This originally appeared at https://pubmed.ncbi.nlm.nih.gov/33464670/
Authors Hollis C Karoly 1 2 , J Megan Ross 3 , Mark A Prince 2 , Alexandra E Zabelski 1 , Kent E Hutchison 1 3 4
PMID: 33464670 PMCID: PMC8286984 DOI: 10.1111/add.15407
Abstract
Background and Aims.
Cannabis is commonly used among people who drink alcohol, but evidence suggests a nuanced relationship between alcohol consumption and cannabis use. In particular, among individuals undergoing alcohol treatment, the impact of cannabis on alcohol intake may depend on cannabis use frequency. We aimed to test the effects of within-day cannabis use on total drinks consumed and likelihood of binge drinking on a given day across all participants and compare these relationships between males and females and between individuals who reported infrequent and frequent cannabis use.
Design.
This observational study is a sub-study of a larger randomized controlled trial (RCT). Individuals were included from the RCT if they reported any cannabis use and were divided into groups based on cannabis use patterns. Alcohol use was compared within and between groups.
Setting.
Individuals were recruited from 2016–2020 from community and university settings in Denver and Boulder, Colorado, USA.
Participants.
Of the 182 individuals enrolled in the RCT, N=96 cannabis-using subjects were included in these analyses.
Measurements.
Subjects completed a Timeline Followback (TLFB) at baseline, 4 weeks, 8 weeks (end of treatment), and 20 weeks. Daily data on alcohol and cannabis use from the TLFB at all timepoints were analyzed.
Findings.
Across the sample (N=96), individuals drank approximately 29% fewer drinks (95% Confidence Interval (C.I.) 18%−39%, p<.001) and were 2.06 times (95% C.I. 1.37–3.08, p<.001) less likely to have a binge-drinking episode on days that cannabis was used compared with days cannabis was not used. These patterns were observed in males, females and the infrequent and frequent cannabis use groups. Findings were inconclusive regarding differences in the association between cannabis use and alcohol outcomes when comparing males and females and when comparing infrequent and frequent cannabis use groups.
Conclusions.
Heavy drinkers engaged in treatment to reduce their alcohol consumption who also use cannabis appear to increase their cannabis use on days when they reduce their alcohol consumption.
Keywords: alcohol, cannabis, substitution, complementarity, multilevel modeling, intervention
Introduction
Cannabis is commonly used among people who drink alcohol (1), and the existing epidemiologic literature supports both “substitution” effects, whereby cannabis is substituted for alcohol, thereby benefiting public health through decreasing alcohol use, as well as “complementarity” effects, whereby cannabis is associated with increased alcohol use (2,3). Ecological Momentary Assessment (EMA) (e.g., using twice-daily online assessments of alcohol and cannabis use) and longitudinal survey studies have been the primary methods for exploring alcohol and cannabis co-use. Co-use includes both simultaneous (using both substances such that their effects overlap) and concurrent (using both substances at non-overlapping time) use patterns. Notably, no existing studies have directly tested the effects of simultaneous vs. concurrent use patterns on the likelihood that cannabis will serve as a substitute or complement for alcohol, but it research suggests that temporal relationships between cannabis and alcohol use matter, such that simultaneous use may promote complementarity while concurrent use patterns may be associated with substitution. (2). Substitution has also been supported in states that have more lenient policies surrounding cannabis use (3,4). This is particularly true among medical cannabis users, as 40% of medical users report using cannabis to decrease alcohol intake (5), alcohol consumption has decreased in medical cannabis states (6), and medical users drink less and have fewer alcohol-related problems than non-medical users (7).
A number of individual-level variables also appear to influence the relationship between alcohol and cannabis use. In particular, males may be more likely to use cannabis as a complement to alcohol under lenient cannabis policies (8), and tobacco use may increase the likelihood of alcohol and cannabis use (9).
Current evidence also suggests a nuanced relationship between alcohol and cannabis among individuals in alcohol treatment, depending upon severity of cannabis use. Specifically, individuals who used cannabis infrequently during AUD treatment had fewer days abstinent from alcohol than cannabis abstainers, but there was no relationship between alcohol and cannabis among the most frequent cannabis users (10). Lack of a significant relationship between cannabis and alcohol in the most frequent cannabis users was replicated in a sample of individuals who had previously been to AUD treatment (11). This study demonstrated that only midlevel cannabis use (using more than monthly but less than weekly) was related to greater alcohol-related harm and persistent AUD compared to no cannabis use. Heavier (more than weekly) or lighter (less than monthly) cannabis use was not associated with worse alcohol outcomes compared to cannabis abstinence. Notably, these studies focused on how between-subjects differences in cannabis use impacted abstinence and alcohol use over the course of 12 or more months (10,11), and neither study examined the effects of cannabis on alcohol consumption within a given substance-using day. Several EMA studies have explored within-subjects associations between cannabis and alcohol use, but they have been focused on adolescent and young adult participants, and thus do not shed light on the temporal relationships between cannabis and alcohol consumption in adult heavy drinkers or individuals engaged in AUD treatment (12–15). Given the nuanced relationships between cannabis and alcohol use (2) and the lack of within-subjects data on this topic, it is important to examine how cannabis impacts alcohol use within a particular substance-using day.
It is also important to directly test these relationships in an AUD treatment sample, given that individuals seeking AUD treatment are often encouraged to stop using other substances (including cannabis) prior to beginning treatment, presumably due to evidence suggesting that cannabis use has a detrimental effect on alcohol abstinence (10,16). However, alcohol abstinence is not always the ultimate goal for individuals with AUD, and although complete cannabis and alcohol abstinence may be ideal for many patients, some individuals may experience significant benefits from simply reducing their daily alcohol consumption (17,18). At present, we lack data on the within-day effects of cannabis on alcohol consumption during AUD treatment. Because cannabis cessation is challenging for many individuals, stopping cannabis use prior to AUD treatment may be a treatment barrier for some habitual cannabis users (19,20). Thus, it is important to determine whether cannabis cessation is truly necessary for AUD treatment to be effective. Empirical data are needed to shed light on these questions and inform treatment recommendations.
The present study thus explores the within-day effects of cannabis use on alcohol consumption and binge drinking before, during, and after an 8-week manualized intervention in a sample of heavy-drinking individuals who also use cannabis and are located in Colorado. There has been little research to date exploring the effects of cannabis use on alcohol outcomes in treatment-seeking drinkers in locations where recreational and medical cannabis is legal and accessible, and there is pressing need for studies exploring this relationship. The aims of this study were to:
- Test whether using cannabis on a given substance-using day predicts the number of alcoholic drinks consumed or binge-drinking likelihood on that day in a sample of treatment-engaged, heavy drinkers who also use cannabis.
- Test whether the relationship between within-day cannabis and alcohol consumption and binge drinking likelihood is moderated by sex.
- Test whether the relationship between within-day cannabis and alcohol consumption and binge drinking likelihood is moderated by frequency of cannabis use (infrequent vs. frequent use).
Given the lack of prior data to inform directionality of hypotheses about the relationships between cannabis use and alcohol consumption in this population, the associations investigated here are considered exploratory, thus no hypotheses were pre-registered.
Method.
Participants were recruited from the Boulder-Denver metro area using social media postings and mailed flyers. Research assistants screened participants via phone. Participants were included if they met the following criteria: aged 21–60; no more than 10 days since last drink; interested in reducing drinking, reported drinking over 14 drinks/week on average over past 3 months, and have a breath alcohol level of 0.00 g/dL at baseline (to sign consent form). Participants were excluded if they were undergoing alcohol withdrawal, taking medications to treat bipolar or psychotic disorders, endorsed current suicidality, met criteria for psychotic disorder, bipolar disorder or a current major depressive episode, had a positive urine drug screen (for illicit drugs besides cannabis), reported using illicit drugs in the 30 days prior to beginning the study, or were pregnant. Eligible individuals completed a baseline appointment followed by eight weeks of therapy (with additional data collection at weeks 4 at 8) and follow up visits at 20 and 32 weeks. The present study is a sub-study of a larger randomized clinical trial (RCT), which is an alcohol treatment study for which cannabis use was not a primary or secondary measure of interest. Note that cannabis use was not an eligibility requirement for participation in the RCT, and N=96 of the N=182 individuals enrolled in the RCT were cannabis users. Thus, the present study only includes the N=96 cannabis users from the larger RCT sample. In addition, the RCT was paused before data collection was complete due to COVID-19 restrictions on research. At the time of this writing, the trial has not resumed data collection and it is currently projected that no additional data will be collected for this project. Note that when the trial was paused, 45.5% of participants had not yet completed their 32-week follow-up data collection session. Given the high amount of missing data for this timepoint, all 32-week data was excluded from the present study.
At baseline, participants provided informed consent and were administered a Breathalyzer (Intoximeter, Inc., St. Louis, MO) and urinalysis test, and females took a pregnancy test. Participants were administered the Clinical Institute Withdrawal Assessment, and those who scored over an 8 were referred to a doctor for follow-up, and instructed to reach out to the study coordinator after completing medically supervised detoxification if they were still interested in participating. Participants completed questionnaires measuring health, demographics, history of drinking, smoking and drug use, depression, anxiety, and alcohol craving. They also completed additional measures pertinent to other study aims not relevant for the present analysis (e.g., neurocognitive measures, blood draws).
Participants were randomly assigned to receive either 8 weeks of manualized Mindfulness Based Relapse Prevention (MBRP) or Relapse Prevention (RP) treatment (21). After completing an initial goal setting session, participants met with their assigned therapist (a graduate student or post-doctoral therapist with a PhD in clinical psychology) weekly for an individual session lasting one hour. At weeks 4 and 8, participants completed an assessment of their alcohol and cannabis use over the past month. Additional follow-up assessments took place three months and six months after the end of the 8-week treatment phase (20-weeks and 32-weeks after starting treatment, respectively). However, because the 32-week data collection timepoint was not possible for nearly half of all study participants due to COVID-19 research restrictions, the 32-week follow-up data was not included in the present analyses. At each follow-up, participants reported on their alcohol and cannabis use over the past month.
Measures
Demographics Questionnaire.
This questionnaire was administered at the baseline appointment to collect information on age, sex, marital status, yearly income, occupation, education, and race.
The Beck Depression Inventory-II (BDI-II)
(22) is a 21-item measure of depression symptom severity over the past two weeks. BDI-II scores range between 0 and 63. Baseline BDI total score was included as a potential covariate.
The Beck Anxiety Inventory (BAI)
(23) describes 21 common symptoms of anxiety. The items are summed to obtain a total score that can range from 0 to 63.The baseline BAI total score was included as a potential covariate.
The Timeline Follow-Back (TLFB)
(24) is a calendar-based retrospective recall measure of daily use of cannabis, tobacco, alcohol and other substances for 30-days prior to each study appointment. The TLFB has been demonstrated to have good psychometric characteristics and can generate variables that provide a wide range of information about an individual’s drinking (25). It has also been used for daily-level analysis of alcohol and cannabis consumption and co-use (26,27). The total number of drinks per day was used to quantify alcohol consumption (range 0–29) and whether the participant engaged in a binge-drinking episode on that day (≥4 drinks for women and ≥5 drinks for men) was coded as 0=no binge drinking episode or 1=binge drinking episode. In addition, we created groups based on the average monthly number of days that cannabis was used over the course of the study. Individuals who use cannabis on ≥15 days per month, on average, were categorized as frequent users (n=28) while individuals who use cannabis on <15 days per month, on average, were categorized as infrequent users (n=68). Cannabis and tobacco consumption were quantified by daily endorsement of cannabis or cigarette use.
Analytic Plan and Results
Statistical Analysis
Mplus 8 (28) was used to estimate multilevel structural equation models (29) to estimate the number of drinks consumed or whether the participant endorsed a binge-drinking episode (dichotomized to binge or non-binge episode) on days that cannabis was used compared to days that cannabis was not used within-individual. More specifically, a within-level random slope was calculated predicting either the count of drinks consumed (modeled with a negative binomial regression) or the likelihood of reporting a binge drinking episode (modeled with a logistic regression). The primary predictor variable was whether a person endorsed cannabis use on a given day (yes or no). In addition, relevant baseline covariates were included as between-level predictors of the random slope. Covariates included tobacco use (due to the demonstrated impact of tobacco on alcohol and substance use (9)), age (due to differential rates of alcohol and cannabis use observed across age groups (30)), BDI and BAI (as depression and anxiety are highly correlated with alcohol and cannabis use (31,32)), and treatment condition (to ensure that treatment condition did not differentially impact study outcomes). This type of cross-level interaction was calculated using the random coefficient prediction method and is notated as a 2 X (1–1) moderation because the moderators are on the between level (i.e., level 2), and the random slope is composed of both a within level predictor (i.e., level 1) and a within level outcome (i.e., level 1) (33). Baseline covariates included, treatment condition (MBRP or RP), sex, age, BDI total score, and BAI total score. In addition, daily endorsement of cigarette use (yes or no) was included as a within level predictor of the random slope in a 1 X (1–1) moderation to control for the daily effects of cigarette use on the random slope of cannabis use day predicting alcohol drinks or binge likelihood.
To further explore differences in the primary relation between cannabis use day and alcohol use, we ran two additional MSEMs. In these models, we examined the within level random slope across 1) sex (males n=58, females n=38)and 2) cannabis use frequency (infrequent compared to frequent cannabis use groups). Because of the negative binomial and logistic regression specifications, models were run with maximum likelihood with robust standard errors and with a Monte Carlo integration algorithm (28). These specifications preclude traditional multi-group analyses in Mplus, so the multi-group analyses needed to be conducted in a multilevel mixture modeling framework using a knownclass specification for the grouping variables. In this framework, a single latent class is specified along with a categorical knownclass variable to indicate groupings. This allows for the estimation of within and between level effects across groups in the same MSEM, as well as a direct comparison of the effects across groups using the model constraint command. More specifically, model constraints were used to calculate the absolute difference between the effects for each group and then a z-test was conducted to determine if the effects statistically differed from each other. In the present study, we used the format for both sex and cannabis use frequency group in two separate models.
In sum, we ran a series of MSEMs to address the study hypotheses. First, we tested the random within level slope of cannabis use day predicting alcohol drinks or binge likelihood (2 models). Next, we re-ran these models adding baseline covariates (2 models). None of the covariates were significantly associated with number of daily drinks or binge drinking episodes (p>.181). Thus, all covariates were removed from subsequent analyses. Finally, we ran the multi-group analyses estimating the within level random effects by sex and cannabis-use status (2 models). For ease of interpretation, we report incident rate ratios for negative binomial regression paths (34) and odds ratios for logistic regression paths by exponentiating the unstandardized regression coefficients. Because study aims focused on testing the effects of cannabis on alcohol consumption within a given substance-use day, days in which no substance use occurred were not of interest. Analyses excluded days in which participants did not use alcohol or cannabis but included all days in which participants reported using alcohol, cannabis or both. Little’s MCAR test indicated that data were missing completely at random (χ2(3) = 0.699, p=.873). Missing data was handled with maximum likelihood estimation with robust standard errors.
Post hoc power analyses followed recommendation from Arend and Schafer (35). Using simulation studies Arend and Schafer examined small, medium, and large ICC values and random slope variances for level 1 units ranging in size from 3 to 30 and level 2 cluster sizes ranging from 30 to 200. Thus, a maximum number of observations of 30*200=6000. They determined that these ranges were sufficient to detect a range of effects for all but cross-level interactions with small random slope variances. The analytic sample had 96 level 2 clusters (i.e., cannabis using participants) and an average of 76.68 level 1 units (i.e., reporting days) per cluster (total observations=7,361). The intraclass correlation (ICC) in the present sample for likelihood of binge drinking was medium sized (ICC=.31) and was large for the count of drinks per drinking day (ICC=.56). The random slope variance for the count of drinks per day was large (random slope variance=.24), and the random slope variance for binge was not calculated because it was treated as a categorical outcome. Taken together, the present study was adequately powered for all MSEMs presented.
Sample Demographics
Demographic information is included in Table 1. The sample contained 182 participants, including 96 individuals who use cannabis. Among the individuals who use cannabis, there was a total of 7,361 observations across the 96 participants. A majority of the people who use cannabis in this study were male (60%) with a mean age of 46 years, and 51% of those who use cannabis were randomized to the MBRP treatment. Most individuals in the sample were white (95%), had at least a bachelor’s degree (73%), reported an annual income of at least $40,000 (72%), and were employed full-time (76%). Given the lack of variability in race, education, income, and employment, these variables were not included as covariates in any analyses.
Table 1.
Participant Characteristics
| Variable | Total sample (N=96) | Infrequent cannabis use (n=68) | Frequent cannabis use (n=28) | Male (n=58) | Females (n=38) |
|---|---|---|---|---|---|
| Sex (% male) | 60.42 | 61.76 | 57.14 | 100.00 | 0.00 |
| Age | 45.45 (10.51) | 44.75 (10.52) | 47.14 (10.48) | 45.55 (10.61) | 45.29 (10.49) |
| Treatment condition (% MBRP) | 51.00 | 50.00 | 53.57 | 53.45 | 47.37 |
| BDI | 13.64 (9.10) | 13.61 (9.25) | 13.70 (8.87) | 14.14 (9.75) | 12.86 (8.06) |
| BAI | 7.49 (6.61) | 7.53 (6.74) | 7.38 (6.38) | 7.33 (6.53) | 7.71 (6.80) |
| Mean % of drinking episodes that are binges | 48.79 (30.23) | 49.94 (28.98) | 46.00 (33.46) | 48.24 (30.38) | 49.62 (30.39) |
| Mean number of drinks/month | 104.21 (87.79) | 104.11 (93.31) | 104.46 (74.24) | 116.81 (98.82) | 84.98 (64.15) |
| Mean % of cannabis use days | 35.03 (35.35) | 14.10 (13.37) | 85.88 (13.00) | 34.24 (33.20) | 36.24 (38.85) |
| Mean % of nicotine use days | 20.28 (37.39) | 15.91 (33.49) | 30.90 (44.39) | 19.56 (37.34) | 21.38 (37.94) |
Note: All values are means (standard deviations) unless noted. MBRP=Mindfulness Based Relapse Prevention, BDI=Beck Depression Inventory, and BAI=Beck Anxiety Inventory. Group differences were examined between individuals who use cannabis infrequently compared to frequently and males compared to females. Bold indicates significant group differences on that variable.
Within-Individual Effects of Cannabis Use on Drinking Outcomes
Individuals drank, on average, 29% fewer drinks [95% Confidence Interval, 18%−39%], p<.001, on cannabis use days compared to days that cannabis was not used. Similarly, individuals were 2.06 less likely [1.37–3.08], p<.001, to engage in a binge-drinking episode on days that cannabis was used compared to non-cannabis use days.
Within-Individual Effects of Cannabis on Drinking Outcomes by Sex
Females drank 32% [10%−49%], p=.007, and males drank 28% [13%−40%], p=.001, fewer drinks on cannabis use days compared to non-cannabis use days. Males and females were not significantly different in terms of the effect of cannabis use on number of drinks (4.8% difference, p=.684).
On cannabis use days, females were 2.89 times less likely [1.74–4.77] to binge drink, p<.001, compared to non-cannabis use days. Males were 1.78 times less likely [1.10–2.87], p=.019, to binge drink on cannabis use days compared to non-cannabis use days. There was not a significant difference in the likelihood of a binge drinking episode on cannabis use days compared to non-cannabis use days between males and females, ORdifference=1.11, p=.182. See Table 2 for results summary.
Table 2.
Comparison of the within-individual effects of cannabis use on number of drinks consumed and likelihood of a binge-drinking episode
| Sample | Number of drinks Percent [95% CI intervals] (p-value) | Binge drinking episode Odds ratio [95% CI intervals] (p-value) |
|---|---|---|
| Whole sample analysis (N=96) | 29.04% [18.13% – 38.49%] (p<.001) | 2.06 [1.37–3.08] (p<.001) |
| Type of cannabis user as a moderator | ||
| Infrequent cannabis use (n=68) | 28.04% [5.92% – 44.95%] (p=.016) | 2.26 [1.44–3.55] (p<.001) |
| Frequent cannabis use (n=28) | 28.18% [9.97% – 42.76%] (p=.004) | 1.16 [0.40–3.35] (p=.786) |
| Gender as a moderator | ||
| Males (n=58) | 27.60% [12.63% – 39.95%] (p=.001) | 1.78 [1.10–2.87] (p=.019) |
| Females (n=38) | 32.29% [9.88% – 49.14%] (p=.007) | 2.89 [1.74–4.77] (p<.001) |
Note: Percent = percent reduction in number of drinks on cannabis use days compared to non-cannabis use days. Odds ratio = Likelihood of not having a binge drinking episode on cannabis use days compared to non-cannabis use days. Frequent cannabis use = average monthly use ≥ 15 days, infrequent cannabis use = average monthly use <15 days.
*indicates significant group differences. Significant effects are bolded.
Within-Individual Effects of Cannabis on Drinking Outcomes by Cannabis Use Group
Individuals who use cannabis infrequently drank 28% fewer drinks [6%−45%], p=.016, and individuals who use cannabis frequently also drank 28% fewer drinks [10%−43%], p=.004, on cannabis use days compared to non-cannabis use days. Individuals who reported frequent and infrequent cannabis use reduced their daily alcohol consumption by similar amounts (<1% difference, p=.990) on cannabis use days compared to non-use days.
Individuals who use cannabis infrequently were 2.26 times less likely [1.44–3.55], p<.001, to engage in a binge-drinking episode on cannabis use days compared to non-cannabis use days. However, there was not a significant effect of cannabis use on binge-drinking episodes for individuals who use cannabis frequently (OR = 1.16, CI = [0.40–3.35]). In addition, there was not a significant difference in the odds of reporting a binge-drinking episode between those who use cannabis infrequently compared to frequently on cannabis use days compared to non-cannabis use days, ORdifference=1.10, p=.387.
Discussion.
The present study demonstrated that in a sample of treatment-involved heavy drinkers, cannabis use on a given day is associated with consuming fewer drinks on that day and lower odds of binge-drinking. These findings are consistent with the literature supporting cannabis substitution in an environment with legalized recreational cannabis. For example, in the years during and immediately following recreational cannabis legalization in Washington state, there was an increase in cannabis use and a decrease in alcohol-related harms (4).
Notably, in contrast to some prior work suggesting a detrimental effect of cannabis on drinking outcomes (2,27,36), both the infrequent and frequent cannabis use group in this study consumed 28% fewer drinks on days when cannabis was used. Conversely, one recent study suggested that among individuals engaged in alcohol treatment who also use cannabis, those who used cannabis once or twice per-month had fewer alcohol-abstinent days at the end of treatment, whereas for those who used more frequently, cannabis was not associated with increased alcohol use during treatment compared to cannabis abstainers (10).This finding was replicated in a study of individuals who had previously been to alcohol treatment (11).
These previous studies suggest that cannabis use during alcohol treatment may be detrimental, at least for individuals who use cannabis relatively infrequently (once or twice per month). In contrast, our data suggest that cannabis use on a given day is associated with decreased drinking on that day for both the frequent and infrequent use groups, and decreased binge drinking likelihood on that day for the infrequent use group (note that the odds ratio for reporting a binge episode on cannabis use days compared to non-cannabis use days was nearly double the size for infrequent users compared to frequent users, despite the group difference being non-significant). One important factor that may contribute to our pattern of results is the fact that cannabis substitution often occurs under lenient cannabis policies (37). Subjects in the present study were living in Colorado, where recreational cannabis was legalized in 2014, whereas prior studies have drawn from national samples and included individuals living under strict cannabis policies. Another explanation for the disparate findings among infrequent users is that prior studies were testing between-subjects differences in overall percentage of days abstinent from alcohol, while this study explored within-day effects of cannabis on alcohol consumption (10,11). Although cannabis use among infrequent users use may decrease the likelihood of abstinence overall (as suggested in prior studies), our results suggest that cannabis use on a specific day appears to be associated with decreased alcohol consumption on that particular day for both the frequent and infrequent use groups. It is also worth noting that the people who used infrequently in our sample used cannabis more frequently (average of 4.2 days/month of cannabis use) than the mid-level users (1–2 times/month) in the prior studies. Perhaps cannabis use among people who are very infrequent users (i.e., those who use cannabis once or twice per month as in prior studies (10,11)) is likely to occur on special occasions when greater alcohol consumption is also more likely to occur. In contrast, cannabis use among individuals who use at least once per week (including the infrequent and frequent users our sample) may be more likely to take the place of alcohol (as demonstrated by the negative association between cannabis and alcohol consumption in the present sample) or not impact alcohol consumption at all (as demonstrated by prior studies and the lack of relationship between cannabis and binge-drinking in the frequent cannabis use group in the present study).
Across the whole study sample, none of the covariates of interest were significantly associated with number of daily drinks or binge-drinking episodes, thus all covariates were removed from primary analyses. In contrast, prior literature has showed associations between depression and anxiety and increased drinking (31,38), and between tobacco use and drinking outcomes (39). One explanation for the tobacco finding may be that overall tobacco use in this sample was relatively low, and effects of tobacco use on drinking outcomes are more likely to emerge at higher levels of tobacco use (39). Our observed lack of association between depression and anxiety and drinking outcomes may be due to the fact that BDI scores were in the mild range and BAI scores were in the minimal range in this sample. It is unsurprising that treatment group was not associated with alcohol use outcomes, as both RP and MBRP have demonstrated good efficacy in decreasing alcohol use and improving alcohol-related outcomes (40,41). Finally, our lack of age effects on drinking are likely due to the developmentally restricted age range of participants in the current study (age 21–60). When compared to adolescents or older adults, age-related differences in drinking patterns may be more likely to emerge.
The present study also tested whether sex moderates the relationship between cannabis and alcohol use, given substantial evidence that men drink more frequently than women, on average (42) and that men use cannabis more frequently and in higher quantities than women (43). In the present sample, both males and females drank less and were less-likely to binge-drink on days when cannabis was used. Although significant differences did not emerge between males and females on these outcomes, the magnitude of the effect of cannabis on alcohol consumption was larger for females compared to males. Thus, further exploring sex and gender differences in the effects of cannabis on alcohol use is warranted. Existing studies also support the need for future research on sex and gender differences in this area, as research suggests that women develop CUDs at a faster rate than men (44), have an increased rate of cannabis withdrawal and report greater severity of certain withdrawal symptoms compared to men (45) and are more likely to have a comorbid mood or anxiety disorder (46).
In sum, these results suggest that it may not be necessary for individuals seeking AUD treatment to stop using cannabis prior to treatment. Specifically, these data suggest that cannabis does not have a detrimental impact on within-day alcohol use during treatment or in the three months following treatment. This pattern holds for both males and females, as well as for infrequent and frequent cannabis users. However, there are a number of limitations that should be considered when interpreting these results, and future studies are warranted to better understand the nuanced relationship between cannabis and alcohol use across various populations. From a policy perspective, it will be important to compare results of studies testing the relationships between cannabis and alcohol use across states with stricter cannabis policies and states with more lenient policies.
Limitations and Future Directions
This study had several limitations of note. First, alcohol and cannabis use data were collected using the TLFB, which is subject to the pitfalls of retrospective recall data. Further, a strength of this study was our examination of the within-day effects of cannabis on alcohol use, however even more fine-grained information could be gleaned from future studies using Ecological Momentary Assessment (EMA) approaches that explore the effects of cannabis on alcohol consumption in a particular substance-use session. From the data collected in the present study, it is also not possible to determine the reasons for participants’ reduction in drinking on cannabis use days. For example, were participants deliberately using cannabis to fill the ‘spaces’ left by reduced alcohol consumption, or was cannabis facilitating a reduction in drinking by some other mechanism (e.g., reducing the reward value of alcohol, reducing desire to socialize, impacting overall level of subjective intoxication, etc.)? Future EMA or daily diary studies could explore this question through directly asking participants about their cannabis and alcohol use motives during each substance use session.
Also, heavy drinkers were not required to meet DSM criteria for alcohol use disorder in order to participate in this study. Findings need to be replicated among individuals with AUD diagnoses. Additionally, individuals in this study were recruited from the Denver and Boulder, Colorado region and were 95% white. It is unknown whether these results would generalize to other populations. Future work in this area should prioritize the inclusion of racially and ethnically diverse populations. Finally, the present study did not assess the cannabinoid content of the cannabis used by participants. Given the different effects of delta-9-tetrahydrocannabinol (THC; the primary psychoactive constituent in the Cannabis plant used for medical and recreational purposes) and cannabidiol (CBD; a non-psychoactive cannabinoid also commonly present in recreational and medical cannabis) on alcohol-related outcomes (47–52), it will also be important for future studies to directly assess the amount of THC versus CBD consumed by participants on the TLFB or other self-report measures.
Conclusions
The present study suggests that among individuals in Colorado, where recreational cannabis has been legal and accessible for over 5 years, the use of cannabis during alcohol treatment is associated with decreased within-day alcohol intake for both males and females and among individuals who use cannabis infrequently and frequently. Future work is needed to determine how cannabis impacts alcohol use at later timepoints post-treatment (e.g., 1 year or more post-treatment), as well as how these patterns may differ across different populations and among those with more severe AUD.
Funding
This work was supported by R01AA024632 to KEH. JMR was supported by National Institutes of Health grant T32DA017637. The content is solely the responsibility of the authors and does not represent the opinion of the National Institutes of Health.
Footnotes
Conflicts of interest: None
Clinicaltrials.gov registration number: NCT02994043
References
1. Subbaraman MS, Kerr WC. Simultaneous versus concurrent use of alcohol and cannabis in the National Alcohol Survey. Alcohol Clin Exp Res [Internet]. 2015. May [cited 2019 Jan 6];39(5):872–9. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25872596 [PMC free article] [PubMed] [Google Scholar]
2. Subbaraman MS. Substitution and Complementarity of Alcohol and Cannabis: A Review of the Literature. Subst Use Misuse [Internet]. 2016. September 18 [cited 2018 Nov 15];51(11):1399–414. Available from: http://www.tandfonline.com/doi/full/10.3109/10826084.2016.1170145 [PMC free article] [PubMed] [Google Scholar]
3. Guttmannova K, Lee CM, Kilmer JR, Fleming CB, Rhew IC, Kosterman R, et al. Impacts of Changing Marijuana Policies on Alcohol Use in the United States. Alcohol Clin Exp Res [Internet]. 2016. January [cited 2019 Jan 6];40(1):33–46. Available from: http://www.ncbi.nlm.nih.gov/pubmed/26727520 [PMC free article] [PubMed] [Google Scholar]
4. Subbaraman MS, Kerr WC. Subgroup trends in alcohol and cannabis co-use and related harms during the rollout of recreational cannabis legalization in Washington state. Int J Drug Policy. 2020. January 1;75:1–7. [PMC free article] [PubMed] [Google Scholar]
5. Reiman A Cannabis as a substitute for alcohol and other drugs. Harm Reduct J. 2009;6(1):35. [PMC free article] [PubMed] [Google Scholar]
6. Mark Anderson D, Hansen B, Rees DI. Medical Marijuana Laws, Traffic Fatalities, and Alcohol Consumption. J Law Econ [Internet]. 2013. May 18 [cited 2018 Nov 19];56(2):333–69. Available from: https://www.journals.uchicago.edu/doi/10.1086/668812 [Google Scholar]
7. Subbaraman MS, Kerr WC. Alcohol Use and Risk of Related Problems Among Cannabis Users Is Lower Among Those With Medical Cannabis Recommendations, Though Not Due To Health. J Stud Alcohol Drugs [Internet]. 2018. November 21 [cited 2019 Jan 14];79(6):935–42. Available from: https://www.jsad.com/doi/10.15288/jsad.2018.79.935 [PMC free article] [PubMed] [Google Scholar]
8. Williams J, Mahmoudi P. Economic Relationship Between Alcohol and Cannabis Revisited. Econ Rec [Internet]. 2004. March 1 [cited 2020 Mar 30];80(248):36–48. Available from: http://doi.wiley.com/10.1111/j.1475-4932.2004.00123.x [Google Scholar]
9. Roche DJO, Bujarski S, Green R, Hartwell EE, Leventhal AM, Ray LA. Alcohol, tobacco, and marijuana consumption is associated with increased odds of same-day substance co- and tri-use. Drug Alcohol Depend. 2019. July 1;200:40–9. [PMC free article] [PubMed] [Google Scholar]
10. Subbaraman MS, Metrik J, Patterson D, Swift R. Cannabis use during treatment for alcohol use disorders predicts alcohol treatment outcomes. Addiction [Internet]. 2017. April 1 [cited 2019 Apr 22];112(4):685–94. Available from: http://doi.wiley.com/10.1111/add.13693 [PMC free article] [PubMed] [Google Scholar]
11. Subbaraman MS, Barnett SB, Karriker‐Jaffe KJ. Risks Associated with Mid level Cannabis Use Among People Treated for Alcohol Use Disorder. Alcohol Clin Exp Res [Internet]. 2019. April 4 [cited 2019 Apr 22];43(4):690–4. Available from: https://onlinelibrary.wiley.com/doi/abs/10.1111/acer.13973 [PMC free article] [PubMed] [Google Scholar]
12. Patrick ME, Terry-McElrath YM, Lee CM, Schulenberg JE. Simultaneous alcohol and marijuana use among underage young adults in the United States. Addict Behav [Internet]. 2019. January 1 [cited 2019 Jan 14];88:77–81. Available from: https://www.sciencedirect.com/science/article/pii/S030646031830515X [PMC free article] [PubMed] [Google Scholar]
13. Terry-McElrath YM, Patrick ME. Simultaneous Alcohol and Marijuana Use Among Young Adult Drinkers: Age-Specific Changes in Prevalence from 1977 to 2016. Alcohol Clin Exp Res. 2018. November 1;42(11):2224–33. [PMC free article] [PubMed] [Google Scholar]
14. White HR, Kilmer JR, Fossos‐Wong N, Hayes K, Sokolovsky AW, Jackson KM. Simultaneous Alcohol and Marijuana Use Among College Students: Patterns, Correlates, Norms, and Consequences. Alcohol Clin Exp Res [Internet]. 2019. July 28 [cited 2020 Feb 3];43(7):1545–55. Available from: https://onlinelibrary.wiley.com/doi/abs/10.1111/acer.14072 [PMC free article] [PubMed] [Google Scholar]
15. Patrick ME, Fairlie AM, Lee CM. Motives for simultaneous alcohol and marijuana use among young adults. Addict Behav. 2018. January 1;76:363–9. [PMC free article] [PubMed] [Google Scholar]
16. Aharonovich E, Liu X, Samet S, Nunes E, Waxman R, Hasin D. Postdischarge cannabis use and its relationship to cocaine, alcohol, and heroin use: A prospective study. Am J Psychiatry [Internet]. 2005. August 1 [cited 2020 Oct 29];162(8):1507–14. Available from: http://psychiatryonline.org/doi/abs/10.1176/appi.ajp.162.8.1507 [PubMed] [Google Scholar]
17. Rosenberg H, Grant J, Davis AK. Acceptance of non-abstinence as an outcome goal for individuals diagnosed with substance use disorders: A narrative review of published research. J Stud Alcohol Drugs. 2020. August 17;81(4):405–15. [PubMed] [Google Scholar]
18. Witkiewitz K, Wilson AD, Roos CR, Swan JE, Votaw VR, Stein ER, et al. Can Individuals With Alcohol Use Disorder Sustain Non-abstinent Recovery? Non-abstinent Outcomes 10 Years After Alcohol Use Disorder Treatment. J Addict Med. 2020; [PMC free article] [PubMed] [Google Scholar]
19. Zvolensky MJ, Rogers AH, Manning K, Hogan JBD, Paulus DJ, Buckner JD, et al. Anxiety sensitivity and cannabis use problems, perceived barriers for quitting, and fear of quitting. Psychiatry Res. 2018. May 1;263:115–20. [PubMed] [Google Scholar]
20. Budney AJ, Vandrey RG, Hughes JR, Thostenson JD, Bursac Z. Comparison of cannabis and tobacco withdrawal: Severity and contribution to relapse. J Subst Abuse Treat. 2008. December 1;35(4):362–8. [PMC free article] [PubMed] [Google Scholar]
21. Bowen S, Witkiewitz K, Clifasefi SL, Grow J, Chawla N, Hsu SH, et al. Relative efficacy of mindfulness-based relapse prevention, standard relapse prevention, and treatment as usual for substance use disorders: a randomized clinical trial. JAMA psychiatry. 2014;71(5):547–56. [PMC free article] [PubMed] [Google Scholar]
22. Beck AT, Steer RA, Carbin MG. Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clin Psychol Rev. 1988;8(1):77–100. [Google Scholar]
23. Beck AT, Steer RA. Manual for the Beck anxiety inventory. 1990;
24. Sobell LC, Sobell MB. Timeline follow-back. In: Measuring alcohol consumption. Springer; 1992. p. 41–72. [Google Scholar]
25. Donohue B, Azrin NH, Strada MJ, Silver NC, Teichner G, Murphy H. Psychometric evaluation of self-and collateral timeline follow-back reports of drug and alcohol use in a sample of drug-abusing and conduct-disordered adolescents and their parents. Psychol Addict Behav. 2004;18(2):184. [PubMed] [Google Scholar]
26. Gunn R, Jackson K, Borsari B, Metrik J. A longitudinal examination of daily patterns of cannabis and alcohol co-use among medicinal and recreational veteran cannabis users. Drug Alcohol Depend. 2019. December 1;205. [PMC free article] [PubMed] [Google Scholar]
27. Metrik J, Gunn RL, Jackson KM, Sokolovsky AW, Borsari B. Daily Patterns of Marijuana and Alcohol Co-Use Among Individuals with Alcohol and Cannabis Use Disorders. Alcohol Clin Exp Res [Internet]. 2018. June [cited 2018 Dec 5];42(6):1096–104. Available from: http://www.ncbi.nlm.nih.gov/pubmed/29656401 [PMC free article] [PubMed] [Google Scholar]
28. Muthén LK, Muthen B. Mplus user’s guide: Statistical analysis with latent variables, user’s guide. Muthén & Muthén; 2017. [Google Scholar]
29. Preacher KJ, Zyphur MJ, Zhang Z. A general multilevel SEM framework for assessing multilevel mediation. Psychol Methods. 2010. September;15(3):209–33. [PubMed] [Google Scholar]
30. Schulenberg J, Johnston L, O’Malley P, Bachman J, Miech R, Patrick M. Monitoring the Future national survey results on drug use, 1975–2019: Volume II, college students and adults ages 19–60. 2020.
31. Boden JM, Fergusson DM. Alcohol and depression. Addiction [Internet]. 2011. May 1 [cited 2020 Mar 30];106(5):906–14. Available from: http://doi.wiley.com/10.1111/j.1360-0443.2010.03351.x [Google Scholar]
32. Lev-Ran S, Roerecke M, Le Foll B, George T, McKenzie K, Rehm J. 892 – The association between cannabis use and depression: a systematic review and meta-analysis of longitudinal studies. Eur Psychiatry. 2013. January;28(S1):1. [Google Scholar]
33. Preacher KJ, Zhang Z, Zyphur MJ. Multilevel structural equation models for assessing moderation within and across levels of analysis. Psychol Methods. 2016;21(2):189–205. [PubMed] [Google Scholar]
34. Hilbe JM. Negative binomial regression. Cambridge University Press; 2011. [Google Scholar]
35. Arend MG, Schäfer T. Statistical power in two-level models: A tutorial based on monte carlo simulation. Psychol Methods. 2019. February 1;24(1):1–19. [PubMed] [Google Scholar]
36. Subbaraman MS, Metrik J, Patterson D, Stout RL. Cannabis use during alcohol treatment is associated with alcohol-related problems one-year post-treatment. Drug Alcohol Depend. 2018. December 1;193:29–34. [PMC free article] [PubMed] [Google Scholar]
37. Risso C, Boniface S, Subbaraman MS, Englund A. Does cannabis complement or substitute alcohol consumption? A systematic review of human and animal studies. J Psychopharmacol [Internet]. 2020. September 1 [cited 2020 Oct 29];34(9):938–54. Available from: http://journals.sagepub.com/doi/10.1177/0269881120919970 [PubMed] [Google Scholar]
38. Schuckit MA, Hesselbrock V. Alcohol Dependence and Anxiety Disorders. Focus (Madison) [Internet]. 2004. July 1 [cited 2020 Mar 30];2(3):440–53. Available from: http://psychiatryonline.org/doi/abs/10.1176/foc.2.3.440 [Google Scholar]
39. Falk DE, Yi H, Hiller-Sturmhöfel S. An epidemiologic analysis of co-occurring alcohol and tobacco use and disorders: findings from the National Epidemiologic Survey on Alcohol and Related Conditions. Alcohol Res Health [Internet]. 2006. [cited 2020 Mar 30];29(3):162–71. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17373404 [PMC free article] [PubMed] [Google Scholar]
40. Witkiewitz K, Marlatt GA. Relapse prevention for alcohol and drug problems: That was Zen, this is Tao. In: Addictive behaviors: New readings on etiology, prevention, and treatment. American Psychological Association; 2009. p. 403–27. [PubMed] [Google Scholar]
41. Witkiewitz K, Marlatt GA, Walker D. Mindfulness-based relapse prevention for alcohol and substance use disorders. Vol. 19, Journal of Cognitive Psychotherapy. Springer Publishing Company; 2005. p. 211–28. [Google Scholar]
42. Wilsnack RW, Wilsnack SC, Kristjanson AF, Vogeltanz-Holm ND, Gmel G. Gender and alcohol consumption: patterns from the multinational GENACIS project. Addiction [Internet]. 2009. September 1 [cited 2020 Mar 30];104(9):1487–500. Available from: http://doi.wiley.com/10.1111/j.1360-0443.2009.02696.x [PMC free article] [PubMed] [Google Scholar]
43. Cuttler C, Mischley LK, Sexton M. Sex Differences in Cannabis Use and Effects: A Cross-Sectional Survey of Cannabis Users. Cannabis Cannabinoid Res [Internet]. 2016. December 28 [cited 2019 Jun 19];1(1):166–75. Available from: http://www.liebertpub.com/doi/10.1089/can.2016.0010 [PMC free article] [PubMed] [Google Scholar]
44. Ehlers CL, Gizer IR, Vieten C, Gilder DA, Stouffer GM, Lau P, et al. Cannabis dependence in the San Francisco Family Study: Age of onset of use, DSM-IV symptoms, withdrawal, and heritability. Addict Behav [Internet]. 2010;35(2):102–10. Available from: 10.1016/j.addbeh.2009.09.009 [PMC free article] [PubMed] [CrossRef] [Google Scholar]
45. Schlienz NJ, Budney AJ, Lee DC, Vandrey R. Cannabis Withdrawal: a Review of Neurobiological Mechanisms and Sex Differences. Curr Addict Reports. 2017. June;4(2):75–81. [PMC free article] [PubMed] [Google Scholar]
46. Khan SS, Secades-Villa R, Okuda M, Wang S, Pérez-Fuentes G, Kerridge BT, et al. Gender differences in cannabis use disorders: Results from the National Epidemiologic Survey of Alcohol and Related Conditions. Drug Alcohol Depend. 2013. June 1;130(1–3):101–8. [PMC free article] [PubMed] [Google Scholar]
47. Viudez-Martínez A, García-Gutiérrez MS, Navarrón CM, Morales-Calero MI, Navarrete F, Torres-Suárez AI, et al. Cannabidiol reduces ethanol consumption, motivation and relapse in mice. Addict Biol [Internet]. 2018. January [cited 2018 Sep 27];23(1):154–64. Available from: http://doi.wiley.com/10.1111/adb.12495 [PubMed] [Google Scholar]
48. Viudez-Martínez A, García-Gutiérrez MS, Fraguas-Sánchez AI, Torres-Suárez AI, Manzanares J. Effects of cannabidiol plus naltrexone on motivation and ethanol consumption. Br J Pharmacol [Internet]. 2018. August [cited 2018 Sep 27];175(16):3369–78. Available from: http://doi.wiley.com/10.1111/bph.14380 [PMC free article] [PubMed] [Google Scholar]
49. Viudez-Martínez A, García-Gutiérrez MS, Manzanares J. Gender differences in the effects of cannabidiol on ethanol binge drinking in mice. Addict Biol [Internet]. 2019. May 9 [cited 2019 May 14];e12765. Available from: http://doi.wiley.com/10.1111/adb.12765 [PubMed] [Google Scholar]
50. Chait L, Perry J. Acute and residual effects of alcohol and marijuana, alone and in combination, on mood and performance. Psychopharmacology (Berl). 1994. July;115(3):340–9. [PubMed] [Google Scholar]
51. Ballard ME, de Wit H. Combined effects of acute, very-low-dose ethanol and delta(9)-tetrahydrocannabinol in healthy human volunteers. Pharmacol Biochem Behav [Internet]. 2011. February [cited 2018 Nov 15];97(4):627–31. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21110996 [PMC free article] [PubMed] [Google Scholar]
52. Ramaekers JG, Theunissen EL, de Brouwer M, Toennes SW, Moeller MR, Kauert G. Tolerance and cross-tolerance to neurocognitive effects of THC and alcohol in heavy cannabis users. Psychopharmacology (Berl) [Internet]. 2011. March 30 [cited 2018 Dec 4];214(2):391–401. Available from: http://link.springer.com/10.1007/s00213-010-2042-1 [PMC free article] [PubMed] [Google Scholar]
