This cohort study examines associations of vaping with long-term tobacco cessation among US cigarette smokers who used electronic nicotine delivery systems.
Key Points
Question
Is daily or nondaily vaping associated with increased smoking cessation after 4 years of follow-up?
Findings
In this cohort study of 6013 cigarette smokers in the US, compared with matched controls who did not vape, smoking cessation rates were significantly lower (5.3 percentage points) among those who vaped nondaily; there was no statistically significant difference for those who vaped daily. Abstinence from both vaping and smoking was significantly lower (14.7 percentage points) among those who vaped daily.
Meaning
In this cohort study, neither daily nor nondaily vaping among US smokers was associated with increased smoking cessation; however, each was associated with reduced tobacco abstinence, suggesting that vaping may prolong both smoking and nicotine dependence among US smokers.
Abstract
Importance
An important public health goal is to increase tobacco cessation, but there is limited research on associations of vaping with tobacco cessation.
Objective
To estimate the association of vaping with long-term tobacco cessation among US cigarette smokers who used electronic nicotine delivery systems (ENDS; ie, e-cigarettes) in 2017.
Design, Setting, and Participants
This cohort study used a nationally representative sample of US cigarette smokers from the Population Assessment of Tobacco and Health cohort at wave 4 (2017) with follow-up at wave 6 (2021). Data were analyzed from June 2023 to June 2024.
Exposure
Wave 4 ENDS use: daily, nondaily, or no use.
Main Outcomes and Measures
The primary outcomes were 12 or more months’ abstinence from cigarette smoking and 12 or more months’ abstinence from both cigarette and ENDS use, ascertained at wave 6. Propensity score matching was used to control confounding on 14 potential confounders, including interest in quitting, income, age, education, nondaily smoking, and presence of a smoke-free home.
Results
A total of 6013 smokers were included in the sample (3634 aged ≥35 years [weighted percentage, 65.2%]; 3182 female [weighted percentage, 46.5%]). Among smokers who vaped daily (228 individuals), an estimated 20.9% (95% CI, 15.0% to 26.8%) were abstinent from cigarette smoking at follow-up, compared with 14.3% abstinence (95% CI, 13.0% to 15.5%) among smokers who did not vape (5070 individuals) and 12.6% abstinence (95% CI, 9.8% to 15.4%) among smokers who vaped nondaily (715 individuals). Compared with similar propensity score–matched smokers who did not vape, smoking cessation was 4.1 percentage points lower among those who vaped daily (95% CI, −11.9 to 3.6 percentage points; P = .30), a nonsignificant difference. Smoking cessation was 5.3 percentage points lower among those who vaped nondaily (95% CI, −9.1 to −1.5 percentage points; P = .01) compared with similar propensity score–matched controls. Considering abstinence from both smoking and vaping, compared with matched controls, smokers who vaped daily had lower abstinence at follow-up by 14.7 percentage points (95% CI, −20.2 to −9.2 percentage points; P < .001), and those who vaped nondaily had lower abstinence by 7.2 percentage points (95% CI, −10.7 to −3.8 percentage points; P < .001).
Conclusions and Relevance
In this representative cohort study of US smokers who used ENDS, neither daily nor nondaily vaping was associated with increased smoking cessation, and each was associated with reduced tobacco abstinence, suggesting that careful adjustment of confounding is critical in studies of ENDS and smoking cessation.
Introduction
From 2014 to 2016, US smokers trying to quit were more likely to report using an e-cigarette than a US Food and Drug Administration–approved cessation aid to help with the quit attempt.1 E-cigarettes include all electronic nicotine delivery systems (ENDS).2 By 2020, 68% of current US smokers believed that ENDS use might help with smoking cessation.3 However, there is no corresponding consensus in the scientific literature. While a recent Cochrane review4 of randomized trials concluded that e-cigarettes increase smoking cessation more than nicotine replacement therapy, a meta-analysis5 that also included observational studies noted substantial heterogeneity and concluded that e-cigarette use is not generally associated with smoking cessation, although daily vaping might be helpful. A 2021 US Preventive Services Task Force review6 concluded that evidence of benefits from vaping was both limited and inconsistent.
There may be several reasons for such heterogeneity of results. Commercially available e-cigarettes have changed dramatically over time, notably with the widespread adoption of e-cigarettes with high concentrations of nicotine in the US in 2017.7 It is possible that more potent e-cigarettes may be more effective for smoking cessation.8 Different studies have used different outcomes, with some using short-term abstinence (ie, <6 months) as the outcome measure9,10 and others using at least 6 months’ abstinence.11,12 Differences in inclusion criteria for the comparison group, particularly on baseline level of interest in quitting cigarettes, may also potentially confound results.13
In this study, we used data from all adult current smokers from wave 4 (2017) of the nationally representative US Population Assessment of Tobacco and Health (PATH) cohort14,15,16 to investigate the association of daily and nondaily vaping in 2017 with 12 or more months’ abstinence at 4-year follow-up in 2021 (wave 6). We considered 2 abstinence outcomes: 12 or more months’ abstinence from cigarette smoking and 12 or more months’ abstinence from both smoking and vaping. We used propensity score matching because of its relative transparency and ease of interpretation, and, following statistical recommendations,17 matched on a comprehensive list of baseline potential confounders, while being careful to exclude potential mediating variables. We used the matching results to demonstrate the degree of potential confounding in unadjusted analyses from this observational study. As a complementary analysis, we used multivariable logistic regression to investigate these same comparisons.
Methods
Data Source
The PATH Study obtained written participant informed consent and was overseen by the Westat Institutional Review Board. This cohort study was a secondary data analysis that used deidentified data and had no contact with human participants, which met the criteria for exemption from further human participants’ protection review by the University of California, San Diego institutional review board. The reporting of the study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. The PATH cohort study sampled from a representative address-based list of US households, oversampling tobacco users, young adults aged 18 to 24 years, and African American individuals; respondents are regularly reinterviewed using computer-assisted methods.14 At wave 4, 27 757 adults from wave 1 were reinterviewed from December 1, 2016, through January 3, 2018, along with an additional randomly selected replenishment sample (6065 individuals).15 This wave 4 sample was reinterviewed in 2019 (wave 5) and 2021 (wave 6; March to September). Household screener response rates were 54% for the initial sample and 52.8% for the replenishment sample. Adult survey response rates were greater than 60% for all surveys. We use the PATH restricted use data files in this study.16
Current Tobacco Use Status at Wave 4
For each PATH survey, lifetime cigarette smokers (ie, those who have smoked 100 or more cigarettes) were asked if they “currently smoke every day, some days, or not at all.” Those responding every day or some days were categorized as current established cigarette smokers. Respondents were also asked, “Do you now use electronic nicotine products every day, some days, or not at all?”
Study Outcomes
Study outcomes were defined as abstinence for at least 12 months from (1) cigarette smoking and (2) cigarette and/or ENDS use at wave 6. Respondents were asked if they had ever used any electronic nicotine products, if they currently smoke or use electronic nicotine products every day or some days, and if they had smoked cigarettes or used an electronic nicotine product within the past 12 months.
Potential Confounders at Wave 4
As in prior studies,18,19 potential confounders associated with smoking behavior included sociodemographic variables (age, sex, education, and income), self-reported race and ethnicity chosen from among survey options, cigarette smoking status (daily or nondaily), e-cigarette vaping status (daily, nondaily, or no use), interest in quitting cigarettes, having made a quit attempt in the past year, the perceived harm of cigarettes, insurance status, the presence of a smoke-free home, the existence of externalizing and/or internalizing mental health symptoms, and diagnosis of a smoking-related disease. We categorized race and ethnicity into White and other, with the latter including African-American or Black, Asian, Hispanic, and multiracial participants; race and ethnicity were included because smoking behaviors vary across different racial and ethnic groups.20 Measurement details are in the eAppendix in Supplement 1.
Statistical Analysis
Analyses were conducted in R version 4.4.2 (R Project for Statistical Computing) from June 2023 to June 2024. Population estimates used the wave 6 all-waves longitudinal weights, which adjust for baseline nonresponse and longitudinal study dropout; inference was conducted using replicate weights with balanced repeated replication and Fay adjustment.21 Statistical significance was assessed at the 5% level (ie, a 2-tailed P < .05). Differences among weighted proportions were assessed by Pearson χ2 test (R package, survey). Prior to conducting the analysis, we used multivariate imputation by chained equations (R package, mice) to impute missing data in covariates.
Differences in cessation rates between ENDS users and propensity score–matched controls who did not vape were used to estimate the average causal effect of vaping at baseline on cigarette smoking abstinence at follow-up among the population of US cigarette smokers who use ENDS for 3 exposure groups: daily ENDS use, nondaily ENDS use, and any ENDS use (daily or nondaily). We estimated each of the 3 propensity scores separately using logistic regression with the corresponding exposure outcome, the subsample of exposed and control respondents, and the complete list of potential baseline confounders. We then matched each participant in the exposed group with 2 controls with a similar propensity score using 2:1 nearest-neighbor matching without replacement and a caliper of 0.1 (R package, MatchIt); the difference in abstinence rates between the exposed group and the matched controls (R package, survey) is the estimated average causal effect of e-cigarette use in the exposed group.22 Conceptually, for each ENDS user (the index case) the average outcome of the pair of matched control nonusers is used to impute the cessation outcome that would have occurred if the index case had not used ENDS. Hence, for these comparisons we used the survey weights of the index case.23 To assess the quality of the matching, we utilized a Love plot to visualize the absolute standardized mean difference between cases and controls for each potential confounder. As a sensitivity analysis, multivariable logistic regression with survey weights (R package, survey) was used to assess the associations of abstinence outcomes with baseline covariates.
Results
Baseline Characteristics and Vaping Status
At wave 4 (2017), there were 9915 current established cigarette smokers, of whom 6013 (3634 aged 35 years or older [weighted percentage, 65.2%]; 3182 female [weighted percentage, 46.5%]; 3779 White [weighted percentage, 68.0%]; 2234 other race [weighted percentage, 32.0%]) completed wave 5 and wave 6, comprising the study sample (Table 1). An estimated 14.4% (95% CI, 13.4%-15.0%) of current adult US smokers also vaped e-cigarettes in 2017, with 228 (weighted percentage, 3.9%) vaping daily and 715 (weighted percentage, 10.6%) vaping nondaily. At baseline, daily vaping was associated with nondaily vs daily smoking status (χ21 = 158.25; P < .001) and with having made a cigarette quit attempt in the past year (χ22 = 38.68; P < .001). Daily vaping was also associated with age (χ21 = 23.74; P < .001), race and ethnicity (χ21 = 13.49; P = .002), income (χ22 = 14.67; P = .01), perceived harmfulness of cigarettes (χ21 = 12.47; P = .001), moderate or high scores on externalizing mental health symptoms (χ22 = 39.26; P < .001), and high scores on internalizing mental health symptoms (χ22 = 15.31; P = .005). Similar associations were generally observed when examining the subsample of smokers who vaped nondaily, but the magnitude of the differences were less marked.
Table 1. Vaping Among US Smokers by Baseline Characteristics, 2017a.
| Baseline characteristics | Sample size, No. (weighted %) | Weighted % (95% CI)b | |
|---|---|---|---|
| Daily vaping | Nondaily vaping | ||
| Overall | 6013 (100) | 3.9 (3.3-5.0) | 10.6 (9.7-11.0) |
| Age, y | |||
| <35 | 2379 (34.8) | 5.5 (4.4-6.6) | 14.8 (13.3-16.4) |
| ≥35 | 3634 (65.2) | 3.0 (2.3-3.7) | 8.3 (7.2-9.4) |
| Sex | |||
| Male | 2831 (53.5) | 4.0 (3.2-4.9) | 10.6 (9.2-11.9) |
| Female | 3182 (46.5) | 3.7 (2.8-4.5) | 10.6 (9.5-11.7) |
| Educational level | |||
| <High School | 1712 (26.9) | 3.0 (2.0-4.0) | 10.1 (8.5-11.8) |
| High school graduate | 1524 (30.1) | 3.8 (2.5-5.0) | 9.5 (8.0-10.9) |
| Some college or more | 2777 (43.1) | 4.5 (3.5-5.5) | 11.6 (10.4-12.7) |
| Race and ethnicityc | |||
| Non-Hispanic White | 3779 (68.0) | 4.5 (3.7-5.3) | 11.2 (10.0-12.3) |
| Other | 2234 (32.0) | 2.5 (1.7-3.4) | 9.3 (7.9-10.6) |
| Income, $ | |||
| <35 000 | 3607 (56.3) | 3.0 (2.3-3.8) | 11.0 (9.8-12.2) |
| ≥35 000 | 2148 (39.0) | 5.0 (3.9-6.2) | 10.3 (8.7-11.9) |
| Missing | 258 (4.7) | 4.1 (0.8-7.4) | 7.7 (4.2-11.3) |
| Cigarette smoking status | |||
| Daily | 4602 (76.2) | 2.1 (1.7-2.6) | 10.1 (9.1-11.0) |
| Nondaily | 1411 (23.8) | 9.5 (7.7-11.3) | 12.2 (10.4-13.9) |
| No past-year quit attempt and low interest in quitting | 2791 (47.2) | 2.6 (1.7-3.4) | 10.0 (8.7-11.3) |
| No past-year quit attempt and high interest in quitting | 1146 (19.8) | 3.4 (2.3-4.5) | 7.7 (6.1-9.3) |
| Past-year quit attempt | 2076 (33.0) | 6.0 (4.7-7.3) | 13.1 (11.5-14.7) |
| Smoke-free homed | |||
| No | 2707 (43.7) | 3.2 (2.4-4.0) | 11.2 (9.8-12.5) |
| Yes | 3273 (56.3) | 4.4 (3.5-5.4) | 10.0 (8.8-11.3) |
| Perceived harmfulness of cigarettesd | |||
| Low | 425 (6.7) | 7.2 (4.4-9.9) | 13.9 (10.6-17.2) |
| Moderate to high | 5577 (93.3) | 3.6 (3.0-4.3) | 10.3 (9.4-11.3) |
| Health insurance status | |||
| None | 1267 (21.3) | 4.3 (2.8-5.8) | 11.1 (9.3-13.0) |
| Yes | 4746 (78.7) | 3.8 (3.1-4.4) | 10.4 (9.4-11.4) |
| Externalizing mental health symptoms | |||
| Low | 4149 (70.5) | 2.9 (2.2-3.5) | 9.2 (8.2-10.3) |
| Moderate | 1531 (24.7) | 6.2 (4.6-7.7) | 13.3 (11.3-15.3) |
| High | 333 (4.8) | 6.8 (3.8-9.7) | 16.1 (12.2-20.1) |
| Internalizing mental health symptoms | |||
| Low | 2949 (51.5) | 3.3 (2.5-4.0) | 8.9 (7.7-10.0) |
| Moderate | 1435 (23.5) | 3.4 (2.3-4.5) | 11.1 (9.4-12.7) |
| High | 1629 (25.0) | 5.6 (4.0-7.1) | 13.5 (11.4-15.7) |
| Existence of smoking-related disease | |||
| No | 5604 (94.5) | 3.9 (3.3-4.6) | 10.2 (9.3-11.1) |
| Yes | 409 (5.5) | 2.6 (1.0-4.3) | 16.9 (12.3-21.4) |
Data are from wave 4 (2017) of the Population Assessment of Tobacco and Health (PATH) study.14,15,16
Weighted US population estimate using wave 6 (cohort 4) all-wave longitudinal weights.
For race and ethnicity, groups categorized as other were African American or Black, Asian, Hispanic, and multiracial.
Missing values that are less than 1% of the overall sample of 6013 were excluded from the table.
Population Rates of Abstinence at Follow-Up
The unadjusted population rate of 12 or more months’ abstinence from cigarettes in 2021 was estimated among smokers grouped by baseline ENDS use (Table 2). Among US smokers who vaped daily in 2017, an estimated 20.9% (95% CI, 15.0%-26.8%) were abstinent from cigarette smoking for 12 or more months in 2021, a rate more than 6 percentage points higher (95% CI, 0.5-12.8 percentage points) than the cessation rate among all US smokers who did not vape at baseline (weighted percentage, 14.3%; 95% CI, 13.0%-15.5%). Thus, in an analysis not adjusted for potential confounders, daily ENDS use was associated with increased smoking cessation. The unadjusted cessation rate was 1.7 percentage points lower among nondaily ENDS users (weighted percentage, 12.6%; 95% CI, 9.8%-15.4%) compared with those who did not use ENDS (weighted percentage, 14.3%; 95% CI, 13.0%-15.5%; χ22 = 9.72; P = .02). When considering 12 or more months’ abstinence from both cigarettes and ENDS in 2021, the abstinence rate at baseline was 4.6 percentage points lower for both daily (weighted percentage, 7.1%; 95% CI, 3.2%-11.0%) and nondaily vapers (weighted percentage, 7.1%; 95% CI, 4.9%-9.3%) compared with those who did not vape at baseline (weighted percentage, 11.7%; 95% CI, 10.5-12.8%; χ22 = 15.87; P = .002).
Table 2. Abstinence Outcomes Among US Smokers in 2021 by ENDS Use Status in 2017a.
| Status | ENDS use in 2017 based on wave 4 baseline frequency of use, No. (weighted %) | ≥12 Months’ abstinence in 2021 (wave 6), weighted % (95% CI)b | |
|---|---|---|---|
| Abstinence from cigarettes | Abstinence from both cigarettes and ENDS | ||
| No use | 5070 (85.6) | 14.3 (13.0-15.5) | 11.7 (10.5-12.8) |
| Nondaily vaping | 715 (10.6) | 12.6 (9.8-15.4) | 7.1 (4.9-9.3) |
| Daily vaping | 228 (3.9) | 20.9 (15.0-26.8) | 7.1 (3.2-11.0) |
Abbreviation: ENDS, electronic nicotine delivery systems.
Data are from wave 4 (2017) and wave 6 (2021) of the Population Assessment of Tobacco and Health (PATH) Study.14,15,16
Weighted US population estimate using wave 6 (cohort 4) all-wave longitudinal weights. For unadjusted abstinence outcomes by study covariates, see eTable 3 in Supplement 1.
Propensity Score Matching
We considered the 3 exposure groups (daily vaping, nondaily vaping, and any vaping [daily or nondaily]) separately. Each propensity score was estimated from a logistic regression using baseline covariates and the respective combined exposure and control (nonvaping) group (eTable 1 in Supplement 1). We then used the estimated propensity score to match each exposure-group smoker (ENDS user) with 2 similar control smokers who did not vape. Nearly all ENDS users were successfully matched (226 of 228 daily ENDS users; 712 of 715 nondaily ENDS users; 940 of 943 any ENDS users).
Considering the propensity score model for daily vaping (eTable 1 in Supplement 1), the adjusted odds of daily vaping were associated at baseline with the following factors: nondaily smoking (adjusted odds ratio [aOR], 4.33; 95% CI, 3.22-5.84), presence of a past-year quit attempt (aOR, 1.81; 1.34-2.46), interest in quitting smoking (aOR, 1.39; 95% CI, 1.02-1.89), age 50 years or older (aOR, 0.51; 95% CI, 0.33-0.75), race and ethnicity (aOR, 0.39; 95% CI, 0.28-0.54), moderate (aOR, 2.31; 95% CI, 1.63-3.27) and high (aOR, 2.97; 95% CI, 1.69-5.09) externalizing mental health symptoms, and perceived harmfulness of cigarettes (aOR, 0.37; 95% CI, 0.23-0.60). The imbalance in these covariates before matching is apparent in the Love plot (Figure 1). After matching, most cases and controls were very well balanced on all covariates (standardized mean difference <0.1); education and income were also balanced with standardized mean differences less than 0.2 (Figure 1).
Figure 1. Absolute Standardized Mean Differences (Love Plot) for Potential Confounder Variables Comparing Current Smokers Who Vaped Daily vs Those Who Did Not Vape in 2017.
Orange circles are differences in the full sample before matching and blue circles are differences in the 2:1 matched subsample. Distance is defined as the standardized mean difference between the groups in the estimated propensity score. Data are from wave 4 (2017) of the Population Assessment of Tobacco and Health (PATH) study.14,15,16
When considering nondaily vaping, similar baseline associations were generally observed, but differences were less marked (eTable 1 in Supplement 1). In the propensity score model, covariate patterns were largely similar to the model for daily vaping, but effect sizes were smaller. Importantly, however, interest in quitting was not associated with nondaily vaping (aOR, 0.97; 95% CI, 0.81-1.16). Also, smokers with an existing smoking-related disease were more likely to report nondaily vaping (aOR, 1.62; 95% CI, 1.21-2.14), an association that was not observed with daily vaping. After matching, all covariates had standardized mean differences less than 0.1, indicating excellent balance (eFigure 1 in Supplement 1). Results for any vaping were similar to the results for nondaily vaping (eTable 1 and eFigure 2 in Supplement 1).
Matched Analysis of Vaping and Abstinence Outcomes Among Smokers Who Used ENDS
To estimate the association of ENDS use with smoking cessation among those who did use ENDS, we estimated the population rate of 12 or more months’ cigarette abstinence in 2021 among the ENDS users and then compared the abstinence rate among their matched controls (Figure 2A). Smokers who vaped daily at baseline had a smoking cessation rate that was 4.1 percentage points lower than their matched controls (95% CI, −11.9 to 3.6 percentage points; P = .30), although this difference was not statistically significant. For nondaily vapers, vaping was associated with a reduction in smoking cessation of 5.3 percentage points (95% CI, −9.1 to −1.5 percentage points; P = .01). For the combined daily and nondaily vaping exposure group, the estimated effect size of ENDS use was much smaller and not statistically significant (percentage point difference, −1.13; 95% CI, −4.30 to 2.03; P = .48).
Figure 2. Estimated Differences in Abstinence Rates Between US Smokers Who Vape and Their Matched Controls Who Do Not Vape for Daily Vaping, Nondaily Vaping, and Any Vaping (Daily or Nondaily).
Vaping assessed at baseline (2017) and abstinence assessed at follow-up (2021), for the outcomes of 12 or more months’ abstinence from cigarette smoking (A) and 12 or more months’ abstinence from both cigarette smoking and vaping (B). Risk difference (RD) indicates the difference in abstinence rates, smokers who vape minus controls. ENDS indicates electronic nicotine delivery systems.
When considering the outcome of 12 or more months’ cessation of both cigarettes and e-cigarettes, daily vaping among US smokers who vaped was associated with a reduction in smoking cessation of 14.7 percentage points (95% CI, −20.2 to −9.2 percentage points; P < .001) (Figure 2B). Nondaily vaping among US smokers who vaped was associated with a reduction in cessation of smoking and vaping of 7.2 percentage points (95% CI, −10.7 to −3.8 percentage points; P < .001) and any vaping among US smokers was associated with a reduction in cessation of smoking and vaping of 5.77 percentage points (95% CI, −8.24 to −3.31 percentage points; P < .001).
Sensitivity Analyses
As a complementary analysis, we estimated the association of daily and nondaily vaping with abstinence outcomes using 2 weighted multivariable logistic regression models on the whole sample (Table 3 and eTable 2 in Supplement 1). The unadjusted estimates of covariates are given in eTable 3 in Supplement 1. Considering 12 or more months’ abstinence from cigarettes in 2021, the daily vaping results were similar to the matching analysis; smokers who vaped daily did not have statistically different odds of abstinence compared with those who did not vape (aOR, 0.94; 95% CI, 0.62-1.41; P = .76). For smokers who vaped on a nondaily basis, the direction of the results was consistent with the matched analysis; however, this finding was not statistically significant in the logistic regression model (aOR, 0.77; 95% CI, 0.58-1.02; P = .06). (Table 3)
Table 3. Multivariable Logistic Regression Model of 12 or More Months’ Abstinence From Cigarettesa.
| Baseline characteristics (2017) | Abstinence, weighted % (95% CI)b | Cigarette abstinence in 2021 | |
|---|---|---|---|
| aOR (95% CI) | P value | ||
| E-cigarette vaping status | |||
| No use | 14.3 (13.0-15.5) | 1 [Reference] | NA |
| Nondaily use | 12.6 (9.8-15.4) | 0.77 (0.58-1.02) | .06 |
| Daily use | 20.9 (15.0-26.8) | 0.94 (0.62-1.41) | .76 |
| Cigarette smoking status | |||
| Daily | 9.7 (8.7-10.8) | 1 [Reference] | NA |
| Nondaily | 29.1 (26.3-31.9) | 3.17 (2.55-3.95) | <.001 |
| Quitting history and interest in quitting | |||
| No past-year quit attempt and low quit interest | 11.9 (10.2-13.7) | 1 [Reference] | NA |
| No past-year quit attempt and high quit interest | 14.7 (12.0-17.4) | 1.46 (1.08-1.95) | .01 |
| Past-year quit attempt | 17.6 (15.9-19.4) | 1.34 (1.08-1.68) | .01 |
| Smoke-free home | |||
| No | 10.2 (8.8-11.6) | 1 [Reference] | NA |
| Yes | 17.7 (16.1-19.2) | 1.36 (1.13-1.64) | .001 |
| Age, y | |||
| ≥35 | 12.8 (11.4-14.2) | 1 [Reference] | NA |
| <35 | 17.3 (15.7-18.9) | 1.23 (1.02-1.49) | .03 |
| Education | |||
| <High school | 11.4 (8.8-14.0) | 1 [Reference] | NA |
| High school graduate | 11.7 (9.5-13.9) | 0.99 (0.73-1.36) | .97 |
| At least some college | 18.0 (16.4-19.7) | 1.43 (1.09-1.88) | .01 |
| Sex | |||
| Male | 15.0 (13.6-16.4) | 1 [Reference] | NA |
| Female | 13.6 (11.9-15.3) | 0.85 (0.70-1.05) | .13 |
| Race and ethnicityc | |||
| Non-Hispanic White | 14.1 (12.7-15.6) | 1 [Reference] | NA |
| Other | 14.8 (13.0-16.7) | 0.90 (0.74-1.10) | .31 |
| Income, $ | |||
| <35 000 | 12.5 (11.2-13.8) | 1 [Reference] | NA |
| ≥35 000 | 16.8 (14.9-18.6) | 1.01 (0.82-1.24) | .94 |
| Harmfulness of cigs | |||
| Low | 12.4 (8.9-15.9) | 1 [Reference] | NA |
| Moderate to high | 14.5 (13.3-15.7) | 0.94 (0.66-1.34) | .75 |
| Insurance status | |||
| No | 11.7 (10.0-13.4) | 1 [Reference] | NA |
| Yes | 15.1 (13.8-16.3) | 1.19 (0.97-1.46) | .10 |
| Externalizing mental health symptoms | |||
| Low | 14.4 (13.0-15.8) | 1 [Reference] | NA |
| Moderate | 14.6 (12.3-16.9) | 0.95 (0.74-1.23) | .70 |
| High | 12.0 (8.2-15.9) | 0.79 (0.51-1.21) | .27 |
| Internalizing mental health symptoms | |||
| Low | 14.8 (13.4-16.1) | 1 [Reference] | NA |
| Moderate | 14.3 (11.5-17.1) | 0.92 (0.71-1.18) | .48 |
| High | 13.5 (11.6-15.4) | 0.95 (0.75-1.20) | .66 |
| Existent smoking-related disease | |||
| No | 14.4 (13.2-15.5) | 1 [Reference] | NA |
| Yes | 14.1 (10.1-18.1) | 1.06 (0.71-1.58) | .78 |
Abbreviations: aOR, adjusted odds ratio; NA, not applicable.
Data are from wave 4 (2017) and wave 6 (2021) of the Population Assessment of Tobacco and Health (PATH) study.14,15,16
Weighted US population estimate of 12 or more months’ abstinence from cigarettes using the wave 6 cohort 4 all-wave longitudinal weights.
For race and ethnicity, groups categorized as other were African American or Black, Asian, Hispanic, and multiracial.
Considering 12 or more months’ abstinence from both cigarettes and ENDS at 2021 follow-up, smokers who vaped daily had lower odds of abstinence for both products compared with smokers who did not vape (aOR, 0.36; 95% CI, 0.19-0.69; P = .002). The same was true for smokers who vaped nondaily (aOR, 0.54; 95% CI, 0.37-0.78; P = .001) (eTable 2 in Supplement 1).
Discussion
In this representative cohort study, as in prior studies, unadjusted abstinence rates suggested that daily e-cigarette vaping was associated with increased long-term smoking cessation.9,19,24 However, once confounders were adjusted for, daily vaping at baseline was not associated with increased cigarette abstinence at follow-up as compared with similar smokers who did not vape. In a propensity score–matched analysis, there was a nonsignificant difference of 4.1 percentage points (95% CI, −11.9 to 3.6 percentage points; P = .30) among those who vaped daily and a significant reduction in smoking cessation of 5.3 percentage points (95% CI, −9.1 to −1.5 percentage points; P = .01) among those who vaped nondaily. Logistic regression models applied to the US population of smokers gave qualitatively similar results. Considering the outcome of 12 or more months’ abstinence from both smoking and vaping, results of both analytic approaches were similar but estimated effect sizes were larger. In the propensity score–matched analysis, daily vaping reduced tobacco abstinence rates by 14.7 percentage points (95% CI, −20.2 to −9.2 percentage points; P < .001) compared with matched controls, and nondaily vaping by 7.2 percentage points (95% CI, −10.7 to −3.8 percentage points; P < .001), suggesting that many dual smokers and ENDS users remain dependent on e-cigarettes even if they manage to quit smoking.25,26
Our analyses confirmed that a comparison of unadjusted smoking cessation rates between ENDS users (daily or nondaily) and nonusers in the US population is confounded, even in prospective analysis.13 The degree of confounding was such that the estimated effect size of ENDS use changed from seemingly beneficial in unadjusted analysis to deleterious once accounting for confounding. Smokers with baseline characteristics associated with future smoking cessation,18,19 such as recent quit attempts, interest in quitting, or nondaily smoking status, were much more likely to use ENDS daily. Additional identified confounders included age, race and ethnicity, mental health symptoms, presence of a smoke free home, and perceived harmfulness of cigarettes.
There are several related analyses of the PATH data, which largely address the older style ENDS, which were commercially available prior to 2017. Chen et al19 reported that 12 or more months’ cigarette abstinence among those who used ENDS to help with a quit attempt was substantially and significantly lower than for similar smokers who made a quit attempt without using an aid. Pierce et al27 considered cigarette smokers at baseline who had quit cigarettes 1 year later; those who switched to e-cigarettes (whether daily or not) had similar rates of relapse to cigarette smoking over the following year as those who were tobacco-free at the first follow-up. Kasza, et al10 studied daily smokers who used ENDS between 2014 and 2019 and concluded that using ENDS daily (vs nondaily) was associated with greater cessation success. Each of these studies is consistent with our results. Harlow et al28 selected wave-1 PATH respondents who smoked cigarettes but did not vape and reported 12 or more months’ cigarette abstinence at wave 4. Using marginal structural models and inverse probability weighting, they estimated that the less than 2% of the sample who vaped daily at both wave 2 and wave 3 had double the cessation rate of others. A concern was that adjustment for covariates did not result in substantial differences between adjusted and unadjusted estimates, suggesting inadequate control of confounding.
Strengths and Limitations
A major strength is the use of the state-of-the-art US PATH longitudinal cohort that provides population-based estimates of tobacco use changes over time. Use of such high-quality observational data increases the generalizability of reported results to the US population. To address self-selection bias, we matched each e-cigarette user to a nonuser on a comprehensive list of baseline variables that are associated with both e-cigarette use and smoking cessation and effectively controlled for confounding. Nevertheless, this study has limitations. Undoubtedly, there was still some unmeasured residual confounding between product user groups. Following statistical recommendations, we did not include mediator variables (such as cigarette dependence measures) in the set of control variables. We assumed that most cigarette dependence variables are intermediate variables on the causal pathway, likely to be affected by ENDS use. We chose to control for daily and nondaily smoking status, assuming that most daily smokers did not switch to nondaily status because of ENDS use. These assumptions, however, should be investigated in future work.
Conclusions
In this cohort study of a representative sample of US adult smokers, undertaken when high-nicotine e-cigarettes were readily available on the market, e-cigarette use was not associated with increased future smoking cessation. Although unadjusted cessation rates in the population appeared to favor the daily use of e-cigarettes, when major confounders were controlled for, the estimated effect size indicated a decrease in future smoking cessation among smokers who vape, whether they vaped daily or nondaily. Furthermore, e-cigarette use was associated with a large and significant decrease in future abstinence from tobacco use (cigarettes or ENDS). These nationally representative data suggest that vaping prolongs both smoking and nicotine dependence among US smokers.26
eAppendix. Measurement Detail for Study Covariates
eTable 1. Logistic Regression Propensity Score Models of ENDS Use at Baseline
eFigure 1. Absolute Standardized Mean Differences (Love Plot) for Potential Confounder Variables Between Current Smokers Who Vaped on a Non-Daily Basis Versus Those Who Did Not Vape
eFigure 2. Absolute Standardized Mean Differences (Love Plot) for Potential Confounder Variables Between Current Smokers Who Vaped (Either Daily or Non-Daily) Versus Those Who Did Not Vape
eTable 2. Logistic Regression Modeling 12+ Month Abstinence From Both Cigarettes and E-Cigarettes at Wave 6 (2021) by Baseline Covariates
eTable 3. 12+ Month Abstinence at Wave 6 (2021) by Baseline (W4, 2017) Characteristics
Data Sharing Statement
References
- 1.Caraballo RS, Shafer PR, Patel D, Davis KC, McAfee TA. Quit methods used by US adult cigarette smokers, 2014-2016. Prev Chronic Dis. 2017;14:E32. doi: 10.5888/pcd14.160600 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Centers for Disease Control and Prevention . E-cigarettes (vapes). Updated May 15, 2024. Accessed January 23, 2025. https://www.cdc.gov/tobacco/e-cigarettes/index.html#:~:text=Electronic%20cigarettes%20(e%2Dcigarettes),and%20the%20risks%20of%20vaping
- 3.Gravely S, Yong HH, Reid JL, et al. Do current smokers and ex-smokers who use nicotine vaping products daily versus weekly differ on their reasons for vaping? Findings from the 2020 ITC four country smoking and vaping survey. Int J Environ Res Public Health. 2022;19(21):14130. doi: 10.3390/ijerph192114130 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lindson N, Butler AR, McRobbie H, et al. Electronic cigarettes for smoking cessation. Cochrane Database Syst Rev. 2024;1(1):CD010216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wang RJ, Bhadriraju S, Glantz SA. E-cigarette use and adult cigarette smoking cessation: a meta-analysis. Am J Public Health. 2021;111(2):230-246. doi: 10.2105/AJPH.2020.305999 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.US Preventive Services Task Force . Tobacco smoking cessation in adults including pregnant persons: interventions. Published January 19, 2021. Accessed January 23, 2025. https://www.uspreventiveservicestaskforce.org/uspstf/recommendation/tobacco-use-in-adults-and-pregnant-women-counseling-and-interventions
- 7.Jackler RK, Ramamurthi D. Nicotine arms race: JUUL and the high-nicotine product market. Tob Control. 2019;28(6):623-628. doi: 10.1136/tobaccocontrol-2018-054796 [DOI] [PubMed] [Google Scholar]
- 8.Brandon KO, Simmons VN, Meltzer LR, et al. Vaping characteristics and expectancies are associated with smoking cessation propensity among dual users of combustible and electronic cigarettes. Addiction. 2019;114(5):896-906. doi: 10.1111/add.14551 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kalkhoran S, Chang Y, Rigotti NA. Electronic cigarette use and cigarette abstinence over 2 years among U.S. smokers in the population assessment of tobacco and health study. Nicotine Tob Res. 2020;22(5):728-733. doi: 10.1093/ntr/ntz114 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kasza KA, Tang Z, Seo YS, et al. Divergence in cigarette discontinuation rates by use of electronic nicotine delivery systems (ENDS): longitudinal findings from the United States PATH study waves 1-6. Nicotine Tob Res. 2022;27(5):236-243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hughes JR, Keely JP, Niaura RS, Ossip-Klein DJ, Richmond RL, Swan GE. Measures of abstinence in clinical trials: issues and recommendations. Nicotine Tob Res. 2003;5(1):13-25. doi: 10.1080/1462220031000070552 [DOI] [PubMed] [Google Scholar]
- 12.Pierce JP, Gilpin EA. A minimum 6-month prolonged abstinence should be required for evaluating smoking cessation trials. Nicotine Tob Res. 2003;5(2):151-153. doi: 10.1080/0955300031000083427 [DOI] [PubMed] [Google Scholar]
- 13.Pierce JP, Leas EC, Benmarhnia T, et al. E-cigarettes and cessation: the introduction of substantial bias in analyses of PATH study. Nicotine Tob Res. 2021;23(5):876-877. doi: 10.1093/ntr/ntaa234 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Hyland A, Ambrose BK, Conway KP, et al. Design and methods of the population assessment of tobacco and health (PATH) study. Tob Control. 2017;26(4):371-378. doi: 10.1136/tobaccocontrol-2016-052934 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Opsomer JD, Dohrmann S, DiGaetano R, et al. Update to the design and methods of the PATH study, wave 4 (2016-2017). Tob Control. 2024;33(6):733-738. doi: 10.1136/tc-2022-057851 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.National Addiction and HIV Data Archive Program . Population assessment of tobacco and health (PATH) study [United States] restricted-use files (ICPSR 36231). Regents of the University of Michigan. Published April 8, 2024. Accessed January 23, 2025. https://www.icpsr.umich.edu/web/NAHDAP/studies/36231/versions/V38
- 17.Robins J. The control of confounding by intermediate variables. Stat Med. 1989;8(6):679-701. doi: 10.1002/sim.4780080608 [DOI] [PubMed] [Google Scholar]
- 18.Pierce JP, Benmarhnia T, Chen R, et al. Role of e-cigarettes and pharmacotherapy during attempts to quit cigarette smoking: the PATH Study 2013-16. PLoS One. 2020;15(9):e0237938. doi: 10.1371/journal.pone.0237938 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chen R, Pierce JP, Leas EC, et al. Effectiveness of e-cigarettes as aids for smoking cessation: evidence from the PATH study cohort, 2017-2019. Tob Control. 2023;32(e2):e145-e152. doi: 10.1136/tobaccocontrol-2021-056901 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Sakuma KK, Pierce JP, Fagan P, et al. Racial/Ethnic Disparities Across Indicators of Cigarette Smoking in the Era of Increased Tobacco Control, 1992-2019. Nicotine Tob Res. 2021;23(6):909-919. doi: 10.1093/ntr/ntaa231 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Judkins DR. Fay’s method for variance estimation. J Off Stat. 1990;6(3):223
- 22.Ho DE, Imai K, King G, Stuart EA. Matching as nonparametric preprocessing for reducing model dependence in parametric causal inference. Polit Anal. 2007;15(03):199-236. doi: 10.1093/pan/mpl013 [DOI] [Google Scholar]
- 23.Lenis D, Nguyen TQ, Dong N, Stuart EA. It’s all about balance: propensity score matching in the context of complex survey data. Biostatistics. 2019;20(1):147-163. doi: 10.1093/biostatistics/kxx063 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Berry KM, Reynolds LM, Collins JM, et al. E-cigarette initiation and associated changes in smoking cessation and reduction: the population assessment of tobacco and health study, 2013-2015. Tob Control. 2019;28(1):42-49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Farsalinos KE, Barbouni A. Association between electronic cigarette use and smoking cessation in the European Union in 2017: analysis of a representative sample of 13 057 Europeans from 28 countries. Tob Control. 2021;30(1):71-76. doi: 10.1136/tobaccocontrol-2019-055190 [DOI] [PubMed] [Google Scholar]
- 26.Martinez U, Simmons VN, Sutton SK, et al. Targeted smoking cessation for dual users of combustible and electronic cigarettes: a randomised controlled trial. Lancet Public Health. 2021;6(7):e500-e509. doi: 10.1016/S2468-2667(20)30307-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Pierce JP, Chen R, Kealey S, et al. Incidence of cigarette smoking relapse among individuals who switched to e-cigarettes or other tobacco products. JAMA Netw Open. 2021;4(10):e2128810. doi: 10.1001/jamanetworkopen.2021.28810 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Harlow AF, Stokes AC, Brooks DR, et al. Prospective association between e-cigarette use frequency patterns and cigarette smoking abstinence among adult cigarette smokers in the United States. Addiction. 2022;117(12):3129-3139. doi: 10.1111/add.16009 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eAppendix. Measurement Detail for Study Covariates
eTable 1. Logistic Regression Propensity Score Models of ENDS Use at Baseline
eFigure 1. Absolute Standardized Mean Differences (Love Plot) for Potential Confounder Variables Between Current Smokers Who Vaped on a Non-Daily Basis Versus Those Who Did Not Vape
eFigure 2. Absolute Standardized Mean Differences (Love Plot) for Potential Confounder Variables Between Current Smokers Who Vaped (Either Daily or Non-Daily) Versus Those Who Did Not Vape
eTable 2. Logistic Regression Modeling 12+ Month Abstinence From Both Cigarettes and E-Cigarettes at Wave 6 (2021) by Baseline Covariates
eTable 3. 12+ Month Abstinence at Wave 6 (2021) by Baseline (W4, 2017) Characteristics
Data Sharing Statement