The prospective association between the use of E-cigarettes and other psychoactive substances in young people: A systematic review and meta-analysis

The use of electronic cigarettes by young people has increased exponentially in the past decade due to various health and social influences. E-cigarettes, particularly those containing nicotine, can cause health complications and addiction, which may result in a subsequent initiation of psychoactive substance use. This systematic review and meta-analysis evaluated the prospective association between e-cigarette use and subsequent use of psychoactive substances in young people aged 10 – 24 years. Pooling of data from the identified longitudinal studies showed that ever e-cigarette users have an increased likelihood for subsequent cannabis, alcohol, and unpre-scribed Ritalin/Adderall use compared to never e-cigarette users. The findings indicate a need for interventions to reduce e-cigarette use in adolescents and young adults.


Introduction
The use of electronic cigarettes has gained an alarmingly exponential increase in popularity among young people in the last decade.The WHO defines young people as those individuals with an age range of 10-24 years (WHO, 2022).From a neurobiological point of view, the brain does not reach full maturation until 25 years (Arain et al., 2013).According to the CDC, 8.1 million US adults used an e-cigarette in 2018, with the highest prevalence among individuals in the age range of 18-24 years (Villarroel et al., 2020).The rise in e-cigarette use among young people skyrocketed between 2011 and 2015, when a 900% increase was reported among adolescents in the United States (U.S.Department of Health and Human Services, 2016).Usage of disposable e-cigarettes increased significantly during 2019-2020 among young people who attended middle school (from 3.0% to 15.2%) and high school (from 2.4% to 26.5%).E-cigarette use remained high in 2020, when an estimated 3.6 million (13.1%) US middle and high school students reported using e-cigarettes within the past 30 days.In 2021, 11.3% of high school students (1.72 million) and 2.8% (320,000) of middle school students reported e-cigarette use (Park-Lee et al., 2021).In the United Kingdom, the number of e-cigarette users grew from around 700,000 in 2012-3.6 million in 2019, falling to 3.2 million in 2020, before increasing again in 2021-3.6 million (7.1%) (ASH, 2021).Results from a recent systematic review carried out by Kim et al. (2022) by pooling results from 53 studies conducted in Asian, South American, and European countries, in addition to the US, showed the overall prevalence of international lifetime e-cigarette use among young people to be 15.3%.
Common reasons for using e-cigarettes among young people include but are not limited to taste, entertainment, experimentation, social influence, replacement of tobacco cigarettes, curbing cravings, and/or addiction (Evans-Polce et al., 2018;Newcombe et al., 2021;Khouja et al., 2020).Young people may also consider e-cigarettes a good alternative to conventional tobacco cigarettes, as they are cheaper and thought to contain less toxins than tobacco cigarettes (Etter and Bullen, 2011).Toxins include carcinogenic chemicals like tar, lead, and arsenic.Ecigarette liquid usually consists of nicotine, propylene glycol and/or glycerine, various flavourings, and several other chemicals (ALA, 2020).There are also some e-cigarettes that are advertised to be nicotine-free; however, research shows that nicotine-free e-cigarettes may still contain nicotine (Buettner-Schmidt et al., 2016;Raymond et al., 2018).E-cigarettes have the potential to deliver equal or more nicotine than a tobacco cigarette, but this depends on the experience and puffing behaviour of the user, type of e-cigarette device, and the nicotine concentration chosen (Voos et al., 2019).Notably, the developing brain is more susceptible to the addictive properties of nicotine, even with intermittent exposure, compared to the fully developed adult brain (Abreu-Villaça et al., 2003).Nicotine's primary psychoactive actions are through binding to the nicotinic acetylcholine receptors in the brain to release dopamine, producing psychoactive effects, and its significant role in reward processing and reinforcing behaviour.Nicotine addiction occurs when the user's body adapts to nicotine exposure over time, raising the threshold to receive the dopaminergic effect and needing higher concentrations to attain it, resulting in a vicious cycle of addiction.Nicotine use in young people may lead to addiction to substances in addition to nicotine, as it primes the brain to be more vulnerable to the addicting effects of other psychoactive substances (Richter, 2019).
According to the Gateway Theory of Kandel and Kandel (2014), during the adolescence and young adulthood neurodevelopmental period, "nicotine interacts with other neurotransmitter systems (e.g.endocannabinoid, opioidergic), and as a result increases the rewarding effects of other drugs by enhanced activation of the reward circuitry" (Ren and Lotfipour, 2019, p.700).Therefore, e-cigarettes can potentially act as a gateway to subsequent use of other psychoactive substances.Notably, animal models showed nicotine exposure during adolescence and young adulthood to prime the intake of other stimulant drugs (e.g.cocaine, methamphetamines), and of other substances such as cannabis, opioids, and alcohol (Cardenas and Lotfipour, 2022;Kandel and Kandel, 2014;Ren and Lotfipour, 2019).
This neurobiological gateway effect, as well as other influences such as commodity, availability, and a lack of proper governance allows ecigarette use to possibly cause a domino effect, leading to the use of other psychoactive substances in young people.The public health implications of a gateway effect caused by e-cigarette use during adolescence and young adulthood would be important.In fact, adolescent polysubstance use has been associated with the development of severe mental health problems (e.g.major depression), in addition to poor social and economic outcomes (Vergunst et al., 2021;Felton et al., 2015).Furthermore, "1.5% of global disease burden is attributable to the health consequences of illicit drug and alcohol use" (Doggui et al., 2021, p.208).The abuse of opioids and alcohol has also been associated with an increased risk of lethal cardiopulmonary events in adolescents (Carreiro et al., 2019;Adger and Saha, 2013).Additionally, chronic cannabis use during adolescence and young adulthood has been associated with cognitive decline, school dropouts, and lower IQ (Hasin, 2018).However, despite the growing number of studies conducted in the last few years on e-cigarette users, there is a lack of qualitative and/or quantitative syntheses investigating the prospective relationship between e-cigarette use during adolescence and young adulthood, and the subsequent uptake of other psychoactive substances.Evidence is limited to reviews assessing the prospective relationship between e-cigarette use during adolescence and combustible cigarette smoking initiation (Adermark et al., 2021;O'Brien et al., 2022), and to a review published in 2019 investigating the association between e-cigarettes and cannabis use (Chadi et al., 2019).However, as stated by the same authors, one limitation of this study was the small number of pooled longitudinal studies (n = 3).Another review conducted by Rothrock et al. (2020) investigated the relationship between adolescent e-cigarette and alcohol use by only including retrospective and cross-sectional studies.Therefore, it was not possible to assess the directionality of the association between alcohol and e-cigarette use.Considering the mentioned gap in the literature, this systematic review and metaanalysis aimed to provide a qualitative and quantitative synthesis for the association between e-cigarette use and the subsequent use of other psychoactive substances like alcohol, cannabis, opioids and/or stimulants in young people (adolescents and young adults).Intervention: Observation of ever users of e-cigarettes.

Search strategy
Comparison or control: Never users of e-cigarettes.
Outcome: Subsequent use of psychoactive substances.
The "Preferred Reporting Items for Systematic Review and Meta-Analysis" (PRISMA) guidelines (Liberati et al., 2009) and the "Meta--Analysis for Observational Studies in Epidemiology" (MOOSE) guidelines (Stroup et al., 2000) were utilised to identify and assess relevant papers to include in this review (Conti et al., 2019).Papers were sourced through the PubMed, Ovid Medline, APA PsycInfo, ERIC and Embase databases, which were accessed using the NHS Knowledge Network Database.The following keywords were used to conduct the search: "association OR relationship OR correlation OR link OR connection AND e-cigarette OR e-cigarettes OR electronic cigarette OR electronic cigarettes OR vape OR vaping OR electronic nicotine delivery system AND psychoactive substances OR psychoactive substance OR alcohol OR stimulants OR stimulant OR cannabis OR marijuana OR cocaine OR opioids OR opioid OR drug OR drugs AND adolescent OR adolescents OR young people OR young person OR teenager OR teenagers OR youth".The inclusion and exclusion criteria are outlined in Table 1.Ethical and research governance approvals were not required for this research.

Quality assessment
All papers were quality appraised based on the Critical Appraisal Skills Programme Cohort Study Checklist (CASP, 2020) with some alterations to fit the yes/no criteria.
The following 14 questions were used: 1. Did the study address a clearly focused issue? 2. Was the cohort recruited in an acceptable way? 3. Was the exposure accurately measured to minimise bias? 4. as the outcome accurately measured to minimise bias? 5. .a. Have the authors identified all important confounding factors?b.Have they taken account of the confounding factors in the design and/or analysis?c.Was the follow-up of subjects complete enough? d.Was the follow-up of subjects long enough? 6. .7. Were there clear results to the study?8. Were the results precise? 9. Do you believe the results?10.Can the results be applied to the local population?11.Do the results of this study fit with other available evidence?12.What are the implications of this study for practice?Each question was graded Yes (Y), No (N) or Can't Tell (CT).For consistent grading, confounding factors specified in (5a) included sociodemographic, environmental, and interpersonal factors associated with adolescent subsequent psychoactive drug use.A suitable follow-up of > 80% of participants and of > 3 months, was used to answer questions (6a) and (6b) (Schulz and Grimes, 2002;Setia, 2016).For question (12), 53 implications of the study for practice included whether the study could be used to increase awareness, identify groups of patients at higher risk of subsequent psychoactive substances use, and/or suggest or indicate the implementation of further public health measures.

Quantitative analysis
Meta-analytic calculations were conducted to investigate the association between e-cigarette use and subsequent psychoactive substance use.Particularly, odds ratios (OR) for subsequent psychoactive substance use among e-cigarette ever users compared to never users and confidence intervals (CI) were extracted from seven studies (Lozano et al., 2021;Park et al., 2020;Bentivegna et al., 2021;Evans-Polce et al., 2020;Audrain-McGovern et al., 2018;Dai et al., 2018;Duan et al., 2022), and inserted into the Comprehensive Meta-Analysis Version III software (Borenstein et al., 2013).Four studies did not report the necessary statistical data and therefore were only included in the qualitative analysis section of the current review (Ksinan et al., 2020;Lozano et al., 2017;Seidel et al., 2022;Sun et al., 2022).It was only possible to extract ORs representing the magnitude of the association between e-cigarette use and subsequent alcohol and cannabis use as these were the psychoactive substances most utilised by participants enrolled in the seven studies pooled for the meta-analysis.ORs were adjusted for relevant covariates by authors of the respective studies.Covariates included: biological sex, age, other tobacco products use, other substances use, sensation seeking, and depression.
Considering that studies tested participants at multiple waves/time points, the ORs pooled from each study consisted of those assessing the association between e-cigarette use at the first time point measurement (e.g.wave 1), and cannabis/alcohol use at the second time point measurement (e.g.wave 2).Except for the study conducted by Audrain-McGovern et al. (2018), who reported results for cannabis use during the past 30 days at a 24-month follow-up period, the other studies included in the meta-analysis assessed cannabis and/or alcohol use during the past 30 days at 12 months follow-up (wave 2).The criterion for statistical significance was considered to be p < 0.05 (Cohen, 1994).As stated by Conti et al. (2019), "a random effect model was preferred over a fixed effect model as the studies included in the review were not functionally equivalent and the assumption that the true effect size was the same in all studies was not met" (p.144).Heterogeneity was assessed by I 2 and Q statistics, while publication bias was assessed by Fail-safe N test results (Orwin, 1983) and by the visual inspection of funnel plots.The Fail-safe N test results consist in "the number of additional 'negative' studies (studies in which the intervention effect was zero) that would be needed to increase the P value for the meta-analysis to above 0.05 (Rosenthal, 1979)" (Higgins et al., 2022).
Considering that the study conducted by Park et al. (2020) investigated the association between low and high e-cigarette use at baseline and the subsequent use of both cannabis and alcohol among adolescents, sensitivity analyses were conducted by just including ORs for individuals who reported a low use of e-cigarettes at baseline.

Qualitative analysis
A qualitative 'narrative synthesis' methodology (Dixon-Woods et al., 2005) was also employed to summarise the findings of the studies included in the review.Three authors (AAC, LL, and ZH) identified and pooled key results from each study and provided a descriptive summary of the findings (these are also reported in Table 3).

Screening strategy
The database search yielded 580 results.Limits were introduced to satisfy exclusion criteria, resulting in 520 entries.Titles and abstracts of all papers were then examined and assessed to remove irrelevant or duplicate entries.Full texts of the remaining 49 articles were then read and assessed against the inclusion and exclusion criteria.38 articles were excluded, including one article written in French.This resulted in 11 final articles eligible for the review.Of these 11 articles, 7 articles were included in the meta-analysis.This resulted in the flow diagram depicted in Fig. 1, which was made by following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 (Liberati et al., 2009).

Sociodemographic data
The sample consisted of 11 longitudinal studies, published between 2017 and 2022.Longitudinal studies were chosen to showcase the subsequent effect of using e-cigarettes.Nine of the papers gathered participants from the USA, one from Mexico, and one from Germany.Four of the cohorts were sourced from the Population Assessment of Tobacco and Health Survey (Bentivegna et al., 2021;Dai et al., 2018;Duan et al., 2022;Sun et al., 2022), three from the respective local region's schools (Audrain-McGovern et al., 2018;Lozano et al., 2017;Seidel et al., 2022), one from the local region's general population of adolescents (Park et al., 2020), one from the Monitoring the Future panel study (Evans-Polce et al., 2020), one from the Split for Science project (Ksinan et al., 2020) and one from the Happiness and Health Study (Lozano et al., 2021).The sample size of each cohort ranged from 801 to 19503 participants, and the total number of participants analysed was 67542.There was a total of 3889 ever e-cigarette users at baseline

Table 1
Inclusion and Exclusion Criteria.

Inclusion criteria Exclusion criteria
Longitudinal studies testing the association between ever e-cigarette use and subsequent psychoactive substance* use in young people Longitudinal studies only testing the association between ever e-cigarette use and subsequent use of tobacco products (e.g.tobacco cigarettes, cigars, hookahs) in young people ** Longitudinal studies testing participants with an age range of 10-24 years old Longitudinal studies not excluding nor statistically controlling for participants' baseline psychoactive substance use other than nicotine delivered through ecigarettes Studies written in the English language Studies written in any language other than English Systematic reviews, meta-analyses, conference reports, and meeting abstracts Cross-sectional and non-longitudinal study designs Studies testing participants who were tobacco cigarette smokers prior to starting e-cigarette use measurement, making up 5.8% of the participants, excluding ever e-cigarette users in Ksinan et al. (2020) as this statistic was not reported.51.2% of the total participants were female.The age range of participants at baseline measurement was 9-24 years.One study was conducted over a period of 1 year (Evans-Polce et al., 2020), one over a period of 1.7 years (Lozano et al., 2017), four studies over a period of 2 years (Audrain-McGovern et al., 2018;Dai et al., 2018;Park et al., 2020;Seidel et al., 2022), two over 3 years (Bentivegna et al., 2021;Sun et al., 2022), two over 4 years (Ksinan et al., 2020;Lozano et al., 2021), and one over 5 years (Duan et al., 2022).

Quality assessment of the included papers
Fig. 2 shows the results of the quality assessment conducted on the 11 pooled studies using the CASP Cohort Study Checklist (CASP, 2020).All the articles measured e-cigarette or substance use with surveys.The reliability of responses was not measured or reported in any of the studies.Despite this, questions (3) and (4) were graded as "Yes" if they met other requirements, where if 1) the questions in the survey were objective and/or specific, 2) subjects were classified into respective exposure groups using the same procedure, and 3) the measurement  L. Lau et al. methods were similar for all different groups.Paper quality was presented according to its results from the CASP checklist evaluation as a "Good", "Fair", or "Poor" rating.Table 2 outlines the quality ratings in addition to sociodemographic information and study characteristics.

Quantitative analysis
Cannabis use (Figs. 3 and 4).The pooled effect size (OR) of subsequent cannabis use for adolescent e-cigarettes ever users consisted in OR= 5.15 (95% CI= 2.94-9.00,p < 0.0001) compared to e-cigarette never users.Results of the sensitivity analysis conducted by including only the subgroup of adolescents who reported a low use of e-cigarettes at baseline in the study conducted by Park et al. (2020) consisted in OR= 3.77 (95% CI=2.77-5.13,p < 0.0001).I 2 and Q statistics revealed heterogeneity between the pooled studies (Q= 61.31, p < 0.0001, I 2 =90.21).Fail-safe N test results showed the absence of publication bias (N = 545).This was confirmed by the visual inspection of the Funnel plot (Supplementary Figure 1).
Alcohol use (Figs. 5 and 6).The pooled effect size (OR) of subsequent alcohol use for young e-cigarette ever users consisted in OR= 6.67 (95% CI= 2.85-15.59,p < 0.0001) compared to e-cigarette never users.Results of the sensitivity analysis conducted by including only the subgroup of adolescents who reported a low use of e-cigarettes at baseline in the study conducted by Park et al. (2020) consisted in OR= 4.02 (95% CI=3.33-4.85,p < 0.0001).I 2 and Q statistics revealed heterogeneity between the pooled studies (Q=26.72,p < 0.0001, I 2 =92.51).Fail-safe N test results showed the absence of publication bias (N = 229).This was confirmed by the visual inspection of the Funnel plot (Supplementary Figure 2).Bentivegna et al. (2021) tested the association between e-cigarette use during adolescence and other illicit substance use by collecting data over 3 waves: Wave 1 (September 2013-December 2014), wave 2 (October 2014-October 2015), and wave 3 (October 2015-October 2016).Results of a multivariable regression model (computed by excluding adolescents who reported combustible tobacco cigarettes smoking, any other illicit substances use, and diagnosis of ADHD at wave 1) revealed increased odds of cannabis smoking (OR 2.55, 95% CI: 1.85-3.52),cannabis used in electronic nicotine products (OR 2.05, 95% CI: 1.39-3.01),non-prescribed Ritalin/Adderall use (OR 2.13, 95% CI: 1.10-4.11),and polysubstance use (OR 2.67, 95% CI: 1.66-4.27)at waves 2 and 3 for e-cigarette ever users.Analyses were adjusted for other relevant covariates including age, biological sex, ethnicity, ever use of alcohol and cannabis, and school performance.However, e-cigarette use at wave 1 did not predict future use of non-prescribed sedatives, tranquilisers, and painkillers.Similarly, Seidel et al. (2022) revealed that adolescent e-cigarette users who never smoked/used cannabis at baseline were more likely to initiate cannabis use in the subsequent 18 months compared to adolescents who never used e-cigarettes (ARR = 1.83;CI: 1.48-2.25).Particularly, "34.5% of the ever e-cigarette users compared with 10.4% of the never e-cigarette users tried cannabis during the observation period.The association remained significant, even after adjusting for social-demographic and personality factors, other substance use, and peer cannabis use" (Seidel et al., 2022, p.1).Sun et al. ( 2022) reported e-cigarettes use among 9828 adolescents who were cannabis naive at baseline to be associated with increased likelihoods of both self-reported past 12-month and past 30-day cannabis use 1 year later.Analyses were adjusted for sociodemographic characteristics, other substances use, and environmental factors (e.g.family tobacco use).Lozano et al. (2021) tested the prospective association between e-cigarettes and alcohol use by gathering data every six months from 2013 to 2017.Use of e-cigarettes among adolescents who did not report any alcohol use at baseline was associated with greater odds of subsequent alcohol drinking initiation after adjusting for relevant covariates (e.g., other tobacco products use, biological sex, and age) (OR,3.95 [95%CI,).Previously, Lozano et al. (2017) did  that Z statistics is significantly different than 0; lower limit = lower limit of the 95% confidence interval for the effect size; upper limit = upper limit of the 95% confidence interval for the effect size; na = not applicable.

Qualitative analysis
Fig. 4. Sensitivity meta-analysis showing Odds Ratios for subsequent cannabis use among e-cigarettes never users compared to ever users.Z value = one sample Z statistic; p value = probability that Z statistics is significantly different than 0; lower limit = lower limit of the 95% confidence interval for the effect size; upper limit = upper limit of the 95% confidence interval for the effect size; na = not applicable.
Fig. 5. Odds Ratios for subsequent alcohol use among young e-cigarette never users compared to ever users.Z value = one sample Z statistic; p value = probability that Z statistics is significantly different than 0; lower limit = lower limit of the 95% confidence interval for the effect size; upper limit = upper limit of the 95% confidence interval for the effect size; na = not applicable.The same authors reported a significant correlation (p < .05) between earlier age at e-cigarette use (12-14 years) and subsequent heavier cannabis use.All analyses were controlled for other substance use, sensation seeking, and sociodemographic factors (e.g., age, biological sex).A summary of the above findings is provided in Table 3.

Key findings
This systematic review and meta-analysis was conducted with the aim to provide a qualitative and quantitative synthesis for the prospective association between e-cigarette use and other psychoactive substance use among young people.Qualitative results revealed that young e-cigarette ever users have higher odds of other psychoactive substance use within 1 or 2 years after baseline measurements compared to e-cigarette never users.These psychoactive substances included: stimulants (Ritalin, Adderall), cannabis, and alcohol.Meta-analytic results showed that young e-cigarette users have 5.15 times the odds (OR) of subsequently using cannabis, and 6.67 times the odds (OR) of subsequently using alcohol compared to e-cigarette never users.
These results provide support to previous systematic reviews and meta-analyses that found a cross-sectional association between young people, e-cigarette use and the use of cannabis and alcohol (Chadi et al., 2019;Rothrock et al., 2020).The results of the current review are also in line with the study conducted by McCabe et al. (2018).These authors investigated the association between age at onset of e-cigarettes use, tobacco smoking, and polysubstance use.Results of this study (McCabe et al., 2018) showed a stronger association between e-cigarettes use, tobacco use, and other psychoactive substance use among individuals who started e-cigarettes use in the 9th grade (14-15 years) compared to those who started e-cigarettes use in the 12th grade (17-18 years).
The current review adds to the body of literature by clearly showing a temporal association between e-cigarettes use during adolescence and an increased likelihood of subsequent psychoactive substance use and initiation.
Considering the previously discussed evidence from animal models (e.g., Ren and Lotfipour, 2019), a possible neurobiological gateway effect of nicotine on the adolescent/young adult brain (Kandel and Kandel, 2014) cannot be excluded.It is therefore possible that nicotine delivered through e-cigarettes may prime the developing adolescent brain to the use of other psychoactive substances by interacting with other neurotransmitter systems (e.g., endocannabinoid, opioidergic), and consequently increasing the rewarding effects of other drugs by enhanced activation of the reward circuitry (Kandel and Kandel, 2014) 2019) may provide support to this hypothesis as they found the association between e-cigarettes use and other psychoactive substance use to be stronger for early adolescents (14-15 years) compared to older adolescents and young adults.Indeed, it is well known that the developing early adolescent brain is particularly susceptible to the neuroadaptive and neurotoxic properties of nicotine (Ren and Lotfipour, 2019).
Another view proposed by Khouja and colleagues (2021) suggests that there may be a shared genetic aetiology for the use of e-cigarettes, combustible tobacco cigarettes, and possibly the use of other psychoactive substances.Particularly, these authors found polygenic risk scores (PRSs) for e-cigarettes and combustible tobacco cigarettes smoking initiation in 7859 young adults (mean age = 24 years) to also be associated with other risky behaviours such as gambling and an increased number of sexual partners, in addition to conduct disorder at 7 years of age (Khouja et al., 2021).
Ease of access to drugs may also influence the prospective association between e-cigarette use during adolescence and young adulthood and the subsequent use of other psychoactive substances.Notably, Duan et al. (2022) provided evidence that young e-cigarette ever users from US states where recreational cannabis use is legalised had higher chances of subsequently starting to use it compared to young e-cigarette ever users from states where the use of cannabis is outlawed.This 'common liability' effect has also been used to explain the strong association between e-cigarette use and subsequent tobacco smoking initiation (e.g., Soneji et al., 2017;Chan et al., 2020).Indeed, the neurobiological 'gateway' theory and the 'common liability' effect are Fig. 6.Sensitivity meta-analysis showing Odds Ratios for subsequent alcohol use among e-cigarettes never users compared to ever users.Z value = one sample Z statistic; p value = probability that Z statistics is significantly different than 0; lower limit = lower limit of the 95% confidence interval for the effect size; upper limit = upper limit of the 95% confidence interval for the effect size; na = not applicable.not mutually exclusive, and it is plausible to think that these longitudinal associations may be explained by both components.However, as also stated by Chan et al. (2020), it is not possible to determine how much of these relationships is causal (gateway effect) or due to common liability.

Strengths and limitations
As the first systematic review and meta-analysis on this topic conducted by pooling data only from longitudinal studies, this paper provides a unique overview of the available evidence on the association between e-cigarette use and the subsequent use of psychoactive substances.A comprehensive, systematic search was conducted to collate relevant papers from a wide range of sources.The papers were quality assessed in line with the CASP checklist, and their characteristics were aptly presented.Other strengths include that most of the pooled studies statistically controlled for covariates known to influence the uptake of psychoactive substances in young people (e.g., depression, sensation seeking, impulsivity, other substances use, other tobacco products use, biological sex, ethnicity).Furthermore, most studies excluded e-cigarette users who were cannabis and/or alcohol users at baseline.Another strength is that it was possible to conduct sensitivity analyses by removing the subgroup of high frequency e-cigarette users from the study of Park et al. (2020), which may have driven the overall effect, and by just inserting the subgroup of low-frequency e-cigarette users.Results from these sensitivity analyses still showed that e-cigarette ever users had significantly higher odds of using cannabis and alcohol in the future compared to e-cigarette never users.Publication bias was also low as shown by fail-safe N test results and by visual inspection of the funnel plots.On the other end, one of the main limitations of this study is the exclusion of papers not written in the English language, making it less representative of global data.Another limitation is that although this study explores the association of e-cigarette use with any subsequent psychoactive substance use (apart from conventional cigarettes), the reviewed articles only contained cannabis, alcohol, and Ritalin/Adderall as psychoactive substances.Thus, the potential association between e-cigarettes and subsequent opioid use or stimulants besides Ritalin/Adderall could not be explored.Bentivegna et al. (2021) did collect data on painkillers, but these were not specific to opioids, and the data gathered was combined with data on sedatives and tranquillisers.Another limitation is the low number of studies included in both the review and meta-analysis, which may limit the power of the current findings.Furthermore, most of the studies were conducted in the US, with just two studies in other countries such as Germany and Mexico.This may affect the generalizability of the findings as research carried out in countries with strict legislations for cannabis, alcohol and/or usage of e-cigarettes may yield different results, especially if considered under the 'common liability' effect framework.Additionally, legislation and jurisdictions related to recreational cannabis use may vary between US states and may have changed through the years.This may partially explain the inconsistency between the individual-level effect identified in the current paper (i.e.increased likelihood of subsequent cannabis use for e-cigarette ever users compared to e-cigarette never users) and the population-level effect (i.e.steady cannabis use among young people in the US despite an increase in e-cigarette use in the past decade) identified by the Monitoring The Future Survey (MTF) (Miech et al., 2023).This inconsistency may be also related to the estimated size of effect at the population level.That is, even if the current review reported a strong prospective association (in terms of odds ratios) between electronic cigarette use and other psychoactive substances use, this association may not be present at the population-level as the subgroup of young people included in our meta-analysis (i.e.young electronic cigarette users who have not used psychoactive substances) is small and may possibly represent < 10% of the entire youth population.
The inclusion of studies utilizing odds ratios as a measure of association rather than relative risk in the meta-analysis may be considered another drawback of the current review.According to Norton et al. (2018) " Although for rare outcomes odds ratios approximate relative risk ratios, when the outcomes are not rare, odds ratios always overestimate relative risk ratios, a problem that becomes more acute as the baseline prevalence of the outcome exceeds 10%" (p.84).Considering that the initiation of psychoactive substances is common among young people, the odds ratios computed by the meta-analytic calculations may represent an overestimation of the effect of the intervention and may be misinterpreted.We did not include studies in the meta-analysis that utilized relative risk as a measure of association, and it was not possible to calculate odds ratios directly from relative risk (Norton et al., 2018).Although, studies employing relative risk as a measure of association were included in the qualitative analysis section of the current review.
Finally, the findings of this review do not allow one to infer whether the longitudinal association identified between e-cigarette use during adolescence/young adulthood and other psychoactive substances use is due to a causal neurobiological and/or genetic gateway effect, common liability, or a combination of both.

Clinical relevance and future research
Cannabis, alcohol, and non-prescribed Ritalin/Adderall are well known to pose many health and social risks for adolescents (Volkow et al., 2014;Robison et al., 2017;Rehm et al., 2010).Therefore, the discovery and evidence of an association between e-cigarette use and subsequent psychoactive drug use would be important for clinical practice.E-cigarettes are widely purchasable and accessible, except for a few countries that have banned them in stores.They can be bought online easily without security or age verification methods.This suggests that clinical measures should be implemented to control e-cigarette use among adolescents to prevent the potential use of other psychoactive substances.This may involve public health measures such as discouraging, limiting, or banning those aged under 24 from using e-cigarettes.This could be done by altering advertisements for e-cigarettes to pose a more truthful representation of e-cigarettes or introducing government laws to limit or prevent the sale of e-cigarettes to adolescents.Other prevention strategies that may be effective include educational programs delivered at schools, and interventions using digital and/or interactive components (Liu et al., 2020).As emphasised by a review conducted by Liu and colleagues (2020), the educational interventions that are most effective in preventing or aiding the cessation of e-cigarettes and other tobacco product use among young people are those grounded on behaviour change and youth learning theories (e.g., transtheoretical model of behaviour change, social learning theory).
There is an overall lack of robust research on the prospective association between e-cigarettes and subsequent use of psychoactive substances.
Further longitudinal studies should be conducted to substantiate this temporal association.These studies should also utilize relative risk as a measure of association in order to avoid an overestimation of the effect of the intervention and subsequent misinterpretation that may occur when results are reported in odds ratios (Norton et al., 2018).Research could explore the presence of a causal relationship, the association between the degree of e-cigarette use and subsequent substance use, the influence of the type of e-cigarette, and/or the subsequent use of opioids.
Studies on a more global scale could also explore the role of culture in this association.
Several studies have also reported a cross sectional association between e-cigarettes use and mental health problems, in addition to other psychoactive substances use, in young people.For example, Grant et al. (2019) reported young e-cigarettes users to be more likely to present mental health conditions such as ADHD, PTSD, gambling disorder, anxiety, and to present higher levels of impulsivity compared to young e-cigarettes never users.Similarly, Cwalina et al. (2021) reported that young multiple tobacco product users who use e-cigarettes more often than other tobacco products (e.g.combustible tobacco cigarettes) have greater odds of presenting depressive and anxiety symptoms.Therefore, future research should also explore the prospective association between e-cigarettes use, other tobacco products use, and mental health problems in young people.

Conclusion
In conclusion, there is overwhelming evidence to suggest that the ever use of e-cigarettes as an adolescent or young adult is associated with increased likelihood of subsequent psychoactive substance use, mainly cannabis, alcohol, or Ritalin/Adderall, compared to the never use of e-cigarettes, when adjusted for confounding variables.
This prospective association is stronger in males or in a location with legalised recreational cannabis use.It also differs by country, considering that this association was found to be not significant in Mexico, but significant in the US and Germany.

A
comprehensive systematic search was performed in March 2022 to identify studies published from January 1st 2002 to March 1st 2022.The Population, Intervention, Comparison, and Outcome (PICO) framework (Centre for Evidence Based Medicine, 2022) was utilised to construct the search strategy.The PICO is outlined below: Population: Young people aged 10-24.

Fig. 1 .
Fig. 1.PRISMA Flow Diagram of the Search and Screening Process.

Fig. 3 .
Fig.3.Odds Ratios for subsequent cannabis use among young e-cigarette never users compared to ever users.Z value = one sample Z statistic; p value = probability that Z statistics is significantly different than 0; lower limit = lower limit of the 95% confidence interval for the effect size; upper limit = upper limit of the 95% confidence interval for the effect size; na = not applicable.
e-cigarette users compared to e-cigarette non-users (aOR=3.61,95%CI=1.11,11.71).Individuals who reported cannabis use at baseline were excluded from the analyses.Furthermore, analyses were controlled for biological sex, age, and binge drinking covariates.Likewise,Ksinan et al. (2020) revealed a prospective association between e-cigarette use during late adolescence and young adulthood (18-24 years) and cannabis use within 1, 2, and 3 years after baseline measurements.Analyses were adjusted for other tobacco products use, alcohol use, polysubstance use, depressive symptoms, age, biological sex, and ethnicity(Ksinan et al., 2020).Audrain-McGovern et al. (2018) reported that ever e-cigarette users are more likely to initiate cannabis use at 24 months after baseline measurements compared to never e-cigarette users (OR=3.63;95% CI= 2.69-4.90).Intriguingly, their findings revealed that early adolescent e-cigarette users (age 14-15 yrs) initiated the use of three different forms of cannabis including combustible cannabis, vaping cannabis, and cannabis edibles.In accordance with the previous cited studies, their analyses were conducted by controlling for relevant covariates known to influence cannabis uptake such as sociodemographic characteristics (e.g.age, biological sex, peer tobacco use, familial history of substance use) and interpersonal factors (depressive symptoms, impulsivity).The longitudinal study conducted byDai et al. (2018) also showed that e-cigarette use among adolescents (who reported to be never cannabis users at baseline) at wave 1(2013)(2014) predicted cannabis use at wave 2 (2014-2015)(aOR= 1.9; 95% CI: 1.4-2.5).
. The findings of the studies conducted by Audriain-McGovern et al. (2018) and by McCabe et al. (

Table 2
Demographics, Study Characteristics and Quality Ratings of the Included Papers.

Table 3
Summary of the Papers.

Table 3
(continued ) (continued on next page) L.Lau et al.