Association Between Electronic Cigarette Use and Smoking Reduction in France

Key Points

Question  Is electronic cigarette use associated with smoking reduction in the general population?

Findings  This cohort study found that, among daily smokers in France, regular (daily) electronic cigarette use is associated with a significantly higher decrease in the number of cigarettes smoked per day as well as an increase in smoking cessation attempts. However, among former smokers, electronic cigarette use is associated with an increase in the rate of smoking relapse.

Meaning  Daily electronic cigarette use appears to be helpful in initiating smoking cessation among persons who intend to quit tobacco; however, in the general population, its efficacy with regard to smoking abstinence in the long term is uncertain.

Importance  The electronic cigarette (EC) has become popular among smokers who wish to reduce their tobacco use levels or quit smoking, but its effectiveness as a cessation aid is uncertain.

Objective  To examine the association of regular EC use with the number of cigarettes smoked per day, smoking cessation among current smokers, and smoking relapse among former smokers.

Design, Setting, and Participants  The CONSTANCES (Consultants des Centres d’Examens de Santé) cohort study, based in France, began recruiting participants January 6, 2012, and is currently ongoing. Participants were enrolled in CONSTANCES through 2015, and included 5400 smokers (mean [SD] follow-up of 23.4 [9.3] months) and 2025 former smokers (mean [SD] follow-up of 22.1 [8.6] months) at baseline who quit smoking in 2010, the year in which ECs were introduced in France, or afterward. Analyses were performed from February 8, 2017, to October 15, 2018.

Main Outcomes and Measures  The association between EC use and the number of cigarettes smoked during follow-up was studied using mixed regression models. The likelihood of smoking cessation was studied using Poisson regression models with robust sandwich variance estimators. The association between EC use and smoking relapse among former smokers was studied using Cox proportional hazards regression models. All statistical analyses were adjusted for sociodemographic characteristics, duration of follow-up, and smoking characteristics.

Results  Among the 5400 daily smokers (2906 women and 2494 men; mean [SD] age, 44.9 [12.4] years), regular EC use was associated with a significantly higher decrease in the number of cigarettes smoked per day compared with daily smokers who did not use ECs (–4.4 [95% CI, –4.8 to –3.9] vs –2.7 [95% CI, –3.1 to –2.4]), as well as a higher adjusted relative risk of smoking cessation (1.67; 95% CI, 1.51-1.84]). At the same time, among the 2025 former smokers (1004 women and 1021 men; mean [SD] age, 43.6 [12.1] years), EC use was associated with an increase in the rate of smoking relapse among former smokers (adjusted hazard ratio, 1.70; 95% CI, 1.25-2.30).

Conclusions and Relevance  This study’s findings suggest that, among adult smokers, EC use appears to be associated with a decrease in smoking level and an increase in smoking cessation attempts but also with an increase in the level of smoking relapse in the general population after approximately 2 years of follow-up.

Cigarette smoking has been identified as a cause of cancer incidence and mortality since the end of World War II¹ and remains a major public health problem today.^2,3 Most smokers initiate tobacco use in adolescence⁴ and attempt to quit at around 30 years of age (especially women) or after 50 years of age.⁵ Pharmacotherapies (nicotine replacement therapy [NRT], bupropion hydrochloride, and varenicline tartrate) and behavioral therapies have been shown to be effective in helping smokers quit.^6-8 However, the appeal of smoking cessation aids is relatively low,⁹ and most quit attempts are done “cold turkey” (ie, stopping nicotine consumption all at once), without professional assistance or treatment,^10-13 which may be because smoking cessation aids have a financial cost or because smokers lack knowledge about their effectiveness and safety. There are also other reasons for not using smoking cessation aids; for example, some smokers believe that quitting without help gives them greater satisfaction and a feeling of self-control, strength, and autonomy.¹⁰ However, studies show that smokers who use smoking cessation aids are more likely to remain abstinent.¹⁴

Electronic cigarettes (ECs), sometimes also referred to as electronic nicotine delivery systems, have become popular in recent years. In the United States, approximately 15.3% of adults have used ECs,¹⁵ as have 14.6% of adults in Europe¹⁶ (41.7% of adults in France¹⁷). Approximately 3.2% of persons in the United States use ECs regularly,¹⁵ as do 1.8% of persons in Europe¹⁶ (3.8% of persons in France¹⁷). Electronic cigarettes are generally used by smokers who consider them to be less harmful than conventional cigarettes^18,19 and try to reduce or quit their cigarette consumption.¹² In some countries, such as France, ECs have become the leading smoking cessation method (27% of smokers who try to quit use ECs), ahead of NRT (18%).¹² However, the effectiveness of ECs as a smoking reduction and cessation aid is still a subject of controversy.^20-24 Randomized clinical trials have shown that ECs are as effective as²¹ or more effective than²⁵ NRT with regard to smoking reduction or cessation. On the other hand, there is also evidence that concurrent use of ECs and NRT may hamper smoking cessation.²⁶ However, prior studies have been based on relatively small samples or were conducted for short follow-up periods and have limited external validity.

One of the major concerns regarding the consequences of EC use is that it might reduce smokers’ motivation to quit²⁷ by providing a cue for smoking relapse.²⁸ Thus, paradoxically, EC users might need a larger number of quit attempts to achieve successful smoking cessation. Because former smokers may relapse at different rates, some after only a few days and others after several months,²⁹ it is necessary to follow the consequences of EC use over extended periods of time. To date, population-based evidence of long-term smoking trajectories after EC use is limited.

Moreover, most studies have focused on the association between EC use and smoking cessation among smokers who are trying to stop smoking^21,22 (ie, among those most motivated to quit). However, in the general population, smokers use ECs for various reasons—to reduce smoking level, to “smoke” indoors, to reduce tobacco-related expenses, to reduce health risks, or simply out of curiosity.^30,31 Recent studies have examined the effect of EC use in the general population,^23,32 but they have mostly been cross-sectional or short-term.

The aim of our study, based on the French CONSTANCES (Consultants des Centres d’Examens de Santé) cohort, was to investigate whether, in a community sample with prospective follow-up, EC use is associated with changes in the number of cigarettes smoked, with smoking cessation rates among smokers, and with smoking relapse among former smokers.

The CONSTANCES cohort was designed as a randomly selected sample of 200 000 adults drawn from France’s compulsory health insurance scheme (Caisse nationale d’assurance maladie), which covers about 85% of persons living in France (excluding farmers and self-employed workers). Recruitment started January 6, 2012, and is currently ongoing, among persons 18 to 69 years of age who live throughout France; the sociodemographic and economic characteristics of participants’ districts of residence are very similar to the French average. The sampling base at inclusion is composed of all persons meeting eligibility criteria; to obtain a sample comparable to the French population, an unequal probability sampling scheme overrepresenting men, younger participants, and those belonging to socioeconomically disadvantaged groups, who generally tend to have low participation levels in epidemiologic surveys, was implemented.^33,34 Every year, participants are invited to complete a paper and pencil or web-based questionnaire and additionally undergo a medical examination every 4 years.^33,34 Participants involved in the first wave of recruitment had more follow-up questionnaires than those recruited at later stages. The CONSTANCES cohort received the approval of the French legal authorities (Commission nationale de l’informatique et des libertés) that ensure ethical review, including an evaluation of participants’ written informed consent, data confidentiality, and safety.³³

Our investigation is based on CONSTANCES cohort participants included in the study through 2015, and who had at least 1 completed follow-up questionnaire (n = 40 311). A total of 19 912 participants (49.4%) were nonsmokers, 6423 (15.9%) were current smokers (at least 1 cigarette per day), and 13 976 (34.7%) were former smokers at the time of inclusion in CONSTANCES (eFigure in the Supplement).

We focused on current smokers and former smokers who reported having quit smoking from 2010 onward (the year that ECs were commercially introduced in France; n = 2046). After excluding participants with no data on EC use (1023 current smokers and 21 former smokers), our final analytical sample comprised 5400 current smokers and 2025 former smokers with at least 1 year of follow-up (mean [SD] follow-up of 2.6 [0.7] years for current smokers and 2.5 [0.6] years for). First, among current smokers, we studied the association between EC use and the number of cigarettes smoked as well as smoking cessation. Second, among former smokers, we studied the association between EC use and smoking relapse.

The 4 study outcomes examined are (1) the number of cigarettes smoked per day, (2) the difference between the number of cigarettes smoked per day at baseline and the number of cigarettes smoked per day at follow-up, (3) smoking cessation among smokers (ie, 0 cigarettes per day in any year of follow-up), and (4) cigarette smoking relapse among former smokers (≥1 cigarette per day reported on any follow-up questionnaire).

Participants reported current regular (daily) EC use (yes or no) (822 [15.2%] smokers and 176 [8.7%] former smokers) and the date of initiation of regular EC use, which made it possible to calculate the duration of regular EC use. For each participant, we evaluated EC use prospectively, irrespective of the type of device (rechargeable vs disposable; data on device type were not usable because of missing data). Because data on motives for using EC were not collected, EC use in our study is not restricted to only those who want to stop smoking. Among the 822 smokers who used an EC during the study, 194 (23.6%) had started using ECs prior to study baseline.

The duration of EC use has been shown to be associated with smoking cessation.³⁵ In secondary analyses, we studied the association between the duration of EC use (<1 year vs ≥1 year) and smoking patterns.

Our statistical analyses controlled for covariates previously shown to be associated with either tobacco cessation or EC use: sex, age,³³ marital status (single vs cohabiting or married), educational level³⁶ (≤high school vs higher education), employment status (employed, unemployed, or retired), citizenship (non-French vs French), household income³⁶ (<€1500 [$1694.50], €1500-€2799 [$1694.50-$3162], or≥€2800 [$3163] per month), financial difficulties (yes vs no), alcohol abuse (Alcohol Use Disorders Identification Test score), number of cigarettes smoked per day at the time of inclusion,³⁷ number of pack-years of smoking (lifetime tobacco exposure; a pack-year is defined as 20 cigarettes smoked every day for 1 year), depressive symptoms measured by the Center for Epidemiologic Studies–Depression scale, lifetime history of depression (yes vs no), respiratory problems in the preceding 12 months (yes vs no), lifetime history of cardiovascular disease (yes vs no), and lifetime history of cancer (yes vs no). In addition, we controlled for participants’ year of inclusion in the CONSTANCES cohort, the duration of follow-up, and prior lifetime episodes of smoking cessation³⁷ (none, <1 year, or ≥1 year).

To identify covariates associated with both the study exposure and the study outcomes, we conducted univariate logistic and linear regression analyses. All P values were from 2-sided tests and results were deemed statistically significant at P < .05.

Among daily smokers, the association of EC use with the number of cigarettes smoked per day with the difference in the number of cigarettes smoked per day between baseline and follow-up was estimated using mixed linear models adjusted for sociodemographic characteristics such as sex, age, marital status, educational level, and income; substance use, including alcohol abuse; number of cigarettes smoked per day; number of pack-years of smoking; and health characteristics, such as depressive symptoms and respiratory problems. The variables included in the final model were selected using the least absolute shrinkage and selection operator method.³⁸

To determine the likelihood of smoking cessation being associated with EC use, we used Poisson regression models with robust sandwich variance estimators, adjusted for sociodemographic characteristics, duration of follow-up, and previous smoking cessation attempts. This method was preferred to logistic regression, for which the adjusted odds ratios would have overstated the participants’ relative risk³⁹ of quitting smoking (28% of smokers reported quitting in any year of follow-up). Because the associations between EC use and cigarette smoking can vary with individuals’ sex, age, duration of previous smoking cessation attempts, and educational level, we additionally performed analyses stratified on these characteristics.

To test whether EC use is associated with later smoking relapse, we focused on former smokers who quit tobacco in or after 2010, and we used Cox proportional hazards regression models adjusted for sociodemographic characteristics, including sex, age, marital status, educational level, and income; alcohol use; cigarette use; and health conditions, such as depressive symptoms and respiratory problems. To estimate the time to event (relapse or regular smoking), we calculated the number of months between the inclusion in the CONSTANCES cohort and the follow-up questionnaire in which the participant reported regular smoking. Among former smokers who did not relapse, data were censored at the last follow-up questionnaire available. We verified the proportional hazards assumption both graphically and statistically. Because the level of EC use increased and because the devices used evolved over time, we performed supplementary analyses, stratifying our sample on the year of smoking cessation.

Overall, less than 2% of data were missing, except for data on number of pack-years of smoking, which were unavailable for 718 of 7425 participants (9.7%). Missing data on all covariates were imputed using multiple imputations (10 imputations per missing value) with fully conditional specification.⁴⁰ All data analyses were conducted using SAS, version 9.4 (SAS Institute Inc).

In our study, smokers (n = 5400) were followed up for a mean (SD) period of 23.4 (9.3) months, during which 822 (15.2%) reported regular (daily) use of an EC. As shown in Table 1, univariate analyses show that, compared with the 4578 nonusers, EC users were more likely to be male (423 [51.5%] vs 2071 [45.2%]), older (mean [SD] age, 45.9 [11.6] vs 44.7 [12.5] years), and in a civil partnership or married (403 [49.0%] vs 2142 [46.8%]) and were followed up for a longer period (mean [SD], 26.2 [9.5] vs 22.9 [9.1] months). Electronic cigarette users were heavier smokers (mean [SD], 12.9 [6.8] vs 10.0 [6.6] cigarettes per day; 17.5 [14.1] vs 12.6 [12.1] pack-years of smoking) and were more likely to have previously made an attempt to quit smoking (594 [72.3%] vs 3147 [68.7%]). Electronic cigarette users were also more likely to have depressive symptoms (mean [SD] Center for Epidemiologic Studies–Depression scale score, 14.1 [10.3] vs 12.2 [9.5]), a history of depression (199 [24.2%] vs 911 [19.9%]), or respiratory problems (646 [78.6%] vs 3116 [68.1%]).

Former smokers (n = 2025) were followed up for a mean (SD) period of 22.1 (8.6) months, during which 176 (8.7%) reported regular EC use (Table 1). Electronic cigarette users were more likely than the 1849 non-users to be male (111 [63.1%] vs 910 [49.2%]), have higher levels of tobacco smoking (mean [SD], 16.9 [12.6] vs 12.9 [13.7] pack-years) and lower levels of alcohol-related problems (mean [SD] Alcohol Use Disorders Identification Test score, 16.9 [12.6] vs 12.9 [13.7]), as well as higher levels of depressive symptoms (mean [SD] Center for Epidemiologic Studies–Depression scale score, 12.6 [9.8] vs 10.9 [8.6]).

In a univariate mixed linear model (Table 2), EC users smoked significantly more cigarettes per day than nonusers (11.2 [95% CI, 10.8-11.7] vs 9.8 [95% CI, 9.6-10.0]). However, after controlling for demographic, socioeconomic, substance use–related characteristics, and health characteristics, we found that the estimated number of cigarettes smoked per day was significantly lower among EC users than among nonusers (11.2 [95% CI, 10.5-11.8] vs 12.2 [95% CI, 11.6-12.8]). After adjustment for all covariates, EC users decreased the number of cigarettes smoked significantly more during the course of follow-up than did nonusers (–4.4 [95% CI, –4.8 to –3.9] vs –2.7 [95% CI, –3.1 to –2.4] cigarettes per day).

In both univariate and multivariate models, EC users were more likely to quit smoking during follow-up compared with nonusers (univariate relative risk, 1.59 [95% CI, 1.45-1.76]; multivariate relative risk, 1.67 [95% CI, 1.51-1.84]) (Table 2). In additional analyses, this association was stronger among participants who used ECs for more than 1 year (adjusted relative risk, 2.03 [95% CI, 1.82-2.27]) than among those who used ECs for less than 1 year (adjusted relative risk, 1.33 [95% CI, 1.15-1.54]) (eTable 1 in the Supplement). We found no statistical interaction between EC use and sex, age group, duration of prior smoking cessation, or educational level (eTable 2 in the Supplement).

Overall, compared with former smokers who did not use ECs, those who did were more likely to relapse to smoking (adjusted hazard ratio, 1.70 [95% CI, 1.25-2.30]) (Figure). This hazard ratio decreased with time from 1.70 (95% CI, 1.25-2.30) among persons who quit as of 2010 (n = 2025) to 0.94 (95% CI, 0.57-1.52) among persons who quit as of 2013 (n = 601) (Table 3).

Studying longitudinal associations between EC use and tobacco smoking patterns in a large population-based cohort study, we found that EC use was associated with a reduction in smoking level as well as an increased probability of smoking cessation. However, we also observed that, over time, EC users who quit tobacco tended to relapse to smoking more frequently than did nonusers. Thus, while EC use can help persons reduce their smoking levels in the short term, there is no evidence that it is an efficacious smoking cessation aid in the long term.

Our investigation has weaknesses that need to be acknowledged. First, our study was not designed to test whether ECs are efficacious with regard to tobacco smoking reduction. We had no information on the motives underlying EC use nor the extent to which participants intended to quit smoking. Previous studies have shown that the main reason for EC use among adults is the intention to reduce or quit smoking³⁰ and that ECs are the most used aid for smoking cessation in France (no aid, 52%; ECs, 27%; NRT, 18%).¹² Moreover, we controlled for previous smoking cessation attempts, and our results are consistent with those of other researchers who suggest that EC use is associated with an increase in the reduction of tobacco consumption over time.³² Therefore, it is likely that, among regular smokers, ECs primarily serve to help decrease tobacco use levels.

Second, participants’ nicotine dependence was not measured, but our analyses controlled for the number of cigarettes smoked per day and the number of pack-years of smoking, which can be considered as valid proxies.⁴¹ Similarly, smoking was self-reported, which could induce bias, but such measures are generally considered valid.⁴² Results of the Fagerström test for nicotine dependence were also not available. Third, the mean duration of follow-up was 23 months, which is longer than in most previous studies, but it could be argued that it should be even longer because smokers often need several quit attempts before achieving successful long-term smoking cessation.⁴³

Fourth, participants reported current EC use and the date of initiation, from which we derived the duration of EC use. However, the daily frequency of EC use (eg, number of puffs) was not documented. Previous studies have shown that smoking cessation is primarily associated with extensive EC use.^20,24 Similarly, we were not able to evaluate EC users’ nicotine intake or examine whether it is associated with smoking behavior. Most participants reported using ECs with nicotine, but the information regarding the nicotine dosage of the e-liquid was often missing. In future studies, it will be important to assess the frequency of EC use and associated nicotine levels via questionnaires or other direct means of data collection.

Despite these limitations, our study has important strengths. We assessed the association between EC use and smoking among smokers and former smokers prospectively in a large population sample, for approximately 2 years of follow-up on average. We were able to take into consideration the duration of EC use, which seems to play a role in smoking cessation. However, our main contribution to the existing literature is the finding of an elevated rate of smoking relapse among former smokers who use ECs.

Our results are in line with those of other studies showing that EC use can help reduce tobacco smoking^32,44,45 and encourage smoking cessation.^23,25 The decrease in tobacco consumption among smokers irrespective of EC use observed in national surveys¹⁷ suggests that recent policies, such as the ban on smoking in public places, the reimbursement for NRT, and the increase in the price of tobacco products, have been successful. We found that smokers who used ECs decreased their smoking significantly more than nonusers and that they had a significantly higher probability of quitting smoking during follow-up. A recent randomized clinical trial showed that, among smokers trying to quit smoking, EC use was associated with a higher level of 1-year abstinence compared with NRT (relative risk, 1.83 [95% CI, 1.30-2.58]; P < .001).²⁵ Unfortunately, we had no information on the reasons for EC use, but previous studies indicate that, in France, 82% of smokers and 89% of former smokers who use ECs consider them an aid to quit smoking or prevent a relapse.⁴⁶ It would be interesting to further explore whether this smoking reduction or cessation is observed mainly among smokers who use ECs as a cessation tool or is observed also among those who use ECs for other reasons. In additional analyses, we found that smoking cessation was associated with duration of EC use, which is consistent with findings from previous studies.³⁵

Although the EC users in our study were more likely to be male, there were no sex differences in the association between EC use and smoking cessation. Previous research showed no sex differences⁴⁷ or higher levels of smoking cessation among men,⁴⁸ but these studies were conducted prior to the introduction of ECs. In particular, women are more likely than men to quit smoking before the age of 50 years, while the opposite is true after 50 years.⁴⁷ Because men and women have different patterns of use and expectancies regarding ECs,⁴⁹ future research should focus on possible sex differences with regard to long-term patterns of smoking cessation.

Although EC use among smokers is associated with an increased probability of attempts to quit smoking, its use by former smokers, on the other hand, is linked to a higher likelihood of smoking relapse. This finding may be due to higher nicotine dependency among EC users or the fact that EC use may contribute to maintaining individuals’ levels of nicotine addiction over time. In particular, in the case of technical problems with an EC (eg, low battery or lack of e-liquid) or if an EC does not give the same pleasure as conventional cigarettes,^50-52 individuals may revert to smoking cigarettes.

However, levels of smoking relapse were not increased among former smokers who quit in recent years. Measures of plasma nicotine levels have showed that, compared with older models of ECs, the new generation delivers higher levels of nicotine to the bloodstream.^53,54 This finding may be an explanation as to why smokers who recently quit smoking and switched to ECs are less likely to relapse than those who quit earlier. Although we found a higher probability of relapse among former smokers who used ECs than among nonusers, the question of whether this difference could be associated with a shorter period of follow-up, technical improvements in ECs over time, or a change in the profile of EC users will need to be evaluated in future studies.

Among current smokers, EC use is associated with a decrease in the number of cigarettes smoked and with an increase in cessation attempts, especially if EC use lasts more than 1 year. However, among former smokers, EC use is associated with a higher likelihood of relapse to smoking. Although EC use may help individuals decrease smoking levels and initiate smoking cessation, it is not clear whether it leads to complete long-term cessation.

Accepted for Publication: March 29, 2019.

Corresponding Author: Ramchandar Gomajee, MSc, Inserm Unité Mixte de Recherche en Santé 1136, Pierre Louis Institute of Epidemiology and Public Health, Sorbonne Université, 27 rue Chaligny, 75012 Paris, France (ramchandar.gomajee@inserm.fr).

Published Online: July 15, 2019. doi:10.1001/jamainternmed.2019.1483

Author Contributions: Dr Melchior had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: El-Khoury, Goldberg, Zins, Lemogne, Kousignian, Melchior.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Gomajee, Kousignian, Melchior.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Gomajee, El-Khoury, Goldberg, Kousignian.

Obtained funding: Zins, Lemogne, Kousignian, Melchior.

Supervision: Kousignian, Melchior.

Conflict of Interest Disclosures: Dr Lemogne reported receiving personal fees from Janssen-Cilag, personal fees and nonfinancial support from Lundbeck, and nonfinancial support from Otsuka Pharmaceutical outside the submitted work. Drs Lemogne, Wiernik, Lequy, and Romanello reported receiving grants from INCA during the conduct of the study. No other disclosures were reported.

Funding/Support: The CONSTANCES (Consultants des Centres d’Examens de Santé) Cohort Study was supported and funded by the Caisse nationale d’assurance maladie des travailleurs salariés. The CONSTANCES Cohort Study is an “Infrastructure nationale en Biologie et Santé” is funded by grant ANR-11-INBS-0002 from Agence Nationale de la Recherche. CONSTANCES is also partly funded by MSD, AstraZeneca, and Lundbeck. The present analyses were supported by grant 2016-082 from Institut National du Cancer.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank the “Caisse nationale d’assurance maladie des travailleurs salaries” and the “Centres d’examens de santé” of the French Social Security which are collecting a large part of the data, as well as the “Caisse nationale d’assurance vieillesse,” ClinSearch, Asqualab, and Eurocell, in charge of the data quality control.