Longitudinal transitions of cigarettes and electronic nicotinedelivery systems among adolescents: Construction of aretrospective cohort using recall data from a cross-sectionalsample

INTRODUCTION A major concern regarding non-cigarette tobacco or nicotine products (NCTNPs) is whether they facilitate or mitigate overall tobacco or nicotine use. We examined longitudinal transitions of cigarettes and electronic nicotine delivery systems (ENDS) by constructing a retrospective cohort based on the recall data of a cross-sectional sample. METHODS Using the Korea Youth Risk Behavior Survey, we constructed crosssectional data of 59576 adolescents into retrospective cohort data. Participants were categorized into 4 mutually exclusive tobacco or nicotine use states. We used a multistate Markov model to identify transitions between the states to calculate transition intensity ratios (TIRs), and examined the current use of tobacco or nicotine products to assess both gateway effects to cigarettes, and whether ENDS use helps adolescents quit cigarette smoking. RESULTS Compared with never use, use of ENDS was associated with an increased risk of initiation of cigarette use (TIR=6.8; 95% CI: 4.5-10.2). The risk of transitioning from cigarette ever use to ENDS, compared with never use to ENDS, was even more pronounced (TIR=44.1; 95% CI: 34.1-56.9). The prevalence of current cigarette smoking was higher among those who started ENDS then cigarettes, compared to those who began cigarette use without experimenting with ENDS (43.1% vs 35.8%). Moreover, 27.8% (95% CI: 23.6-32.0%) of adolescents who experimented first with cigarettes then moved to ENDS were current users of cigarettes, and 46.4% (95% CI: 42.1-51.1%) of these adolescents were current users of both cigarettes and ENDS. CONCLUSIONS Based on the recall data of a cross-sectional sample, we demonstrate that ENDS experimentation increases the likelihood of cigarette smoking initiation. A significant proportion of these adolescents continue to use cigarettes. Moreover, those who experimented with cigarettes then ENDS also continue smoking cigarettes or both cigarettes and ENDS. We suggest comprehensive tobacco control policies for all tobacco/nicotine products and monitoring the timing of NCTNP initiation in cross-sectional surveys.