-46-63]	1.1	9.33 [4.98–17.50]	10.31 [5.33–19.94]
Current smoker (adolescents)	9	33.97 [20.09–57.45]	35.81 [19.36–66.23]
Nondaily smoker (adults)							4	1.00	1.00
Daily smoker, <5 cigs (adults)							5	1.20 [0.76–1.90]	1.51 [0.93–2.44]
Daily smoker, 5+ cigs (adults)							6	1.66 [1.08–2.52]	2.69 [1.79–4.03]
Quit attempt last 12 month
No							4	1.00	1.00
Yes							5	1.05 [0.71–1.55]	1.01 [0.65–1.55]

^a. Education: For adolescents, education of the head of household was used, with the same categories used as for adults

^b. Wealth index: was represents household assets, with Principal Component Analysis scores used to divide households into quintiles.

Higher education was associated with a greater likelihood of using e-cigarettes among adult nonsmokers (i.e., university or more vs primary or less: AOR = 8.48, 95% CI 2.25 – 31.93) and smokers (i.e., Technical vs primary or less: AOR = 1.96, 95% CI 1.10 – 3.50). Highest compared to lowest wealth was associated with greater likelihood of using e-cigarettes among adolescents (AOR = 5.20, 95% CI 1.91 – 14.15) and adult smokers (AOR = 2.54, 95% CI 1.29 – 5.01).

Among adolescents, e-cigarette use was higher among smokers (AOR =35.81, 95% CI 19.36 – 66.23), former smokers (AOR = 17.91, 95% CI 6.88 – 46.63) and experimental smokers (AOR=5.13, 95% CI 2.40–10.99) than among never smokers. Former smokers had a higher likelihood of using e-cigarettes than never smokers (AOR = 10.31, 95% CI 5.33–19.94). Among adult current smokers, daily smokers were more likely to use e-cigarettes than non-daily smokers (i.e., daily smoker +5 vs nondaily smokers: AOR = 2.69, 95% CI 1.79 – 4.03).

Among current users, the most frequent source of e-cigarettes was the supermarket (18.7%) and other people (25%) (Appendix-Table 1). Results from the sensitivity analyses using multinomial logistic regression (Appendix-Table 2a-2c,), multiple imputation (i.e., Appendix-Table 3a-3c) and the analyses of the subsamples (all adults and adults current smokers with recent former smokers) (i.e., Appendix-Table 4a-4c), were qualitatively and quantitatively consistent with the findings from the models described herein.

DISCUSSION

Our results for this national sample of Mexicans is consistent with the pattern of e-cigarette awareness, trial and use found in prior research in high income countries.^{1,3,26,27,44,45} Older adolescents (16–17 vs 12 years) and younger adults (18–24 vs 45–65 years) were more likely to have heard, tried and be current users of e-cigarettes. Furthermore, adolescent and adult current smokers, especially those who smoke more than 5 cigarettes per day, were more likely to have heard of, tried and currently use e-cigarettes. Indeed, our results are similar to other studies that indicate that being a current smoker is one of the strongest predictors of these outcomes across countries.^2.26,44,45

As in other studies,^20,24,25 male adolescents and male adult former smokers were more likely than their female counterparts to have heard, tried and be current users of e-cigarettes. However, contrary to expectations, female adult smokers were more likely than males to currently use e-cigarettes. Future longitudinal research on e-cigarettes in Mexico by gender is necessary to better understand this finding.

Smokers who made a quit attempt in the past year were more likely to have heard of e-cigarettes compared to current smokers who had not tried to quit, but no differences between these groups were found for trial or current use of e-cigarettes. These results contrast with previous studies that have found greater use amongst smokers who want to quit.^2,3,26 Furthermore, this strengthens the hypothesis that in Mexico e-cigarettes are used in conjunction with conventional cigarettes (i.e., dual use) rather than for quitting smoking. However, our study suggest that former smokers adolescents and adults were more likely to have heard of, tried and use e-cigarettes compared to never smokers. This finding is consistent with previous studies.^{26, 27} Further research is needed in Mexico to evaluate the use and ways to enhance the potential effectiveness of e-cigarettes for smoking cessation. Indeed, the pattern of smoking amongst Mexicans^36,36,42 and among US Latinos of Mexican ancestry⁴⁵ is substantially “lighter” than other populations (i.e., more nondaily smokers; fewer cigarettes per day), and the implications of this pattern for the substitution of e-cigarettes for conventional cigarettes is unclear. Such studies can help inform specific recommendations around policies to benefit public health.

As expected, living in urban areas was associated with greater e-cigarette awareness among all three groups and, among adolescents, with greater e-cigarette trial. However, there were no significant rural-urban differences for trial among adults or for current use in any group studied. These non-significant relationships were primarily found in the final models that adjusted for education and wealth, which likely explain the impact of rurality on these outcomes. Indeed, higher education increased the likelihood of awareness, trial and current use among adult nonsmokers and smokers. For adolescents, however, higher educational attainment of the head of the household was positively associated with e-cigarette awareness, as in prior research with Mexican adolescents.³⁰ Greater household wealth was independently associated with higher awareness and trial of e-cigarettes among adolescents, current smokers, and non-smokers adults. These findings are consistent with previous studies.^20,23,45

Several study limitations should be acknowledged. For instance, we were not able to examine many potential risk factors for trial and use of e-cigarettes (i.e., friend or family use; risk perceptions; beliefs about usefulness for smoking cessation; internet ad exposure and use of e-cigarettes with nicotine). Future research that assesses these factors may help better understand the importance of nicotine in facilitating or impeding both adult cessation and youth initiation.

Another limitation concerns determination of causal relationships given the cross-sectional nature of the data, although reverse causality is unlikely for many of the sociodemographic variables we examined (i.e., age, sex, education, wealth, and rurality). Nevertheless, longitudinal research is needed to better understand the potential consequences of e-cigarette trial and use for smoking progression among youth and nonsmokers, as well as for smoking cessation among adults.

Conclusions

This nationally-representative survey allows extrapolation of results to the general population of Mexicans waged 12 to 65 years old, providing the most detailed assessment of the prevalence and correlates of awareness, trial and use of e-cigarettes in LMICs to date. Although the importation, production, distribution, and marketing of e-cigarettes have been banned in Mexico,³⁰ the prevalence of trial and use of e-cigarette is significantly higher than expected. Moreover, e-cigarette users easily acquire this product, suggesting that enforcement of the ban is weak. This is alarming given that e-cigarettes not only contain nicotine but also other toxic constituents^{5,47,48,49,50} and have significant cardiovascular effects.^51,52 Studies are needed of the contents (i.e., nicotine, additives, flavorings) and emissions to determine whether they are more hazardous when illegal.

Despite the ban on e-cigarette sales, a substantial number of Mexicans have access to e-cigarettes. Decision makers should consider e-cigarette regulations that ban advertising, promotion and sponsorship, regulate the e-cigarette content and emissions, and ban e-cigarette use in smoke-free places, as recommended by the WHO.⁵ At the same time, the country needs to invest in the infrastructure needed to enforce these regulations