Prospective association between e-cigarette use frequency patterns and cigarette smoking abstinence among adult cigarette smokers in the United States

Addiction. 2022 Dec;117(12):3129-3139. doi: 10.1111/add.16009. Epub 2022 Aug 12.


Aims: To estimate the association of longitudinal patterns of e-cigarette use with cigarette smoking abstinence, after accounting for time-dependent confounding and selection bias.

Design: Secondary analysis of longitudinal national cohort data. Using marginal structural models and four waves of the population assessment of tobacco and health (wave 1, 2013-14; wave 2, 2014-15; wave 3, 2015-16; wave 4, 2016-18), we estimated the association of vaping frequency across waves 2 and 3 with 12-month sustained cigarette smoking abstinence at wave 4, adjusting for time-dependent confounders at waves 1 and 2 and selection bias due to drop-out with inverse probability of treatment and censoring weights.

Setting: United States.

Participants/cases: A total of 5699 adults (18+ years) who smoked cigarettes and did not vape at wave 1.

Measurements: The exposure was vaping frequency at waves 2 and 3 (non-use, non-daily use, daily use), representing nine possible combinations of vaping frequency across two waves. Non-use at both waves was the exposure reference group. The primary outcome was sustained 12-month cigarette smoking abstinence at wave 4.

Findings: Among 5699 adults who smoked cigarettes at wave 1, a total of 560 (9.8%) reported smoking abstinence at wave 4. Compared with nonuse at both waves, daily vaping at both waves [risk ratio (RR) = 3.82, 95% confidence interval (CI) = 2.59-5.64] and non-use at wave 2 followed by daily vaping at wave 3 (RR = 2.50, 95% CI = 1.66-3.77) were positively associated with smoking abstinence; non-daily vaping at both waves was inversely associated with smoking abstinence (RR = 0.28, 95% CI = 0.11-0.75). Results persisted after accounting for misclassification of e-cigarette use and cigarette smoking abstinence and after restricting to participants with plans to quit smoking.

Conclusions: In a US cohort of adult smokers, longitudinal patterns of vaping frequency appear to predict smoking abstinence, even after accounting for several sources of systematic error. Consistent daily vaping is associated with increased chances of cigarette smoking abstinence, while consistent non-daily vaping is associated with decreased chances of smoking abstinence.

Keywords: Confounding; e-cigarettes; marginal structural models; misclassification; selection bias; smoking cessation; vaping.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Adult
  • Cigarette Smoking* / epidemiology
  • Electronic Nicotine Delivery Systems*
  • Humans
  • Nicotiana
  • Smokers
  • Smoking Cessation* / methods
  • Tobacco Products*
  • United States
  • Vaping* / epidemiology