Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jan 9.
Published in final edited form as: Drug Alcohol Depend. 2018 Mar 10;190:235–241. doi: 10.1016/j.drugalcdep.2018.01.020

Marijuana use predicts onset of current little cigar use in a national sample of US young adults

Amy M Cohn a,b,*, Amanda L Johnson c,1, Craig S Fryer d, Andrea C Villanti e,f
PMCID: PMC12780974  NIHMSID: NIHMS2132792  PMID: 30064060

Abstract

Background:

This study examined whether young adult marijuana use increases risk of subsequent large cigar (LC) and little cigar/cigarillo (LCC) use among naïve users.

Methods:

Data were from 8 waves of the Truth Initiative Young Adult Cohort, a national sample of US young adults aged 18–34 assessed every 6 months. Discrete-time survival analyses examined whether baseline ever marijuana use among never cigar users predicted onset of past 30-day LC and LCC use and whether baseline ever LC and LCC use among never marijuana users predicted onset of past 30-day marijuana use. Models adjusted for demographics, past 30-day alcohol use, past 30-day tobacco product use, and menthol tobacco use.

Results:

In adjusted models, baseline ever marijuana use predicted onset of past 30-day LCC but not LC use. Cumulative risk ratios showed that 23% of ever marijuana users at baseline reported past 30-day use of LCCs by the end of wave 8 compared to just 3% of baseline never marijuana users. Race and past 30-day use of specific tobacco products also predicted onset of past 30-day LC and LCC use. Past 30-day use of alcohol uniquely predicted onset of past 30-day LCC use but not LC use. Baseline ever LC and LCC use did not predict onset of past 30-day marijuana use in models that adjusted for demographics, alcohol, and specific tobacco product use.

Conclusions:

Ever marijuana use among US young adults may be an important predictor of onset of regular LCC use. Findings suggest different pathways linking marijuana to different cigar sub-types.

Keywords: Young adults, Cigars, Little cigars, Marijuana, Tobacco products, Alcohol, Survival analysis

1. Introduction

While cigarette smoking has declined in the US (Jamal et al., 2015; Lawrence and Williams, 2016), cigar use has become more prevalent, particularly among youth and young adults (defined as ages 18–24; Cullen et al., 2011; Fix et al., 2014; Messer et al., 2014). Nearly 40% of young adults have ever used a cigar in their lifetime, and the dual use of large cigars (LCs) and little cigars/cigarillos (LCCs) in this age group is more prevalent than use of either product alone (Richardson et al., 2013). While cigar use among US young adults showed an approximately 10% increase from 2011 to 2012 (Richardson et al., 2014), rates of use dropped slightly from 2014 to 2015 in this age group according to recent data from the National Survey of Drug Use Health (Center for Behavioral Health Statistics and Quality, 2016). Recent population-based estimates from the first wave of data from the Population Assessment of Tobacco and Health (PATH) study indicate that 15% of US young adults reported past 30-day use of cigars (inclusive of both LCs and LCCs), a rate that is nearly 2 times higher than that of youth and 3 times higher relative to adults 25 years of age or older (Kasza et al., 2017). Reasons for cigar use among young adults could be due to a variety of reasons, including lower cost relative to cigarettes (Gammon et al., 2016; Pickworth et al., 2016), inclusion of a variety of flavors in cigars and colorful packaging (Glasser et al., 2017; Kostygina et al., 2014), targeted marketing (Biener and Albers, 2004; Ganz et al., 2016), lower perceptions of harm (Bernat et al., 2017), and until recently, no regulatory actions. Cigar smoking can result in similar levels of toxicant exposure as cigarettes (Caruso et al., 2015; Koszowski et al., 2015; Koszowski et al., 2017; Pickworth et al., 2016; Rosenberry et al., 2016) and many of the same negative health outcomes as cigarette smoking, including cancer risk, coronary heart disease, and COPD (Boffetta et al., 1999; Ghosh et al., 2016; Iribarren et al., 1999; Malhotra et al., 2017; Nonnemaker et al., 2014; Stallones, 2015), making their use in young adults a significant public health concern.

As of 2015, nearly 20% of US young adults report marijuana use in the past month, and, while rates of cigar use are slowly declining in this age group (Center for Behavioral Health Statistics and Quality, 2016), the co-use of marijuana with cigars is becoming increasingly popular (Cohn et al., 2016; Schauer et al., 2016; Schauer et al., 2017). In fact, recent US national data suggest that marijuana and cigar co-use in the form of blunt smoking (e.g., removing some, or all the tobacco from the cigar and replacing it with marijuana) is more popular than exclusive cigar use among young adults (Cohn et al., 2016; Schauer et al., 2017). Correlates of marijuana and cigar co-use include being male, African-American, young adult, using flavored tobacco products, and using cigarettes, alcohol, or illicit drugs (Cohn et al., 2016; Montgomery and Oluwoye, 2016; Schauer et al., 2016; Schauer et al., 2017). Of concern are recent data showing that blunt smoking does expose users to nicotine, as cigar wrappers contain between 1.2–6 mg of nicotine content per cigar (Peters et al., 2016), thus increasing their abuse liability. Our recent non-longitudinal cross-sectional analysis of the association between marijuana use and emerging tobacco product use in a national sample of young adults showed that current marijuana use (some days/everyday use) was most strongly associated with LCC use relative to the use of cigarettes and other emerging tobacco products, like hookah and e-cigarettes. Further, our results showed that marijuana use was more strongly correlated with LCC use than current alcohol use (Cohn et al., 2015). However, this study did not examine the longitudinal and causal sequencing of cigar versus marijuana use to determine whether use of one product increases risk for use of the other at a later time point.

As of 2017, 28 US states and the District of Columbia have either legalized, decriminalized, or medicalized the use of marijuana; these changes are coupled with increases in marijuana and cigar co-use, particularly among young adults (Cohn et al., 2016; Peters et al., 2012; Schauer et al., 2015; Schauer et al., 2017). As such, an important public health question is whether marijuana use is one factor driving increases in cigar use among young adults. This study used eight waves of data collected from a national sample of young adults aged 18–34 to examine two primary aims. Aim 1 examined whether and to what degree ever marijuana use at baseline among never cigar users (at baseline) predicts later onset of past 30-day LC and LCC use, controlling for factors associated with cigar and marijuana use and co-use. Aim 2 examined whether and to what degree ever LC or LCC use at baseline among never marijuana users (at baseline) predicts later onset of past 30-day marijuana use. Our primary interest in past 30-day use as the outcome of interest was to observe the impact of marijuana and/or cigar use on subsequent regular use behavior, rather than initiation, which could indicate ever use but not consistent use.

2. Methods

2.1. Respondents

Data were drawn from Waves 1 (July 2011) through 8 (July 2015) of the Truth Initiative Young Adult Cohort, a national probability-based sample of young adults aged 18–34 assessed every 6 months collected through GfK’s KnowledgePanel®. The panel was recruited via address-based sampling, a sampling method that provides statistically valid representation of the U.S. population including cell phone-only households. Households without internet access were provided a free netbook computer and internet service to reduce response bias. African-American and Hispanic individuals were oversampled. Prior published work using this panel demonstrates its representativeness to the US population of individuals in this age group (Rath et al., 2012).

After Wave 1, the sample for each wave consisted of people who completed the previous wave plus a refreshed sample to replace individuals lost to follow-up and to ensure the sample did not age-out of the cohort. The full methodology has been previously reported (Rath et al., 2012). The study was approved by the Independent Investigational Review Board, Inc. for Waves 1–3 and Chesapeake Institutional Review Board, Inc. for Waves 4–8.

The present analysis focused on three different subpopulations. For Aim 1, the analyses focused on a subset of young adults aged 18–34 who were never users of either LC or LCC at Wave 1 (baseline), had non-missing ever marijuana use status at baseline, and reported information on either past 30-day LC use (n = 2096; n = 9585 observations) or past 30-day LCC use (n = 2217; n = 10,125 observations) in a subsequent wave. For Aim 2, the analysis focused on a subset of young adults aged 18–34 who were never users of marijuana at baseline, had non-missing ever LC or LCC use status at baseline, and reported information on past 30-day marijuana use in a subsequent wave (Aim 2; n = 2756; n = 12,316 observations). In addition, respondents included in any analysis were required to have no missing baseline sociodemographic covariates. Respondents who reported past 30-day LC use, LCC use, or marijuana use at Wave 1 were excluded from the subpopulation for the respective analysis, as they were considered no longer at risk.

2.2. Measures

2.2.1. Demographics

At baseline, respondents provided information on age (grouped as 18–24 and 25–34), gender, race/ethnicity (White, non-Hispanic; Black, non-Hispanic; Other, non-Hispanic; and Hispanic), and financial situation (does not meet basic needs, just meets basic needs, meets needs with a little left, and lives comfortably).

2.2.2. Cigar use, cigarette use, and menthol tobacco product use

At baseline, respondents were categorized as ever users of LC or LCCs if they reported ever smoking the product, even one puff, at Wave 1. At each wave, respondents were asked about past 30-day use of ten different tobacco products (cigarettes, LCs, pipe tobacco, LCCs, e-cigarettes, chew, dip/snuff, snus, dissolvable tobacco, and hookah). A respondent who reported use of a specific tobacco product in the past month was defined as a past 30-day user in that particular wave. For each product used, respondents were also asked to identify the typical brand used in the past 30 days and whether it was menthol-flavored. A respondent who reported that any of their typical brands were mentholated was defined as a current menthol tobacco user in that particular wave. Examples of usual brands were included in the question to further assist with the query (Terchek et al., 2009; Trapl et al., 2011). Past 30-day non-menthol tobacco use and past 30-day menthol tobacco use were wave-specific variables. Thus, over time, an individual could be both a past 30-day non-menthol tobacco user or a past 30-day menthol tobacco user, depending on the wave.

2.2.3. Alcohol and marijuana use

Current alcohol use was assessed at each wave. In waves 1–6, all respondents were asked about the frequency of their current use, with response options “every day”, “some days”, and “not at all”. Those who reported using alcohol “some days” or “every day” were defined as current users in that particular wave. For waves 7 and 8, current alcohol use was determined by two items. The first item asked about frequency of drinking alcohol in the past year (“never”, “monthly or less”, “2–4 times per month”, “2–3 times per week”, or “4 or more times per week”). Those who reported any use of alcohol were then queried about the frequency of use in the past 30 days, with respondents using ≥1 days in the past month defined as current users. Those with missing data (< 2% of all observations) were defined as non-current users in that particular wave.

At baseline, respondents were categorized as never marijuana users if they reported using marijuana “Not at all” or ever marijuana users if they reported using marijuana “Some days” or “Every day” at Wave 1. At each subsequent wave, past 30-day marijuana use was defined as using marijuana on at least one day in the past 30 days.2 Those with missing data (< 2% of all observations) or who reported “not at all” or “0 days” were defined as non-current users in that particular wave.

2.3. Analyses

Analyses were performed using Stata/SE 14.2 using the svy prefix command to account for sampling weights in survey data. Taylor Linearization was used to estimate standard errors. Post-stratification weights were used to offset non-response or non-coverage bias to produce nationally representative estimates specific to each wave. We first examined the baseline characteristics of never users of LCs and LCCs by baseline never and ever marijuana use. Next, separate discrete time survival analyses were conducted to examine the impact of baseline ever (vs. never) marijuana use on the outcomes of time-to-first past 30-day use of LCs or time-to-first past 30-day use of LCCs (Aim 1) and the impact of baseline ever (vs. never) LC or LCC use on the outcome of time-to-first past 30-day use of marijuana (Aim 2), controlling for study wave, demographics, and past 30-day alcohol, past 30-day tobacco product-specific use, and past 30-day menthol-flavored tobacco use.

Results were reported as incident rate ratios (IRRs): the incidence rate of the past 30-day outcome in those with the exposure during the time at risk divided by the incidence rate of the past 30-day outcome in those without the exposure during the time at risk. The discrete time proportional hazard model assumes constant continuous-time hazard within each time-period. The glm command in Stata with the link function = cloglog and family = binomial was used to estimate IRRs. Cumulative risk ratios were also calculated to estimate the risk of past 30-day LC, LCC, or marijuana use onset by baseline ever (vs. never) marijuana, LC, or LCC use over the study period. Similar analyses were conducted to present the risk of past 30-day onset of marijuana use among baseline LC or LCC users.

The time origin for the analyses was Wave 1, with Wave 2 being the earliest time point respondents could achieve the outcome. Time “at risk” was defined as waves following Wave 1 in which the respondent completed the survey but did not report the past 30-day outcome. Respondents could have gaps in waves of study participation and remain in the analytic sample. Person-periods in which the respondent was not known to be at risk (i.e., waves of non-response, person-periods in which the exposure and/or outcome of interest were not reported) or in which the respondent was no longer at risk (i.e., person-periods following the first report of past 30-day use) were removed from the analysis.

Baseline demographics and ever marijuana use (or cigar use) were time-fixed factors, and past 30-day use covariates were time-varying factors.

3. Results

3.1. Correlates of baseline marijuana use among never users of LCs and LCCs

Table 1 shows characteristics of baseline never LC users (n = 2096) and never LCC users (n = 2217) by baseline never and ever marijuana use. Across both never LC and never LCC users, baseline marijuana users were primarily younger (18–24 vs. 25–34), White, female, and met basic needs financially or lived comfortably. Three quarters of ever marijuana users reported past 30-day alcohol use, one third reported past 30-day cigarette use, 80% reported past 30-day marijuana use, and 16% to 19% reported past 30-day menthol tobacco use.

Table 1.

Baseline characteristics of never use of large cigars (LCs) and little cigars/cigarillos (LCCs) by baseline marijuana use.

Never users of LCs Never users of LCCs
Never marijuana user (n = 1958) Ever marijuana user (n = 138) Total (n = 2096) Never marijuana user (n = 2088) Ever marijuana user (n = 128) Total (n = 2217)
Age
 18–24 42.2% 57.3% 43.2% 39.5% 53.1% 40.3%
 25–34 57.8% 42.8% 56.8% 60.5% 46.9% 59.7%
Sex
 Male 33.3% 39.9% 33.7% 38.8% 47.7% 39.2%
 Female 66.8% 60.1% 66.3% 61.3% 52.3% 60.8%
Race
 White 52.7% 50.0% 52.5% 54.6% 61.7% 55.0%
 Black 14.9% 20.3% 15.2% 12.3% 14.1% 12.4%
 Other 5.5% 7.3% 5.6% 5.3% 6.3% 5.3%
 Hispanic 27.0% 22.5% 26.7% 27.9% 18.0% 27.3%
Financial Situation*
 Does not meet basic expenses 8.0% 13.8% 8.4% 7.6% 11.7% 7.8%
 Just meets basic expense 29.3% 36.2% 29.7% 27.7% 33.6% 28.0%
 Meets needs with a little left 37.7% 38.4% 37.8% 39.1% 41.4% 39.2%
 Lives comfortably 24.9% 11.6% 24.1% 25.6% 13.3% 24.9%
Past 30-day use
 Alcohol 38.4% 74.6% 40.8% 39.8% 78.1% 42.0%
 Marijuana – 82.6% 5.4% – 80.5% 4.7%
 Cigarettes 8.2% 33.1% 9.8% 8.3% 35.4% 9.8%
 Large cigars – – – 1.3% 3.1% 1.4%
 Pipe tobacco 0.0% 0.0% 0.0% 0.1% 0.8% 0.1%
 Little cigars/cigarillos 0.4% 11.6% 1.1% – – –
 E-cigarettes 0.3% 2.2% 0.4% 0.3% 0.8% 0.3%
 Chew 0.2% 0.0% 0.2% 0.3% 0.0% 0.3%
 Smokeless tobacco 0.3% 0.0% 0.3% 0.5% 0.8% 0.5%
 Snus 0.2% 0.0% 0.1% 0.1% 0.8% 0.2%
 Dissolvable tobacco 0.0% 0.0% 0.0% 0.1% 0.0% 0.0%
 Hookah 0.2% 2.2% 0.3% 0.2% 3.1% 0.3%
 Past 30-day menthol tobacco use 4.1% 18.8% 5.1% 3.8% 15.6% 4.5%
Open in a new tab

Note:

*

Missingness: n = 1 among never users of LCs; n = 1 among never users of LC

Results of unadjusted and adjusted discrete time survival analyses of baseline ever marijuana use (vs. never use) predicting time-to-first past 30-day use of LCs and LCCs are shown in Table 2.

Table 2.

Discrete time survival analyses of baseline ever marijuana use (vs never use) predicting time-to-first past 30-day use of large cigars (LCs) and little cigars/cigarillos (LCCs).

Time to first past 30-day use of LCs Time to first past 30-day use of LCCs
Unadjusted Adjusted Unadjusted Adjusted
IRR 95% CI aIRR 95% CI IRR 95% CI aIRR 95% CI
Baseline ever marijuana use 2.35 (1.01, 5.04) 0.29 (0.06, 1.46) 5.22 (2.51, 10.85) 2.79 (1.12, 6.95)
Age
 18–24 – – – –
 25–34 0.81 (0.37, 1.73) 0.52 (0.22, 1.24)
Sex
 Male – – – –
 Female 0.68 (0.33, 1.39) 1.14 (0.49, 2.65)
Race
 White (ref) – – – –
 Black 0.44 (0.14, 1.43) 4.50 (1.60, 12.66)
 Other 0.11 (0.03, 0.37) 0.42 (0.10, 1.73)
 Hispanic 1.79 (0.87, 3.67) 0.90 (0.38, 2.11)
Financial Situation
 Does not meet basic expenses 0.56 (0.12, 2.57) 1.72 (0.57, 5.18)
 Just meets basic expenses 1.03 (0.40, 2.66) 1.02 (0.43, 2.41)
 Meets needs with a little left – – – –
 Lives comfortably 1.31 (0.45, 3.85) 1.34 (0.45, 3.96)
Past 30-day alcohol use 2.16 (0.98, 4.77) 3.52 (1.11, 11.20)
Past 30-day cigarette use 1.42 (0.22, 9.19) 3.48 (1.07, 11.30)
Past 30-day other cigar usea 41.53 (8.50, 203.00) 5.07 (1.20, 21.51)
Past 30-day pipe use 19.18 (6.21, 59.27) 2.06 (0.22, 19.42)
Past 30-day e-cigarette use 1.05 (0.20, 5.52) 3.53 (0.61, 20.52)
Past 30-day chewing tobacco use 5.63 (1.09, 28.99) 0.58 (0.03, 13.46)
Past 30-day smokeless tobacco use 1.60 (0.26, 9.85) 2.34 (0.62, 8.91)
Past 30-day snus use 0.30 (0.04, 2.29) 3.45 (0.61, 19.53)
Past 30-day dissolvable tobacco use 0.94 (0.08, 11.53) 6.96 (0.81, 59.70)
Past 30-day hookah use 0.84 (0.12, 6.12) 5.19 (1.21, 22.32)
Past 30-day menthol tobacco use 2.53 (0.35, 18.40) 1.64 (0.53, 6.51)
Open in a new tab

Note: Models control for Wave. Data represent hazard ratios, interpreted as incident rate ratios (IRR). Items in bold are significantly different at p < .05.

a

Other cigar use includes little cigars and cigarillos for time to first past 30-day use of LCs, and large cigars for time to first past 30-day use of LCCs.

Among baseline never LC users, the risk of reporting past 30-day LC use was significantly higher among baseline ever versus never marijuana users in the unadjusted model. This effect was no longer significant after adjusting for demographics, alcohol, past 30-day product-specific use, and menthol tobacco use. Baseline never LC users who identified as a non-Hispanic race other than Black had significantly reduced risk of reporting onset of past 30-day LC use compared to respondents who identified as White (aIRR = 0.11; CI: 0.03, 0.37). Those who reported past 30-day use of LCC (aIRR = 41.53; CI: 8.50, 203.00), pipe tobacco (aIRR = 19.18; CI: 6.21, 59.27), and chew (aIRR = 5.63; CI: 1.09, 28.99) had the greatest risks of reporting onset of past 30-day LC use.

Among baseline never LCC users, both unadjusted and adjusted models show that baseline ever marijuana users had elevated risk of reporting past 30-day LCC use compared to never marijuana users at baseline (cIRR = 5.22; CI: 2.51,10.85; aIRR = 2.79; CI: 1.12, 6.95). Further, Black vs. White race (aIRR = 4.50; CI: 1.60, 12.66), past 30-day alcohol use (aIRR = 3.52; CI: 1.11, 11.20), past 30-day cigarette use (aIRR = 3.48; CI: 1.07, 11.30), other cigar use (aIRR = 5.07; CI: 1.20, 21.51), and past 30-day hookah use (aIRR = 5.19; CI: 1.21, 22.32) were also associated with elevated risk of reporting onset of past 30-day LCC use. Age, gender, financial situation, and menthol tobacco use were unrelated to increased risk of reporting onset of past 30-day LCC use.

Fig. 1 shows the cumulative risk of past 30-day LC and LCC use by baseline ever marijuana use among baseline never LC users (Panel A) and baseline never LCC users (Panel B). By the end of Wave 8, 13% of respondents who were never users of LCs and ever users of marijuana at baseline reported past 30-day LC use compared with 4% of respondents who were never users of both LCs and marijuana at baseline. From Waves 5 through 8, there were no new reports of past 30-day LC use among ever marijuana users, as indicated by the flat line and the cumulative percent (13%) remaining stable over these time points. The cumulative risk ratios for LCC use showed large differences between ever and never marijuana users. Among baseline never users of LCCs who reported ever (vs. never) using marijuana at baseline, 23% (vs. 3%) reported onset of past 30-day LCC use by the end of Wave 8. At each wave, the figure shows an increase in the proportion of young adults reporting onset of past 30-day LCC use from Wave 1 to Wave 7, while the proportion from Wave 7 to Wave 8 remained unchanged.

Fig. 1.

Fig. 1.

Open in a new tab

(Panels A and B). Cumulative risk of past 30-day large cigar (LC) and little cigar/cigarillo (LCC) use by baseline ever marijuana use.

Note: 2012 = Waves 2–3; 2013 = Waves 4–5; 2014 = Waves 6–7; 2015 = Wave 8.

3.2. Discrete-time survival analyses: baseline LC or lcc use predicting time-to-first past 30-day marijuana use

The second aim of the study was to examine whether baseline ever LC or LCC use predicts time-to-first past 30-day use of marijuana among baseline never users of marijuana. Results from adjusted models also controlling for demographics, alcohol, and tobacco product use revealed no significant risk differential in reporting past 30-day use between baseline ever and never LC users or between baseline ever and never LCC users. The cumulative risk ratios (Fig. 2, Panel A and B) showed that the risk of reporting past 30-day marijuana use was 16% among ever LC users and 19% among ever LCC users compared to 11% for never users of both.

Fig. 2.

Fig. 2.

Open in a new tab

(Panels A and B). Cumulative risk of past 30-day marijuana use by baseline (wave 1) large cigar (LC) use or little cigar/cigarillo (LCC) use. Multivariable models showed no significant difference in risk of reporting past 30-day marijuana use between baseline never and ever LC users or between never and ever LCC users.

Note: 2012 = Waves 2–3; 2013 = Waves 4–5; 2014 = Waves 6–7; 2015 = Wave 8.

4. Discussion

Using a national sample of US young adults, this study examined the predictive associations between ever marijuana use and onset of past 30-day cigar use among never cigar users and between ever cigar use and onset of past 30-day marijuana use among never marijuana users. Among young adults, 23% of marijuana ever users reported onset of past 30-day use of LCCs, while 13% reported onset of past 30-day LC use. A small proportion of never marijuana users became past-month LC or LCC users (3%–4%). After adjustment for demographics, alcohol, tobacco product use, and menthol flavored tobacco use, ever marijuana users at baseline showed a 2.79 increased risk of reporting past 30-day use of LCCs, an incident rate that was significantly greater than never marijuana users. There was no impact of baseline ever vs. never marijuana use on onset of past 30-day LC use after adjusting for other factors. In models that tested the opposite pathway among never users of marijuana, ever use of either LCs or LCCs at baseline did not differentially increase risk of onset of past 30-day marijuana use compared to never use of these products at baseline. The non-significant link between marijuana and subsequent LC use was not surprising, given prior research showing that LC use is concentrated among older, high income, college-educated, white males (Cohn et al., 2016; Richardson et al., 2013; Richardson et al., 2012), a demographic profile that does not align with that of marijuana users or marijuana and cigar co-users, who are primarily younger, male, or Black or African American (Cohn et al., 2016; Fairman, 2015; Schauer et al., 2015; Schauer et al., 2017).

Other correlates of LC and LCC use emerged in adjusted models. Among respondents who were naïve LC users at baseline, those who reported past 30-day use of LCCs, pipe, or chewing tobacco use showed significantly elevated risk of reporting onset of past 30-day LC at a subsequent wave, while respondents of Other race (relative to White race) had significantly lower risk of onset of past 30-day LC use at a subsequent wave. Past 30-day LCC use conferred the greatest risk of onset of past 30-day LC use of any other variable in the model, including marijuana. It is important to note that confidence intervals for this variable and other tobacco use correlates were wide, likely due to small sample sizes. Factors uniquely associated with increased risk of the onset of past 30-day LCC use among LCC naïve users at baseline included Black race, past 30-day alcohol use, past 30-day cigarette use, and past 30-day LC use. Because flavored cigars are popular among young adults (Villanti et al., 2017; Villanti et al., 2013), we expected that menthol tobacco use would have been associated with elevated risk of onset of both LC and LCC use; however, this was not the case. Menthol flavored tobacco use was not a predictor of time-to-first past 30-day use of either LCs of LCCs, although it is possible that other characterizing flavor profiles, like fruit or candy, would be. We were not able to examine the impact of non-menthol flavored (i.e., chocolate, fruit flavoring, etc.) tobacco product use on the outcomes of interest, as such questions were only included only in later waves of assessment. More research is needed to examine whether and what specific flavor profiles are associated with the co-use of marijuana and cigars.

This study adds to the literature in several ways. First, this study used a national sample of young adults, a vulnerable population that has the highest rates of marijuana use and increasingly growing rates of marijuana and cigar co-use compared to any other group. Because initiation and escalation of use of these products typically begins during young adulthood (Sussman and Arnett, 2014), it is an ideal time to disseminate public health messages about tobacco and marijuana use harms to prevent escalation to future use and to deter individuals from becoming problem users later in life. Second, this study incorporates a large multi-wave longitudinal design with over 10,000 observations of data. This allowed us to isolate, with a high level of power and precision, the causal association between marijuana use and subsequent cigar use (and vice versa), controlling for confounders. Most large-scale studies to date that have examined links between marijuana and cigar use have done so using correlational analyses or with convenience samples (Cohn et al., 2016; Delnevo et al., 2011; Richardson et al., 2013; Sifaneck et al., 2006), which provide only a snapshot in time and do not allow for a better understanding of the sequencing of substance use behaviors. Other recently published studies that have taken a causal approach to teasing apart the sequencing of marijuana and cigar use behavior have focused more broadly on adults rather than young adults (Fairman and Anthony, 2017). Finally, because of the large sample size, we were able to separately examine the associations of marijuana use with LC and LCC use. Given the unique association between marijuana and subsequent LCC use, findings highlight the need to disaggregate LCs from LCCs in future research studies.

Several limitations should be addressed. First, we did not specifically measure blunt use, nor did we measure non-menthol flavored tobacco use. These items were queried in the final two waves of assessment and were not included due to limited sample size. Second, we did not examine state and local marijuana policies as a potential moderator in our analyses due to the heterogeneity of policies across regions and the rapid changes of these policies over time. Third, cell sizes were too small to examine the moderating impact of race or ethnicity on the links between marijuana and cigar use. Fourth, it is important to note that predictor variables of ever use at baseline could capture regular use or experimentation, and our outcome of past 30-day use did not include a measurement of quantity or frequency of use. It would be important to determine whether and to what degree marijuana and cigar use predict varying intensities of subsequent substance use, such as number of days used in the past 30-days. Additionally, some variables were wave-specific (past 30-day menthol use), and thus we did not assess switching from tobacco to non-use or from menthol to non-menthol use. Fifth, a limitation to survival analysis is the assumption of non-informative censoring, i.e., those who left the study did so for reasons unrelated to the study. Because our study allowed respondents to have gaps between survey participation, those who left and reported past 30-day use upon returning to the study may have had the onset of past 30-day use in prior waves. We did not tease apart movement in and out of marijuana use and cigar use. Future quantitative studies could apply multi-state modeling to predict transitions in and out of young adult substance use behavior, including movement from marijuana-only use, to cigar-only use, and to marijuana and cigar co-use. Lastly, it is worth discussing limitations that are inherent to longitudinal data collection that could impact interpretation of study findings. These include participant willingness to take part in a longitudinal study, differential sample attrition, and potential for measurement reactivity when items are taken at repeated time points. The panel has been shown to be representative of individuals in this age group in the US (Rath et al., 2012), suggesting that study attrition or willingness to participate does not significantly affect generalizability of study findings. With regard to reactivity, published work indicates that reactivity to repeated assessments is unlikely, especially when multiple behaviors are monitored (Barta et al., 2012; Hufford et al., 2002) and when surveys are taken farther apart in time from each other. Each survey included other questions, aside from marijuana and cigar use, which likely diluted the potential impact of repeated administrations on behavior change and reporting of these behaviors. Further, when reactivity does exist, per previous research, it accounts for only a small proportion of the variance in behavior (Clifford et al., 2007; Maisto et al., 2007).

5. Conclusions

Taken together, ever marijuana use in the absence of ever LCC or LC use is an important predictor of onset of past 30-day LCC use but not LC use. This suggests different pathways linking marijuana to these products. Cigar use (both LCs and LCCs) among marijuana-naïve young adults was not a predictor of future marijuana use. Interventions designed to prevent onset of past 30-day LCC use in young adults may want to target early stage marijuana use, perhaps by developing messaging strategies about harms associated with marijuana and tobacco co-use, particularly cigar smoking. It would be important to further refine targets for these interventions by determining whether trajectories of marijuana and cigar use differ or are the same between young adults and youth. If marijuana use similarly predicts onset of past 30-day LCC use in youth, public health prevention campaigns targeting both marijuana and cigar use may be effective at reducing health harms across multiple public health priority groups.

Acknowledgments

Data collection for this study was funded by Truth Initiative. Support for these analyses was provided in part by Truth Initiative (AJ), award R21DA041548 from the National Institute on Drug Abuse (AC, AV, AJ) and the Centers of Biomedical Research Excellence P20GM103644 award from the National Institute on General Medical Sciences (AV). A special thank you to Dr. Erica Peters for her thorough review of this manuscript.

Footnotes

Conflict of interest

No conflict declared.

2

In Waves 2–4 and 6–8, respondents who reported using marijuana “Some days” or “Every day” were subsequently asked, “During the last 30 days, on how many days did you use the following products?”. In Wave 5, marijuana use was only assessed by the question “How often, if ever, do you currently use [marijuana]?” with response choices “Every day”, “Some days”, and “Not at all”. Thus, in Wave 5, past 30-day marijuana use was defined as reporting using marijuana “Every day” or “Some days”.

References

  1. Barta WD, Tennen H, Litt MD, 2012. Measurement reactivity in diary research. In: Mehl MR, Conner TS (Eds.), Handbook of Research Methods for Studying Daily Life. Guilford Press, New York. [Google Scholar]
  2. Bernat JK, Ferrer RA, Margolis KA, Blake KD, 2017. US adult tobacco users’ absolute harm perceptions of traditional and alternative tobacco products information-seeking behaviors, and (mis) beliefs about chemicals in tobacco products. Addict. Behav. 71, 38–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Biener L, Albers AB, 2004. Young adults: vulnerable new targets of tobacco marketing. Am. J. Public Health 94, 326–330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Boffetta P, Nyberg F, Agudo A, Benhamou E, Jockel KH, Kreuzer M, Merletti F, Pershagen G, Pohlabein H, Simonato L, Wichman HE, Saracci R, 1999. Risk of lung cancer from exposure to environmental tobacco smoke from cigars, cigarillos and pipes. Int. J. Cancer 83, 805–806. [DOI] [PubMed] [Google Scholar]
  5. Caruso RV, O’Connor RJ, Travers MJ, Delnevo CD, Stephens WE, 2015. Design characteristics and tobacco metal concentrations in filtered cigars. Nicotine Tob. Res. 17, 1331–1336. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Clifford PR, Maisto SA, Davis CM, 2007. Alcohol treatment research assessment exposure subject reactivity effects: part I: alcohol use and related consequences. J. Stud. Alcohol Drugs 68, 519–528. [DOI] [PubMed] [Google Scholar]
  7. Cohn A, Villanti A, Richardson A, Rath JM, Williams V, Stanton C, Mermelstein R, 2015. The association between alcohol, marijuana use, and new and emerging tobacco products in a young adult population. Addict. Behav. 48, 79–88. [DOI] [PubMed] [Google Scholar]
  8. Cohn A, Johnson A, Ehlke S, Villanti AC, 2016. Characterizing substance use and mental health profiles of cigar blunt, and non-blunt marijuana users from the National Survey of Drug Use and Health. Drug Alcohol Depend. 160, 105–111. [DOI] [PubMed] [Google Scholar]
  9. Cullen J, Mowery P, Delnevo C, Allen JA, Sokol N, Byron MJ, Thornton-Bullock A, 2011. Seven-year patterns in US cigar use epidemiology among young adults aged 18–25 years: a focus on race/ethnicity and brand. Am. J. Public Health 101, 1955–1962. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Delnevo CD, Bover-Manderski MT, Hrywna M, 2011. Cigar, marijuana, and blunt use among US adolescents: are we accurately estimating the prevalence of cigar smoking among youth? Prev. Med. 52, 475. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Fairman BJ, Anthony JC, 2017. Does starting to smoke cigars trigger onset of cannabis blunt smoking? Nicotine Tob. Res 20, 355–361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fairman BJ, 2015. Cannabis problem experiences among users of the tobacco-cannabis combination known as blunts. Drug Alcohol Depend. 150, 77–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fix BV, O’Connor RJ, Vogl L, Smith D, Bansal-Travers M, Conway KP, Ambrose B, Yang L, Hyland A, 2014. Patterns and correlates of polytobacco use in the United States over a decade: NSDUH 2002–2011. Addict. Behav. 39, 768–781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Gammon DG, Loomis BR, Dench DL, King BA, Fulmer EB, Rogers T, 2016. Effect of price changes in little cigars and cigarettes on little cigar sales: USA,Q4 2011–Q4 2013. Tob. Control 25, 538–544. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ganz O, Teplitskaya L, Cantrell J, Hair EC, Vallone D, 2016. Direct-to-consumer marketing of cigar products in the United States. Nicotine Tob. Res. 18, 864–868. [DOI] [PubMed] [Google Scholar]
  16. Ghosh A, Reeber SL, Jania CM, Garrison AJ, Lee SL, Tilley SL, Doerschuck CM, Herring AH, Glish GL, Tarran R, 2016. Little cigars exert more severe effects on airway epithelia than cigarettes. You Are What You Breathe: Airway Cell Biology in COPD. American Thoracic Society. [Google Scholar]
  17. Glasser AM, Johnson AL, Rose S, Ganz O, Cantrell J, Delnevo C, Villanti AC, 2017. Correlates of cigar use by type and flavor among US young adult adults: 2011–2015. Tob. Regul. Sci. 3 (Suppl. 1), 59–71. [Google Scholar]
  18. Hufford M, Shields A, Shiffman S, Paty J, Balabanis M, 2002. Reactivity to ecological momentary assessment: an example using undergraduate problem drinkers. Psychol. Addict. Behav. 16, 205–211. [PubMed] [Google Scholar]
  19. Iribarren C, Tekawa IS, Sidney S, Friedman GD, 1999. Effect of cigar smoking on the risk of cardiovascular disease, chronic obstructive pulmonary disease, and cancer in men. N. Engl. J. Med. 340, 1773–1780. [DOI] [PubMed] [Google Scholar]
  20. Jamal A, Homa DM, O’Connor E, Babb SD, Caraballo RS, Singh T, Hu SS, King BA, 2015. Current cigarette smoking among adults—United States, 2005–2014. Morb. Mortal. Wkly. Rep. 64, 1233–1240. [DOI] [PubMed] [Google Scholar]
  21. Kasza KA, Ambrose BK, Conway KP, Borek N, Taylor K, Goniewicz ML, Cummings KM, Sharma E, Pearson JL, Green VR, Kaufman AR, Bansal-Travers M, Travers MJ, Kwan J, Tworeck C, Cheng YC, Yang L, Pharris-Ciurej N, van Bemmel DM, Backinger CL, Compton WM, Hyland AJ, 2017. Tobacco-product use by adults and youths in the United States in 2013 and 2014. N. Engl. J. Med. 376, 342–353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kostygina G, Glantz SA, Ling PM, 2014. Tobacco industry use of flavours to recruit new users of little cigars and cigarillos. Tob. Control 25, 66–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Koszowski B, Rosenberry ZR, Kanu A, Viray LC, Potts JL, Pickworth WB, 2015. Nicotine and carbon monoxide exposure from inhalation of cigarillo smoke. Pharmacol. Biochem. Behav. 139, 7–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Koszowski B, Rosenberry ZR, Yi D, Stewart S, Pickworth WB, 2017. Smoking behavior and smoke constituents from cigarillos and little cigars. Tob. Regul. Sci. 3, 31–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Lawrence D, Williams JM, 2016. Trends in smoking rates by level of psychological distress—time series analysis of US National Health Interview Survey data 1997–2014. Nicotine Tob. Res. 18, 1463–1470. [DOI] [PubMed] [Google Scholar]
  26. Maisto SA, Clifford PR, Davis CM, 2007. Alcohol treatment research assessment exposure subject reactivity effects: Part II: treatment engagement and involvement. J. Stud. Alcohol Drugs 68, 529–533. [DOI] [PubMed] [Google Scholar]
  27. Malhotra J, Borron C, Freedman ND, Abnet CC, van den Brandt PA, White E, Milne RL, Giles GG, Boffetta P, 2017. Association between cigar or pipe smoking and cancer risk in men: a pooled analysis of five cohort studies. Cancer Prev. Res. 10, 704–709. [DOI] [PubMed] [Google Scholar]
  28. Messer K, White MM, Strong DR, Wang B, Shi Y, Conway KP, Pierce JP, 2014. Trends in use of little cigars or cigarillos and cigarettes among US smokers, 2002–2011. Nicotine Tob. Res. 17, 515–523. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Montgomery L, Oluwoye O, 2016. The truth about marijuana is all rolled up in a blunt: prevalence and predictors of blunt use among young African-American adults. J. Subst. Use 21, 374–380. [Google Scholar]
  30. Nonnemaker J, Rostron B, Hall P, MacMonegle A, Apelberg B, 2014. Mortality and economic costs from regular cigar use in the United States, 2010. Am. J. Public Health 104, e86–e91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Peters EN, Budney AJ, Carroll KM, 2012. Clinical correlates of co-occurring cannabis and tobacco use: a systematic review. Addiction 1078, 1404–1417. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Peters EN, Schauer GL, Rosenberry ZR, Pickworth WB, 2016. Does marijuana blunt smoking contribute to nicotine exposure? Preliminary product testing of nicotine content in wrappers of cigars commonly used for blunt smoking. Drug Alcohol Depend. 168, 119–122. [DOI] [PubMed] [Google Scholar]
  33. Pickworth W, Rosenberry ZR, Koszowski B, 2016. Toxicant exposure from smoking a little cigar: further support for product regulation. Tob. Control 26, 269–276. [DOI] [PubMed] [Google Scholar]
  34. Rath JM, Villanti AC, Abrams DB, Vallone DM, 2012. Patterns of tobacco use and dual use in US young adults: the missing link between youth prevention and adult cessation. J. Environ. Public Health. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Richardson A, Xiao H, Vallone DM, 2012. Primary and dual users of cigars and cigarettes Profiles, tobacco use patterns and relevance to policy. Nicotine Tob. Res. 15, 1729–1736. [DOI] [PubMed] [Google Scholar]
  36. Richardson A, Rath J, Ganz O, Xiao H, Vallone D, 2013. Primary and dual users of little cigars/cigarillos and large cigars: demographic and tobacco use profiles. Nicotine Tob. Res. 15, 1729–1736. [DOI] [PubMed] [Google Scholar]
  37. Richardson A, Williams V, Rath J, Villanti AC, Vallone D, 2014. The next generation of users: prevalence and longitudinal patterns of tobacco use among US young adults. Am. J. Public Health 104, 1429–1436. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Rosenberry ZR, Pickworth WB, Koszowski B, 2016. Large cigars: smoking topography and toxicant exposure. Nicotine Tob. Res. 20, 183–191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Schauer GL, Berg CJ, Kegler MC, Donovan DM, Windle M, 2015. Assessing the overlap between tobacco and marijuana: Trends in patterns of co-use of tobacco and marijuana in adults from 2003–2012. Addict. Behav. 49, 26–32. [DOI] [PubMed] [Google Scholar]
  40. Schauer GL, Berg CJ, Kegler MC, Donovan DM, Windle M, 2016. Differences in tobacco product use among past month adult marijuana users and nonusers: findings from the 2003–2012 National Survey on Drug Use and Health. Nicotine Tob. Res. 18, 281–288. [DOI] [PubMed] [Google Scholar]
  41. Schauer GL, Rosenberry ZR, Peters EN, 2017. Marijuana and tobacco co-administration in blunts, spliffs, and mulled cigarettes: a systematic literature review. Addict. Behav. 64, 200–211. [DOI] [PubMed] [Google Scholar]
  42. Sifaneck SJ, Johnson BD, Dunlap E, 2006. Cigars-for-blunts: choice of tobacco products by blunt smokers. J. Ethn. Subst. Abuse 4, 23–42. [DOI] [PubMed] [Google Scholar]
  43. Stallones RA, 2015. The association between tobacco smoking and coronary heart disease. Int. J. Epidemiol. 44, 735–743. [DOI] [PubMed] [Google Scholar]
  44. Substance Abuse and Mental Health Services Administration, Center for Behavioral Health Statistics and Quality, 2016. 2015 National Survey on Drug Use and Health: Detailed Tables. Substance Abuse and Mental Health Services Administration, Rockville, MD. [Google Scholar]
  45. Sussman S, Arnett JJ, 2014. Emerging adulthood: developmental period facilitative of the addictions. Eval. Health Prof. 37, 147–155. [DOI] [PubMed] [Google Scholar]
  46. Terchek JJ, Larkin EM, Male ML, Frank SH, 2009. Measuring cigar use in adolescents: inclusion of a brand-specific item. Nicotine Tob. Res. 11, 842–846. [DOI] [PubMed] [Google Scholar]
  47. Trapl ES, Terchek JJ, Danosky L, Cofie L, Brooks-Russell A, Frank SH, 2011. Complexity of measuring cigar use in adolescents: results from a split sample experiment. Nicotine Tob. Res. 13, 291–295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Villanti AC, Richardson A, Vallone DM, Rath JM, 2013. Flavored tobacco product use among US young adults. Am. J. Prev. Med. 44, 388–391. [DOI] [PubMed] [Google Scholar]
  49. Villanti AC, Johnson AL, Ambrose BK, Cummings KM, Stanton CA, Rose SW, Feirman SP, Tworek C, Glasser AM, Pearson JL, Cohn AM, Conay KP, Niaura RS, Bansal-Travers M, Hyland A, 2017. Flavored tobacco product use in youth and adults: findings from the first wave of the PATH Study (2013–2014). Am. J. Prev. Med. 53, 139–151. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES