Longitudinal trends in the past 30-day co-use of nicotine/tobacco, alcohol, and cannabis among youth and adults in the PATH study
Center for Alcohol and Addiction Studies, Brown University School of Public Health, USA
Abstract
Introduction:
The co-use of substances confers risks above single-product use and has significant public health implications. This study investigated trends in past 30-day co-use of nicotine/tobacco products with alcohol and cannabis in the US using Population Assessment of Tobacco and Health Study data from Waves 4–6 (December 2016-November 2021).
Methods:
All wave 4–6 PATH participants age 15+ were included in analyses. Changes across wave in past 30-day co-use of cigarettes, e-cigarettes, and other tobacco products (OTP; cigars, filtered cigars, smokeless, hookah, snus, and cigarillo) with alcohol and cannabis, moderated by age (15–17,18–24, 25–34,35–64, 65+), and controlling for demographics were investigated.
Results:
Changes in co-use of tobacco products with cannabis and alcohol varied across age and product. Cigarette and alcohol co-use was most prevalent across all adult ages, with rates declining over time among young adults (18–24, 25–34) but stable in older adults (65+). Rates of e-cigarette and alcohol co-use increased among young adults, possibly supplanting alcohol and cigarette co-use. E-cigarette and alcohol co-use was the most popular pattern of co-use in youth, with initially increasing and then declining prevalence. Co-use of e-cigarette and cannabis increased at Wave 5 among those 15–17, 18–24, and 25–34, although this increase lessened in all groups except those age 25–34 at Wave 6. Cigarette and cannabis co-use rates, and co-use rates of OTP with both cannabis and alcohol were generally stable or decreasing.
Conclusions:
Findings highlight the complex interplay between substance use patterns and developmental stages and the dynamic nature of co-use in ever-evolving tobacco and cannabis marketplaces.
Article notes
Untitled section
Issue date 2026 May.
1.Introduction
Investigating concurrent use (co-use) of nicotine and tobacco products (NTP) with alcohol and cannabis is imperative because combined effects of NTP with other substances can exacerbate adverse health outcomes, including respiratory diseases, (Meier and Hatsukami, 2016; Correa et al., 2020) cardiovascular problems, (Winhusen et al., 2020; Moons et al., 2019) and mental health disorders, (Cohn et al., 2018) among other public health risks. Rates of NTP and cannabis co-use were 6.2% among US adults in 2019, (Rubenstein et al., 2024) and associated with increased risk of negative consequences. Co-use of cigarettes and electronic cigarettes (i.e., e-cigarettes) with cannabis has been linked to poorer smoking cessation outcomes and increased health risks (Wang et al., 2018; Berg et al., 2021). Many individuals who co-use cannabis and NTP use cannabis more heavily (Reboussin et al., 2021) Those who used cannabis and NTP together were also more likely to use cannabis more frequently and in greater quantities, to sell cannabis, and to experience greater negative cannabis-related consequences than those who did not (Tucker et al., 2019). Further, among individuals aged 12–64 years, co-use of cannabis and cigarettes was associated with a greater likelihood of nicotine dependence, with adolescents who co-used being 3.5 times more likely to be dependent on nicotine (Wang et al., 2016).
Similarly, co-use of NTP with alcohol can exacerbate health risks and impair substance use treatment outcomes. Among adolescents in treatment for substance use disorders, those who smoked cigarettes were more likely to return to alcohol and cannabis use relative to individuals who quit smoking (de Dios et al., 2009). College students who used e-cigarettes were more likely to have psychiatric diagnoses, substance use disorders, and to report heavier alcohol use (Hefner et al., 2019). Among adults, individuals who smoke cigarettes and those who formerly smoked cigarettes are more likely to engage in hazardous drinking than adults who never smoked (McKee et al., 2007). Exposure to alcohol and NTP can also interact to reciprocally increase the consumption, craving, and dependence (Verplaetse and McKee, 2017). Further, the co-use of NTP with alcohol creates synergistic risks for numerous adverse health outcomes, including aerodigestive tract cancers (Anantharaman et al., 2011).
Patterns of NTP co-use with alcohol or cannabis are likely to vary across the lifespan, due to differences in social norms, (Cooper et al., 2016) contexts, (Agaku et al., 2020) and substance accessibility (Giovenco et al., 2019; Ribisl et al., 2017). Among adolescents and young adults (age 18–34), co-use patterns may be driven by experimentation motives, (Reboussin et al., 2021) social motives, physiological effects (e.g., relaxation or compensatory effects for other substances), or coping, (Pedersen et al., 2021) each of which vary over the life course. Social contextual factors that precede co-use of cigarettes with alcohol or cannabis may include settings like parties, which could increase co-use due to ubiquitous access (Lajtha and Sershen, 2010) but are more common during adolescence and young adulthood. E-cigarettes are perceived as more socially acceptable in younger individuals, which could result in higher co-use rates of e-cigarettes with cannabis and alcohol, especially in peer-influenced environments (Choi et al., 2017). Research investigating the co-use of e-cigarettes and cannabis has found shared social context to be the most endorsed reason for their co-use (Smith et al., 2021). In adults, patterns of co-use are more varied; some studies report greater co-use rates, particularly in specific age or socio-economic subpopulations (Rubenstein et al., 2024). While tobacco use rates have decreased overall, concomitant decreases in co-use in adults have been attenuated, (Leung et al., 2022) potentially due to the increasing association observed between use of NTP and other substances (Daw et al., 2013).
The Population Assessment of Tobacco and Health (PATH) Study (Hyland et al., 2017) is an ongoing longitudinal cohort study of noninstitutionalized US youth and adults examining life-course transitions in tobacco use and their health sequelae. There are currently 7 waves (Ws) of PATH data and several studies of various combinations of tobacco, alcohol, and cannabis co-use among either youth or adults (Cohn et al., 2018; Cohn et al., 2019). However, few studies have integrated youth and adult data, leveraged recent data, or focused on longitudinal trends in co-use versus intra-individual transitions over time. Thus, in the current study, we aimed to characterize population-representative longitudinal changes in the past 30-day co-use of NTP with alcohol and cannabis in the PATH Study W4-W6, including data from both adults and youth. Specifically, we aimed to measure the co-use trends of combustible cigarettes, e-cigarettes, and other tobacco products (OTP; i.e., cigarillos, cigars, filtered cigars, pipe, hookah, snus, pouches) with alcohol or cannabis. While no a-priori hypotheses were specified, close monitoring of changing co-use patterns is essential to adequately allocate resources for interventions tailored to the subpopulations with greatest need, especially those groups with co-use at high, stable, or increasing rates.
2.Method
2.1.Participants
This project is a secondary analysis of the PATH study. Public Use Files (PUF) were used from W4 (December 2016-January 2018), W5 (December 2018-November 2019), and W6 (March 2021-November 2021). Participants included adults and youth aged 12+. The PATH study involves a nationally representative longitudinal cohort that is replenished every 3 waves (i.e., additional participants enrolled due to attrition and youth aging). As appropriate weighting of the data in PATH is cohort-dependent, investigating longitudinal changes in population cross-sections over time is partially limited by the aging out of individuals in younger age strata within a given cohort (for example, in W6 there are very few remaining 12–14 year olds from the W4 cohort). Due to this limitation, inherent to the design of the PATH study, we limited our analyses to youth and adults age 15+ in W4-W6 to avoid underestimation in the 12–14 age range, and we omitted W7 to avoid having to further restrict the sample.1 Detailed information on survey procedures, response rates, weighting, and participants are published elsewhere (Hyland et al., 2017). Additional details on procedures are publicly available (USDHHS, 2024). All adult participants provided informed consent; assent was requested for all youth participants with parental permission.
2.2.Measures
Cigarette use.
This study-derived variable captured whether participants had smoked at least one “cigarette” in the past 30 days (coded 0 = no; 1 = yes).
E-cigarette use.
This study-derived variable captured whether participants had used an “electronic nicotine product” in the past 30 days (0 = no; 1 = yes).
OTP use.
We derived an OTP use variable (coded 0 = no to all; 1 = yes to one or more) based on responses to items assessing past 30-day participant behavior, including: “smoked a pipe”, “smoked a traditional cigar”, “smoked a cigarillo,” “smoked a filtered cigar”, “smoked hookah”, “used smokeless tobacco”, or “used snus”. Although the youth questionnaire also queried the use of kretek and bid, their use was infrequent and not were queried in the adult questionnaire and thus excluded from the computed OTP use variable.
Alcohol use.
Participants were asked a binary response question “Have you used alcohol within the past 30 days?”, which was dichotomized (0 = no; 1 = yes).
Cannabis use.
Participants were asked “Have you used marijuana, hash, THC, grass, pot or weed within the past 30 days?” which was dichotomized (0 = no; 1 = yes).
Co-use of tobacco, alcohol, and cannabis.
We computed six dichotomized co-use variables based on responses to the above-defined past 30-day alcohol, cannabis, and tobacco variables. A variable was created for each combination of [cigarettes, e-cigarettes, or OTP] with [alcohol or cannabis]. Co-use was coded as 1 if past 30-day use was endorsed for both products (yes/yes), 0 if past 30-day product use was not endorsed for one or more of the products (no/no; yes/no; no/yes), and missing if reporting data were absent on either product.
Time.
Time was indexed by wave (W4-W6), which was treated categorically.
Demographic characteristics.
Demographics included self-reported sex, race, ethnicity, household income, level of education,2 and employment status (see Table 1).3 Self-reported binned age categories at a given wave were further binned into developmental categories broadly corresponding to high school (15–17), early young adulthood (18–24), late young adulthood (25–34), middle adulthood (35–64), and older adulthood (65+). We used older adulthood (65+) as the reference category as this age band generally had the most stable patterns of use and co-use over time.
| Wave 4 (N = 41,094) | Wave 5 (N = 39,396) | Wave 6 (N = 33,830) | ||||
|---|---|---|---|---|---|---|
| Unweighted | Weighted | Unweighted | Weighted | Unweighted | Weighted | |
| N (%) or % (weighted) | N (%) or % (weighted) | |||||
| N (%) or % (weighted) | ||||||
| Age | ||||||
| 15–17 | 7,456 (18.1%) | 4.9% | 6,710 (17.0%) | 4.7% | 4,314 (12.8%) | 4.4% |
| 18–24 | 11,213 (27.3%) | 11.8% | 11,355 (28.8%) | 11.4% | 10,633 (31.4%) | 11.2% |
| 25–34 | 6,874 (16.7%) | 17.0% | 6,928 (17.6%) | 16.3% | 6,443 (19.0%) | 16.1% |
| 35–64 | 12,444 (30.2%) | 47.7% | 11,271 (28.6%) | 47.1% | 9,485 (28.0%) | 46.7% |
| 65+ | 3,107 (7.6%) | 18.6% | 3,132 (8.0%) | 20.4% | 2,955 (8.7%) | 21.7% |
| Sex | ||||||
| Male | 20,327 (49.5%) | 48.3% | 19,296 (49.1%) | 48.3% | 16,283 (48.1%) | 48.1% |
| Female | 20,710 (50.5%) | 51.7% | 20,036 (50.9%) | 51.7% | 17,494 (51.8%) | 51.9% |
| Race | ||||||
| White alone | 28,156 (71.0%) | 76.5% | 26,690 (70.5%) | 76.3% | 22,835 (70.3%) | 76.1% |
| Black alone | 6,646 (16.8%) | 12.7% | 6,417 (16.9%) | 12.8% | 5,448 (16.8%) | 12.8% |
| Other | 4,833 (12.2%) | 10.8% | 4,772 (12.6%) | 10.9% | 4,206 (12.9%) | 11.1% |
| Ethnicity | ||||||
| Hispanic/Latino | 8,902 (22.0%) | 16.3% | 8,903 (23.0%) | 16.5% | 7,654 (23.1%) | 16.8% |
| Non-Hispanic/Latino | 31,553 (78.0%) | 83.7% | 29,762 (77.0%) | 83.5% | 25,463 (76.9%) | 83.2% |
| Household Income | ||||||
| Less than $10 k | 6,019 (15.7%) | 11.4% | 5,469 (14.8%) | 10.4% | 3,424 (10.8%) | 8.0% |
| $10 k – $24,999 | 7,861 (20.4%) | 18.6% | 6,976 (18.8%) | 17.2% | 4,981 (15.7%) | 14.5% |
| $25 k – $49,999 | 8,794 (22.9%) | 22.6% | 8,393 (22.6%) | 22.3% | 7,086 (22.3%) | 21.7% |
| $50 k – $99,999 | 8,988 (23.4%) | 26.4% | 8,811 (23.8%) | 26.6% | 8,365 (26.4%) | 28.5% |
| More than $100 k | 6,791 (17.7%) | 21.0% | 7,424 (20.0%) | 23.5% | 7,882 (24.8%) | 27.4% |
| Education1 | ||||||
| Less than high school | 11,696 (28.6%) | 15.2% | 10,565 (26.9%) | 14.7% | 7,302 (21.6%) | 13.5% |
| GED | 2,096 (5.1%) | 4.8% | 1,917 (4.9%) | 4.9% | 1,304 (3.9%) | 4.1% |
| High school graduate | 8,114 (19.8%) | 22.6% | 8,069 (20.5%) | 21.9% | 7,194 (21.3%) | 21.9% |
| Some college/associates degree | 11,894 (29.0%) | 29.5% | 11,538 (29.4%) | 29.5% | 10,624 (31.5%) | 30.0% |
| Bachelor’s or advanced degree | 7,146 (17.5%) | 27.9% | 7,177 (18.3%) | 29.0% | 7,326 (21.7%) | 30.5% |
| Employed/working | ||||||
| Yes | 25,597 (62.6%) | 63.6% | 25,154 (64.2%) | 64.1% | 21,571 (63.9%) | 61.9% |
| No | 15,285 (37.4%) | 36.4% | 14,045 (35.8%) | 35.9% | 12,179 (36.1%) | 38.1% |
2.3.Analysis plan
Data were “stacked” across W4-W6 and both youth and adult samples. We treated the data as longitudinal cross-sections (a time series of cross-sectional observations) that approximated the US civilian noninstitutionalized population (CNP) by using Cohort 4 single-wave weights (i.e., participant weights varied by the observation wave when stacked), and replicate weights, (McCarthy PJ. Pseudoreplication: further evaluation and applications of the balanced half-sample technique. Vital Health Stat 2., 1969) consistent with recommended procedures. We also used Fay’s method for variance estimation (Fay = 0.3) (Judkins, 1990). Data analyses were conducted in R 4.4.1 (R Core Team. R, 2017) using the survey package (Lumley, 2004).
Sample unweighted and weighted descriptive statistics were computed across W4-W6. Although W7 (January 2022 – April 2023) was not included in the final models analytically, we did compute product use and co-use rates by age group for W7, using the Cohort 4 single-wave weights, for all age bands ≥ age 18 (Supplemental Table 1). We fit six survey-weighted generalized linear models specifying a quasibinomial distribution with a logit link and regressing each of the dichotomous co-use variables onto age, time (Wave), the interaction between age and time, and covariates. Specifying participant IDs as the clustering unit in the design facilitated the inclusion of repeated observations while accounting for the unique covariance between observations within a participant. Weights were rescaled to sum to one for numeric stability. We utilized an unstructured covariance matrix; all available data informed the estimation of the respective correlation matrix. We also refit the models to estimate the simple main effects of time for all age bands (Aiken et al., 1991).
3.Results
Table 1 includes the number of participants in W4-W6 as well as unweighted and weighted sample characteristics. The sample was approximately evenly split across males and females, primarily White Non-Hispanic, and the age band most represented was adults ages 34–64. Income brackets and educational attainment were well-distributed.
Between W4 and W6, prevalence rates (reported as the proportion of youth and adults ages 15+ in the CNP) of cigarette use were highest in adults aged 25–35 but decreased over time (0.29 [W4], 0.20 [W6]), and were least prevalent in youth aged 15–17 (0.05 [W4], 0.01 [W6]). E-cigarette use peaked in W5 and declined in W6 for those aged 15–17 (0.14 [W5], 0.07 [W6]), 18–24 (0.26 [W5], 0.20 [W6]), and 25–34 (0.15 [W5], 0.14 [W6]), while remaining generally stable in low for adults aged 35–64 (0.05 [W4,W6], 0.06 [W5]) and 65+ (0.01 [W4–6]). OTP use declined across all age bands over the three waves, and was generally most prevalent in adults aged 18–24 (0.21 [W4], 0.10 [W6]) and least prevalent in youth aged 15–17 (0.04 [W4], 0.01 [W5]). Although patterns of co-use varied across age bands, patterns involving alcohol and tobacco were generally more prevalent than those involving cannabis and tobacco, although with increasing parity in later waves. A comprehensive summary of single product use and past 30-day co-use rates by age band, including W7 rates for adults age 18+, is presented in Supplemental Table 1.
3.1.Co-use with alcohol
Cigarette and alcohol co-use (Fig. 1a and Supplemental Table 2).
For older adults aged 65+, co-use of cigarettes and alcohol was generally stable over time (W5: OR = 1.18 [0.97,1.43]; W6: OR = 1.10 [0.92,1.32]). Significant interactions were observed for all age * time interactions, indicating that these longitudinal patterns of use differed across age bands (see table for effects). Specifically, simple slopes analyses revealed that the odds of cigarette and alcohol co-use decreased in each wave for youth aged 15–17 (W5: OR = 0.59 [0.43,0.79]; W6: 15–17: OR = 0.24 [0.14,0.42]), adults aged 18–24 (W5: OR = 0.70 [0.64,0.76]; W6: OR = 0.48 [0.44,0.53]), adults aged 25–34 (W5: OR = 0.84 [0.77,0.90]; W6: OR = 0.66 [0.59,0.73]), and adults aged 35–64 (W5: OR = 0.89 [0.85,0.94]; W6: OR = 0.78 [0.74,0.83]).
E-cigarette and alcohol co-use (Fig. 1b and Supplemental Table 3).
For older adults aged 65+, co-use of e-cigarettes and alcohol was generally stable over time, although there was a marginal increase in W5 (W5: OR = 1.61 [1.00,42.61]; W6: OR = 0.98 [0.46,2.10]). All age * time interactions were non-significant, indicating that these longitudinal patterns of use were generally comparable across age bands (see table for effects). Simple slopes analyses echoed these observations, with increased odds of e-cigarette and alcohol co-use in W5 and smaller or non-significant differences in W6 for youth aged 15–17 (W5: OR = 1.53 [1.26,1.87]; W6: OR = 0.85 [0.62,1.16]) and adults aged 18–24 (W5: OR = 1.94 [1.76,2.14]; W6: OR = 1.59 [1.43,1.44]), 25–34 (W5: OR = 1.69 [1.50,1.89]; W6: OR = 1.56 [1.36,1.80]), and 35–64 (W5: OR = 1.21 [1.07,1.36]; W6: OR = 0.91 [0.80,1.05]).
Other tobacco products and alcohol co-use. (Fig. 1c and Supplemental Table 4).
For older adults aged 65+, co-use of OTP and alcohol decreased in both W5 (OR = 0.82 [0.68,1.00]) and W6 (OR = 0.67 [0.53,0.85]). Age * time interactions were negative for youth age 15–17, non-significant (W5) or negative (W6) for adults aged 18–24, non-significant for adults aged 25–34, and positive (W5) or non-significant (W6) for adults aged 35–64, indicating nuanced longitudinal patterns of OTP and alcohol co-use over time across age bands (see table for effects). Simple slopes analyses found decreasing odds of OTP and alcohol co-use over time for youth aged 15–17 (W5: OR = 0.52 [0.37,0.74]; W6: 0.22 [0.12,0.40]) and adults aged 18–24 (W5: OR = 0.73 [0.66,0.81]; W6: 0.43 [0.39,0.48]). For adults aged 25–34 and 35–64, odds were stable at W5 (25–34: OR = 0.98 [0.87,1.10]; 35–64: OR = 1.04 [0.94,1.16]) but decreased in W6 (25–34: OR = 0.84 [0.73,0.96]; 35–64: OR = 0.78 [0.70,0.88]).
3.2.Co-use with cannabis
Cigarette and cannabis co-use. (Fig. 1d and Supplemental Table 5).
For older adults aged 65+, co-use of cigarettes and cannabis was stable but marginally increasing in W5 (OR = 1.39 [0.99,1.95]) and increased in W6 (OR = 2.03 [1.41,2.92]). Age * time interactions were significant and negative for all Waves and age bands except adults age 25–34 and 35–64 at W5, which were non-significant (see table for effects). Simple slopes analyses found decreasing odds of cigarette and cannabis co-use over time for youth aged 15–17 (W5: OR = 0.64 [0.51,0.80]; W6: OR = 0.39 [0.24,0.63]) and adults aged 18–24 (W5: OR = 0.68 [0.61,0.76]; W6: OR = 0.52 [0.45,0.59]), stable odds for adults aged 25–34 (W5: OR = 1.00 [0.92,1.10]; W6: 1.07 [0.94,1.21]), and stable but marginal increased odds for adults aged 35–64 at W5 (OR = 1.08 [1.00,51.17]) and increased odds at W6 (OR = 1.23 [1.13,1.34]).
E-cigarette and cannabis co-use. (Fig. 1e and Supplemental Table 6).
For older adults aged 65+, co-use of e-cigarettes and cannabis was stable over time (W5: OR = 1.67 [0.86,3.27]; W6: OR = 1.09 [0.39,3.02]). All age * time interactions were non-significant, indicating that these longitudinal patterns of use were generally comparable across age bands (see table for effects). Notably, simple slopes analyses showed increased odds of e-cigarette and cannabis co-use relative to W4 for all age bands at both W5 (15–17: OR = 1.92 [1.52,2.45]; 18–24: OR = 1.70 [1.53,1.88]; 25–34: OR = 1.74 [1.49,2.03]; 35–64: OR = 1.41 [1.20,1.65]) and W6 (15–17: OR = 1.55 [1.16,2.07]; 18–24: OR = 1.44 [1.29,1.61]; 25–34: OR = 1.98 [1.69,2.33]; 35–64: OR = 1.31 [1.08,1.58]). However, the non-significant age * time interactions indicated that these effects were not significantly different from the stable rates observed for adults age 65 + .
OTP and cannabis co-use. (Fig. 1f and Supplemental Table 7).
For older adults aged 65+, co-use of OTP and cannabis was stable over time, but marginally decreased in W5 (W5: OR = 0.64 [0.40,1.04]; W6: OR = 0.92 [0.55,1.55]). Age * time interactions were non-significant at W5 except for adults aged 35–64 which was positive, significant and negative for youth aged 15–17 and adults aged 18–24 at W6, and non-significant for adults aged 25–34 and 35–64 at W6, indicating nuanced longitudinal patterns of OTP and cannabis co-use over time across age bands (see table for effects). Simple slopes analysis found decreasing odds of OTP and cannabis co-use for youth aged 15–17 (W5: OR = 0.59 [0.41,0.84]; W6: OR = 0.34 [0.20,0.55]) and adults aged 18–24 (W5: OR = 0.71 [0.64,0.78]; W6: OR = 0.45 [0.39,0.52]), stable (W5: OR = 1.04 [0.91,1.19]) or increasing (W6: OR = 1.24 [1.06,1.45]) odds for adults aged 25–34, and stable odds for adults aged 35–64 (W5: OR = 1.08 [0.94,1.24]; W6: OR = 1.05 [0.89,1.23]).
4.Discussion
This study characterized longitudinal trends in past 30-day NTP co-use with alcohol and cannabis across youth and adulthood. Perhaps not surprisingly, co-use rates were consistently highest among young adults aged 18–24 and 25–34 across almost all past 30-day co-use patterns and waves, followed by adults aged 35–64. Two notable exceptions were the co-use of e-cigarettes with cannabis and alcohol, where rates were higher in adolescents aged 15–17 than adults aged 35–64 across most waves. These prevalence rates likely reflect the acute increase of e-cigarette use in youth and young adults following the introduction of protonated nicotine (Cohn et al., 2019; Vallone et al., 2020; Cobb et al., 2018). Also notable were the generally low and stable, although occasionally increasing (i.e., cigarette and cannabis co-use) rates of co-use among adults aged 65+. Changes in cigarette and cannabis co-use rates in older adults are concerning, especially in light of older adults being less interested in quitting smoking and less likely to attempt quitting (Centers for Disease Control and Prevention. Quitting smoking among adults—United States, 2011) despite being more likely to receive quitting advice (Babb et al., 2017). In light of these trends, interventions targeting older adults should be flexible and responsive to clinical resistance and attentive to the emerging co-use of substances in this age groups, especially given the health benefits of smoking cessation at any age (Gellert et al., 2012).
In our study, 18–24 and 25–34 year olds demonstrated markedly higher levels of co-use relative to other ages, especially co-use of e-cigarettes and OTP with alcohol and cannabis. While the dominant focus in the academic literature has been on the potential use of reduced-harm NTP to aid smoking cessation, (Wang et al., 2020) interventions testing effective ways to facilitate cessation of e-cigarettes and OTPs are also critically needed. The co-use rates observed in this study suggest that young adults may benefit from interventions tailored to the specific needs and delivery modalities most effective for this age group, particularly in areas where gold-standard treatment approaches have yet to be established.
Additionally, although the use patterns of 15–17 year olds tended to deviate significantly from young adults, they also evidenced high rates of e-cigarette co-use with both alcohol and cannabis. Adolescents may require separate, developmentally appropriate interventions with demonstrated efficacy in this population, such as family-based treatment, and innovative strategies such as digital intervention (Fadus et al., 2019). These observations highlight the critical need to identify subpopulations with high, stable, or increasing levels of substance co-use whose physical and mental health consequences may otherwise be overlooked. Interventions targeting these populations must address co-use as a unique phenomenon that is distinct from the independent use of single-substances.
Although the current study focused on characterizing trends in substance co-use across age bands and evaluating differences in trends across those bands, there is also a need to understand the drivers of these differences. For instance, given the known cohort and age-based changes in harm perceptions for substances, (Waddell, 2022; Cheeta et al., 2018) there is a need for research to explore the same harm perceptions for co-use patterns and whether such changes translate to population-level shifts in patterns of co-use. Similarly, advertising campaigns by the tobacco, alcohol, and cannabis industry that have demonstrated significant impact on patterns of substance use particularly in attracting naive youth non-users, (Stroup and Branstetter, 2018) may also have a halo effect on perceived acceptability or harm of co-use.
4.1.Strengths and limitations
Strengths of the PATH study and the current analysis include a large, nationally representative sample, enabling the detection of small population-level effects, as well as the longitudinal design, which permits examination of evolving trends over time. Additionally, the inclusion of individuals across the lifespan enabled exploration of past 30-day co-use patterns across developmental stages. However, the current study is not without limitations. First, while the PATH study is representative of the CNP, its resampling strategy limits longitudinal cross-sectional analyses. Yet, near-approximations of cross-sectional estimates are possible by using wave-specific weights, accounting for covariance in observations over time, and accounting for waves analytically, as was done in the present analyses. Second, the interview nature of PATH data collection could result in under-reporting of substance use, particularly for youth. Third, the amount of work required to collect and prepare the data for release results in a lag between collection and analysis. Analyses may thus not fully capture emerging trends, such as the increased use of oral nicotine products (Kramer et al., 2023). Estimates for OTP co-use may thus underestimate actual co-use rates, which should be examined in future waves. Fourth, W4 and W5 of PATH versus W6 were collected before and after the start of the COVID-19 pandemic, respectively. Changes between W5 and W6 may thus reflect both natural shifts in behavior as well as behavior changes associated with the (at the time) emerging health threat. Last, we were unable to assess the frequency and quantity of co-use, limiting our ability to evaluate the intensity of such patterns. Thus, it is possible that some use reflects experimental rather than established patterns of co-use. This restriction prevents a deeper understanding of how often co-use occurs, whether co-use occurs simultaneously or sequentially, and the amounts consumed. Future investigations should examine these behavioral patterns to characterize high-risk behaviors and tailor interventions to address specific patterns of use.
5.Conclusions
While promising decreases in the co-use of NTP with alcohol and cannabis were observed, these declines were not observed in all age groups or for all co-use patterns, especially older adults aged 65+ whose co-use patterns were generally stable or increasing (i.e., co-use of cigarettes and cannabis). In addition, some age groups demonstrated increases in co-use of e-cigarettes with alcohol and cannabis, albeit with signs that these trends may be slowing or reversing in most - but not all (i.e., adults ages 25–34) - age groups. Rates of co-use of OTP with alcohol and cannabis varied considerably by age, evidencing increases, decreases, and stable rates for individuals at different developmental stages. These findings serve as a strong baseline for comparison of future trends, highlight age groups and patterns of co-use of special concern, demonstrate the complex interplay between substance use patterns and developmental stages, and reflect the dynamic nature of co-use behaviors in ever-evolving tobacco and cannabis marketplaces.
Supplementary Material
Funding
This work was supported by NIH grants K08DA048137 (PI: Soko-lovsky); K01DA048135 (PI: Micalizzi); K23DA045078 (PI: Murphy); and P20GM130414 (PI: Monti).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.org/10.1016/j.addbeh.2026.108633.
Footnotes
Footnote Group
Data availability
Data are publicly available
References
Untitled section
References
- Meier E, & Hatsukami DK (2016). A review of the additive health risk of cannabis and tobacco co-use. Drug and Alcohol Dependence, 166, 6–12.
- Correa JB, Myers MG, Tully LK, & Doran N. (2020). Co-occurring use of cannabis and tobacco and the presence of acute respiratory symptoms among young adult light and intermittent smokers. Substance Use & Misuse, 55(13), 2129–2137.
- Winhusen T, Theobald J, Kaelber DC, & Lewis D. (2020). The association between regular cannabis use, with and without tobacco co-use, and adverse cardiovascular outcomes: Cannabis may have a greater impact in non-tobacco smokers. The American Journal of Drug and Alcohol Abuse, 46(4), 454–461.
- Moons P, Luyckx K, Kovacs AH, et al. (2019). Prevalence and effects of cigarette smoking, cannabis consumption, and co-use in adults from 15 countries with congenital heart disease. The Canadian Journal of Cardiology, 35(12), 1842–1850.
- Cohn AM, Johnson AL, Rose SW, Pearson JL, Villanti AC, & Stanton C. (2018). Population-level patterns and mental health and substance use correlates of alcohol, marijuana, and tobacco use and co-use in US young adults and adults: Results from the population assessment for tobacco and health. The American Journal on Addictions, 27(6), 491–500.
- Rubenstein D, McClernon FJ, & Pacek LR (2024). Trends in cannabis and tobacco co-use in the United States, 2002-2021. Addictive Behaviors, 158(108129), Article 108129.
- Wang JB, Olgin JE, Nah G, et al. (2018). Cigarette and e-cigarette dual use and risk of cardiopulmonary symptoms in the Health eHeart Study. PLoS One1, 13(7), Article e0198681.
- Berg CJ, Krishnan N, Graham AL, & Abroms LC (2021). A synthesis of the literature to inform vaping cessation interventions for young adults. Addictive Behaviors, 119(106898), Article 106898.
- Reboussin BA, Wagoner KG, Ross JC, Suerken CK, & Sutfin EL (2021). Tobacco and marijuana co-use in a cohort of young adults: Patterns, correlates and reasons for co-use. Drug and Alcohol Dependence, 227(109000), Article 109000.
- Tucker JS, Pedersen ER, Seelam R, Dunbar MS, Shih RA, & D’Amico EJ (2019). Types of cannabis and tobacco/nicotine co-use and associated outcomes in young adulthood. Psychology of Addictive Behaviors, 33(4), 401–411.
- Wang JB, Ramo DE, Lisha NE, & Cataldo JK (2016). Medical marijuana legalization and cigarette and marijuana co-use in adolescents and adults. Drug and Alcohol Dependence, 166, 32–38.
- de Dios MA, Vaughan EL, Stanton CA, & Niaura R. (2009). Adolescent tobacco use and substance abuse treatment outcomes. Journal of Substance Abuse Treatment, 37(1), 17–24.
- Hefner KR, Sollazzo A, Mullaney S, Coker KL, & Sofuoglu M. (2019). E-cigarettes, alcohol use, and mental health: Use and perceptions of e-cigarettes among college students, by alcohol use and mental health status. Addictive Behaviors, 91, 12–20.
- McKee SA, Falba T, O’Malley SS, Sindelar J, & O’Connor PG (2007). Smoking status as a clinical indicator for alcohol misuse in US adults. Archives of Internal Medicine, 167(7), 716–721.
- Verplaetse TL, & McKee SA (2017). An overview of alcohol and tobacco/nicotine interactions in the human laboratory. The American Journal of Drug and Alcohol Abuse, 43(2), 186–196.
- Anantharaman D, Marron M, Lagiou P, et al. (2011). Population attributable risk of tobacco and alcohol for upper aerodigestive tract cancer. Oral Oncology, 47(8), 725–731.
- Cooper M, Creamer MR, Ly C, Crook B, Harrell MB, & Perry CL (2016). Social norms, perceptions and dual/poly tobacco use among Texas youth. American Journal of Health Behavior, 40(6), 761–770.
- Agaku IT, Perks SN, Odani S, & Glover-Kudon R. (2020). Associations between public e-cigarette use and tobacco-related social norms among youth. Tobacco Control, 29(3), 332–340.
- Giovenco DP, Spillane TE, & Merizier JM (2019). Neighborhood differences in alternative tobacco product availability and advertising in New York City: Implications for health disparities. Nicotine & Tobacco Research, 21(7), 896–902.
- Ribisl KM, D’Angelo H, Feld AL, et al. (2017). Disparities in tobacco marketing and product availability at the point of sale: Results of a national study. Preventive Medicine, 105, 381–388.
- Pedersen ER, Tucker JS, Davis JP, et al. (2021). Tobacco/nicotine and marijuana co-use motives in young adults: Associations with substance use behaviors one year later. Psychology of Addictive Behaviors, 35(2), 133–147.
- Lajtha A, & Sershen H. (2010). Nicotine: Alcohol reward interactions. Neurochemical Research, 35(8), 1248–1258.
- Choi K, Grana R, & Bernat D. (2017). Electronic nicotine delivery systems and acceptability of adult cigarette smoking among Florida youth: Renormalization of smoking? The Journal of Adolescent Health, 60(5), 592–598.
- Smith DM, Kozlowski L, O’Connor RJ, Hyland A, & Collins RL (2021). Reasons for individual and concurrent use of vaped nicotine and cannabis: Their similarities, differences, and association with product use. J Cannabis Res., 3(1), 39.
- Leung J, Gravely S, Lim C, Hall W, & Chan G. (2022). Age-period-cohort analysis of trends in tobacco smoking, cannabis use, and their co-use in the Australian population. Addiction, 117(10), 2730–2735.
- Daw J, Nowotny KM, & Boardman JD (2013). Changing patterns of tobacco and alcohol co-use by gender in the United States, 1976-2010. Demographic Research, 28, 637–648.
- Hyland A, Ambrose BK, Conway KP, et al. (2017). Design and methods of the population assessment of tobacco and health (PATH) study. Tobacco Control, 26(4), 371–378.
- Cohn AM, Abudayyeh H, Perreras L, & Peters EN (2019). Patterns and correlates of the co-use of marijuana with any tobacco and individual tobacco products in young adults from Wave 2 of the PATH Study. Addictive Behaviors, 92, 122–127.
- McCarthy PJ. Pseudoreplication: further evaluation and applications of the balanced half-sample technique. Vital Health Stat 2. 1969;(31):1–24.
- Judkins DRAD (1990). Fay’s method for variance estimation. Journal of Official Statistics., 6(3), 223.
- R Core Team. R: A Language and Environment for Statistical Computing. Published online 2017. https://www.R-project.org/.
- Lumley T. (2004). Analysis of complex survey samples. Journal of Statistical Software., 9 (1), 1–19.
- Aiken LS, West SG, & Reno RR (1991). Multiple regression: Testing and interpreting interactions. Sage.
- USDHHS. (2024). Population assessment of tobacco and health (PATH) study [United States] public-use files users guide. National Institute on Drug Abuse.
- Vallone DM, Cuccia AF, Briggs J, Xiao H, Schillo BA, & Hair EC (2020). Electronic cigarette and JUUL use among adolescents and young adults. JAMA pediatrics., 174(3), 277–286.
- Cobb CO, Soule EK, Rudy AK, Sutter ME, & Cohn AM (2018). Patterns and correlates of tobacco and cannabis co-use by tobacco product type: Findings from the Virginia Youth Survey. Substance Use & Misuse, 53(14), 2310–2319.
- Centers for Disease Control and Prevention. Quitting smoking among adults—United States, 2001–2010. MMWR Morb Mortal Wkly Rep. 2011;60(44).
- Babb SD, Malarcher A, Schauer GL, Asman KJ, & Jamal A. (2017). Quitting smoking among adults - United States, 2000-2015. MMWR Morbidity and mortality weekly report., 65(52), 1457–1464.
- Gellert C, Schöttker B, & Brenner H. (2012). Smoking and all-cause mortality in older people: Systematic review and meta-analysis: Systematic review and meta-analysis. Archives of Internal Medicine, 172(11), 837–844.
- Wang RJ, Bhadriraju S, & Glantz SA (2020). E-cigarette use and adult cigarette smoking cessation: A meta-analysis. American Journal of Public Health, 111(2), 230–246.
- Fadus MC, Squeglia LM, Valadez EA, Tomko RL, Bryant BE, & Gray KM (2019). Adolescent substance use disorder treatment: An update on evidence-based strategies. Current Psychiatry Reports, 21(10), 96.
- Waddell JT (2022). Age-varying time trends in cannabis- and alcohol-related risk perceptions 2002-2019. Addictive Behaviors, 124(107091), Article 107091.
- Cheeta S, Halil A, Kenny M, et al. (2018). Does perception of drug-related harm change with age? A cross-sectional online survey of young and older people. BMJ Open, 8(11), Article e021109.
- Stroup AM, & Branstetter SA (2018). Effect of e-cigarette advertisement exposure on intention to use e-cigarettes in adolescents. Addictive Behaviors, 82, 1–6.
- Kramer RD, Park-Lee E, Marynak KL, Jones JT, Sawdey MD, & Cullen KA (2023). Nicotine pouch awareness and use among youth, national youth tobacco survey, 2021. Nicotine & Tobacco Research, 25(9), 1610–1613.
Associated Data
Supplementary Materials
Data Availability Statement
Data are publicly available