Introduction

Cigarette smoking is the leading cause of premature death in the United States [1]. Despite the documented negative health effects of use, the tobacco control landscape faces an unprecedented diversity of tobacco and nicotine-containing products [2, 3]. Beyond combustible cigarettes, these products include pipe and hookah, cigars, cigarillos, little cigars, bidis, smokeless tobacco, and newer products, such as electronic nicotine delivery systems (ENDS) and heat-not-burn options. ENDS are currently regulated as a tobacco product under the Food and Drug Administration’s deeming rule [4]. Non-cigarette tobacco products, particularly ENDS, are often perceived as safer than cigarettes and may appeal to individuals, including those youth who have traditionally been considered lower risk for initiating tobacco product use [5, 6]. These changes in the tobacco market coincide with alarming reports of increasing poly use of combustible and non-combustible tobacco products, which are also associated with increases in risky behavior among youth and lower likelihood of quitting tobacco use [6–10].

There is an ongoing debate among tobacco control advocates regarding the potential association of the initial experimentation with various tobacco products and patterns of subsequent youth tobacco use [11]. However, there is limited understanding of whether initiation with a specific tobacco product, including ENDS, increases the odds of using other tobacco products or poly-tobacco use [10, 11]. From a population health perspective, several tobacco control researchers conceptualize that the harm from tobacco use can be viewed from a continuum of risk perspective. On this continuum, ENDS and smokeless tobacco (i.e., non-combustible) are considered less harmful than combustible tobacco [12]. However, recent evidence suggests ENDS use attenuates attempts to quit, maintains addiction to nicotine, and contributes to the development of specific health problems such as cardiovascular disease [13] or vaping-related pulmonary illness [14–16]. There also may be differences in youth perceptions that are associated with the choice of product and ultimately their pattern of use [6]. Additionally, the factors associated with patterns of product use have implications for different health campaigns based on the class of products used (e.g., targeting smokeless tobacco products is unrelated to secondhand smoking effects).

Understanding the association of initial experimentation with a tobacco product and subsequent patterns of tobacco use among youth is important to informing primary and secondary prevention activities for youth in the US. Thus, the purpose of this study is to assess the association between initial experimentation and subsequent tobacco use among a nationally representative sample of US youth. Based on prior research, [7–10, 17–22] we hypothesize that initial experimentation with a particular tobacco product will be associated with youth subsequent tobacco use patterns. For example, experimentation with non-combustible products like ENDS will be associated with the likelihood of continued use of the same product class rather than using other tobacco products (e.g., combustible tobacco).

Materials and methods

Results

Overall, 45.2% were female; 76.8% were heterosexual; 50.9% were aged ≥ 17 years old; 62.5% were NH White; and 22.2% were Hispanic (see Table 1). Among our sample, the most common initial tobacco product tried was ENDS (44.7%), followed by cigarettes (35.0%) and cigars (8.6%). Nearly half of the respondents (43.2%) progressed from experimentation to subsequent tobacco use. Among the youth who first experimented with ENDS (n = 836), one-third (32.8%) continued to be subsequent tobacco users, which is considerably lower than the proportion that continued to subsequent use after experimenting with other products. Approximately half of the youth who first experimented with other products continued to be subsequent tobacco users (47.4% for cigarettes, 61.7% for cigar group [cigars, cigarillos, or little cigars], 59.9% for smokeless dip and chewing tobacco, and 54% for hookah). Further Fig 1 illustrates the tobacco experimentation groups and their subsequent tobacco use pattern. The flow from experimentation to current use behavior shows that a higher proportion of initial ENDS experimenters were current non-users than those who had initially experimented with other combustible tobacco products. Among current users (n = 934), more than half (56%) were poly-tobacco users. More specifically, 31% were ENDS plus combustible poly-users, 8% were combustible only poly-users, and 17% were other tobacco poly-users. Exclusive use accounted for 44% of the sample with 29% exclusive ENDS users and 15% exclusive combustible users (i.e., using only one combustible product).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Initial experimentation by tobacco product and subsequent current tobacco use pattern.

https://doi.org/10.1371/journal.pone.0308964.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sample characteristics and tobacco use initial experimentation history among youth in the United States (2017).

https://doi.org/10.1371/journal.pone.0308964.t001

Significant differences in sex, age, sexual orientation, residential area, public assistance, and first tobacco product tried were found between current tobacco use groups. For example, more exclusive ENDS and ENDS plus combustible tobacco users were male and reported that their families did not receive public assistance. More exclusive ENDS users (71.8%) reported that ENDS were the first type of product used, but cigarette experimenters who continued to use tobacco were mostly exclusive combustible mono-users, ENDS plus combustible poly-users or poly combustible only users (56.6%, 45.1%, 60.9%, respectively, from each use group). In addition, more ENDS plus combustible poly-users (12.7%) reported that hookah was the first type of tobacco used, compared to all other groups (Range: 2.2–9.3%; see Table 1).

Table 2 presents the association of subsequent tobacco use pattern with first tobacco product tried. Those who first tried cigarettes, cigars, or hookah compared to ENDS were less likely to be exclusive ENDS users [Relative Risk Ratio (RRR) = 0.46, 95% CI = 0.28, 0.73; RRR = 0.32, 95% CI = 0.13, 0.81; RRR = 0.33, 95%CI = 0.14, 0.79, respectively]. The relative risk of combustible mono-use was highest among the cigar group experimenters (RRR = 17.3, 95% CI = 5.04, 59.34), followed by hookah (RRR = 7.30; 95% CI = 2.04, 26.15) and cigarettes (RRR = 6.90; 95% CI = 2.62, 18.19). The least risk to be an exclusive combustible or an ENDS plus combustible current user was among the smokeless dip and chewing tobacco experimenters (RRR = 0.18, 95% CI = 0.04, 0.91; RRR = 0.10, 95% CI = 0.01, 0.66, respectively). Those who first experimented with cigarettes, in comparison to ENDS, were consistently more likely to belong to one of the combustible tobacco product groups (RRR range: 2.95–8.63; p < .001). In contrast, compared to non-current tobacco users, those who were currently mono-combustible tobacco product users only were more likely to have first experimented with a combustible tobacco product in comparison to ENDS (RRR = 6.90; 95% CI = 2.62, 18.19 for cigarettes; RRR = 17.3; 95% CI = 5.04, 59.34 for cigars; and RRR = 7.30; 95% CI = 2.04, 26.15 for hookah) as shown in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Association between experimentation and current tobacco use pattern among youth in the United States (2017).

https://doi.org/10.1371/journal.pone.0308964.t002

Discussion

ENDS followed by combustible cigarettes were the most frequently reported first used tobacco products by our nationally representative sample of US youth. About 45% of youth experimenters become subsequent tobacco users. A notable finding is that poly tobacco use is becoming established among youth, accounting for more than 50% of the estimated use prevalence [10]. As anticipated, the subsequent pattern of tobacco use was notably linked to the initial tobacco product tried. Specifically, in our study, initial ENDS use was associated with a greater likelihood of being a subsequent exclusive ENDS user, and initial combustible experimentation was closely associated with subsequent combustible product use.

The evidence aligns with the expectation that individuals are inclined to persist with the first tobacco product they tried [23, 24]. Aligning with our hypothesis, the findings suggest that current tobacco users who initially experimented with a combustible tobacco product were more likely to be current combustible only or poly-tobacco users, and less likely to be exclusive ENDS users. Those who initially experiment with combustible products (cigarettes, cigars or hookah) seem likely to be youth who are either (a) high-risk for transitioning to poly-tobacco use or combustible tobacco products as part of their propensity for sensation-seeking and other risky behavior [25] or (b) prone to become more dependent on nicotine, which eventually leads to use of combustible or combustible plus ENDS products [10]. The initial experimentation phase may be of high importance to the youth tobacco experience and the resulting tobacco use pattern; youth who choose to initially experiment with a certain product (e.g., cigarettes) do not realistically consider its harms or addictive potential and may become dependent on this class of products [25]. More specifically, in our sample, half of the experimenters of non-ENDS products reported currently using tobacco in comparison to one-third of those who initially experimented with ENDS.

In the US, the majority of hookah users concurrently consume other tobacco products [23, 24, 26]. Those who used hookah first were more likely to ultimately be part of the ENDS plus combustible and other poly-tobacco users’ profiles rather than the exclusive combustible use group. Hookah use is a social experience, often taking place in hookah cafes, a setting that may be especially attractive to youth [27, 28]. However, the lack of portability of conventional hookah makes it challenging to sustain nicotine addiction from hookah alone. Thus, hookah may attract youth initially with its social appeal, with users then seeking more readily accessible products (e.g., cigarettes, ENDS) and becoming mixed (ENDS plus combustible) or poly-combustible tobacco users.

Finally, ENDS experimentation is associated with exclusive ENDS use rather than combustible or poly tobacco use. This finding could be interpreted in at least three ways. Perhaps those who experiment with ENDS may remain less dependent on nicotine and thus continue use of ENDS products that yield less nicotine than combustible cigarettes [10]. However, given the high nicotine content of popular pod-based products, such as late-generation ENDS or JUUL, this interpretation seems unlikely [29]. Another possibility is that ENDS users are mainly attracted to flavors and thus continue ENDS use exclusively, rather than switching to or adding combustible products without the variety of attractive flavors [30]. Alternately, as noted in previous work, perhaps youth perceive ENDS as less harmful than combustible tobacco; [31, 32] in other words, lower risk-oriented youth may be less likely to progress to tobacco-related behavior that they perceive as riskier (i.e., concomitant use of combustible or poly-tobacco use) [5]. This study cannot further test the latter hypothesis due to the cross-sectional nature of the data. Nevertheless, we divided the current tobacco use groups by exclusive or poly-use of products across all groups, and ENDS experimenters seemed to have the lowest likelihood for possessing poly-tobacco use profiles. But, given the low prevalence of non-combustible product use other than ENDS, we were not able to conduct necessary statistical analyses to determine predictors of membership for this tobacco use group independently, and there may be specific aspects that are not accounted for when incorporating these participants into the all-other tobacco use profile. Further, being on public assistance and from a rural area was inversely associated with using protobacco use, which may reflect the importance of market availability and affordability of tobacco products among youth. There is mixed evidence regarding ENDS serving as a gateway to subsequent use of other tobacco products, especially cigarettes [33, 34]. Nevertheless, it seems that initial experimentation with any tobacco product is important in shaping the subsequent tobacco use profile of youth.

A recent study found that youth experimenting with a new tobacco product can lead to various outcomes. Over time, users perceive the product as less harmful, and this experience influences their perceptions of harm for other tobacco products. Additionally, trying a new tobacco item increases susceptibility to experimenting with different tobacco products. These tendencies are especially notable among youth, indicating a complex interplay between experimentation, evolving perceptions, and heightened openness to using various tobacco products [35]. Although these connections have been consistently reproduced in existing literature, there hasn’t been an exhaustive exploration of empirical data to thoroughly investigate the potential mechanisms behind the longitudinal shifts in tobacco use among youth. The most recent study on NYTS survey demonstrated that, despite advancements, elements fostering tobacco product use among U.S. youths, such as the presence of flavored options, accessibility to tobacco products, exposure to marketing, and misconceptions about the risks associated with tobacco product use, remained widespread in 2021 among US youth [36].

The cross-sectional study design predisposes the findings to certain limitations. No causal statements can be made about the observed associations. We also included the small number of exclusive chew and smokeless dip users within the other tobacco use profile, which was mainly poly-tobacco users, instead of having a dedicated group, given their low prevalence. Caution is needed in interpreting those RRRs with wide confidence intervals due to small sample size and thus greater uncertainty. Data were self-reported; therefore, answers could have been affected by recall bias. This survey was also conducted prior to the FDA approval of heat-not-burn products in the US market in April 2019 [37], thus such products were not included in our study. In addition, some other aspects may have impacted our findings such as the short duration of the study that may preclude the authors from fully assessing the transition from experimenting phase to current use phase. Lastly, other potentially influential factors, such as parental tobacco use, age of first tobacco experimentation and level of nicotine dependence, were not accounted for in the analysis. Future longitudinal research could appraise the temporality of the association between first tobacco product used and transition to subsequent tobacco use.

The current findings can inform primary prevention activities. Given that youth seem to develop a preference for the tobacco product they first experiment with [23, 24], the tobacco control community needs to carefully consider the context and content of primary prevention efforts that are communicated to youth. Collectively, initial experimentation with tobacco products other than ENDS seemed to be significantly associated with continued use of tobacco products, suggesting that a higher proportion of initial ENDS experimenters (particularly from early cohorts potentially driven by the novelty of these products, i.e., early adopters) do not subsequently become current tobacco users [5]. Other than considering the main mandate of deterring youth from experimenting with any tobacco product, specific prevention messages could target those who tried specific tobacco products, with a focus on preventing them from moving to continued use of tobacco. These efforts could potentially deter lower risk-oriented youth from continuing to experiment with other tobacco products or becoming established tobacco users. Finally, tobacco control advocates should convey clearer assessment of the health implications of specific tobacco products to youth, rather than adopt a generalized less-risky message (e.g., comparison to combustible cigarettes, potential for harm reduction) that may result in youth perceiving ENDS as safer products [6].

Conclusions

Prior research suggested that the pattern of tobacco use among youth was driven by their nicotine addiction [10]. Our findings suggest an additional perspective: The experience of initial experimentation with tobacco products may play a role in developing subsequent current tobacco use patterns among youth. Understanding the reasons behind initial product choice may help in understanding the reasons for subsequent current tobacco product use or the transition in patterns from combustible to non-combustible tobacco products or vice versa. Regulatory authorities and health campaign messaging should convey the risk of specific products to help prevent youth from experimenting with products they may consider safer options. Further understanding the reasons behind initial product choice for experimenting may help in understanding the reasons for subsequent current tobacco product use, thus informing youth prevention efforts.