Problematic substance use is one of the most prevalent health issues in our society. In the last Canadian Alcohol and Drug Use Monitoring Survey (1), an estimated 6 million Canadians met the criteria for a substance use disorder. Statistics on polysubstance use are less accessible, but research has shown that dependent drug users report use of over three substances on average (2). Currently, the three most used substances in Canada are alcohol, cannabis, and nicotine (3, 4), all of which can be purchased legally. Although overall prevalence of use varies among substances [i.e., 76% of Canadians 15 years and older reported past year alcohol use (3), 27% of Canadians 16 years and older reported past year cannabis use (5), and 12 and 5% of Canadians 15 years and older reported past-30-day nicotine use via tobacco and vaping, respectively (4)], it is not uncommon for an individual to use two or more of these substances. Indeed, the use of each of these substances is related to increased probability of co-use of the other substances (6). Further, use of one substance may change the consumption amount and frequency of another: alcohol and cannabis co-use leads to increases in alcohol consumption (7, 8); smoking tobacco leads to increases in urges to smoke cannabis, as well as the inverse (9); smokers consume twice as much alcohol as non-smokers (10); and nicotine has been shown to increase alcohol consumption (11). As there is evidence that the concurrent use of nicotine, alcohol and cannabis may lead to increased use of each substance, gaining a better understanding of who is more likely to use and become dependent on these substances is imperative.

Problematic alcohol, cannabis, and nicotine use have been connected to a variety of risk factors in prior research. Age has been implicated as a factor for dependence in all three substances, with individuals often showing the most dependence symptoms in their late teens and early twenties (12). Age of onset has also been implicated in substance dependence, with earlier use of cigarettes being related to higher dependence scores (13). Other research has shown that, while a later age of onset was related to lower alcohol and cannabis use in general, individuals with a later age of onset co-used alcohol and cannabis more frequently than individuals with an earlier age of onset (14). While age of onset seems to relate to dependence on that substance, researchers have also investigated cross-substance age of onset; however, there appears to be little impact across substances, as neither age of onset of alcohol or nicotine use played a role in cannabis dependence, nor did age of onset of nicotine use play a role in alcohol dependence (13, 15).

Personality has also been implicated in dependence on alcohol, cannabis, and nicotine, particularly the traits measured by the Substance Use Risk Profile Scale (SURPS) (16). The SURPS assesses four personality traits: hopelessness/introversion, anxiety sensitivity, impulsivity, and sensation seeking. In one study, hopelessness, impulsivity, and sensation seeking positively correlated with drinking problems (17). Similarly, higher levels of hopelessness, impulsivity, and sensation seeking were related to polysubstance use (18). Contrarily, anxiety sensitivity appears to be inversely related with cannabis use (18) and drinking levels (17). Polysubstance users tend to be generally more impulsive than single substance users. Binge drinkers who met also used cannabis had higher impulsivity scores than those who did not use cannabis (19). Smokers who also reported additional substance use were more impulsive than individuals who only reported smoking (20, 21).

While there is evidence that all these factors play some role in substance dependence, the magnitude and importance of the role these factors play is unclear, as well as the commonality of the risk factors across different substance users. This is especially true when considering the use of all three substances by the same individual, as much of the literature focuses solely on use of a single substance. Similarly, much of the previous literature focuses on a single type of factor (e.g., demographics, personality, or age of onset). Combining these different types of factors into a larger model can both help to parse apart the roles of each type of factor on dependence, as well as the roles of each individual factor. The generalizability of past research is also brought into question because much of the research has been conducted on young and emerging adults, student, or in-patient populations, as opposed to the general population. It is for this reason that we chose to focus our research on a community sample of polysubstance (alcohol, cannabis, and nicotine) users across Canada. Thus, the purpose of this study is to determine the common risk factors for developing substance use within individuals who use alcohol, cannabis, and nicotine. While this research is mainly exploratory, we hypothesized that age of first use, impulsivity, and dependence on other substances will be key factors in predicting substance dependence.

This study aimed to determine risk factors for developing substance dependence among individuals who use alcohol, cannabis, and nicotine. After including a variety of predictor variables, including demographic variables, personality, age of onset for the three substances, and dependence on the other substances, our analyses revealed several risk factors for dependence on each of the substances. Common predictor variables were responsible for large amounts of variance in each model; specifically, impulsivity and dependence on the other substances were key risk factors for elevated levels of dependence on each of alcohol, cannabis, and nicotine. The role of impulsivity in these models confirms prior research in the field as numerous studies highlight the connection between impulsivity and substance dependence (30–34).

Particularly of note was the strong relationship between levels of cannabis and alcohol dependence. In the model for alcohol dependence, level of cannabis dependence alone predicted 23.7% of the variance; while in the model for cannabis dependence, level of alcohol dependence predicted 23.1% of the variance. When using Cohen's (35) definition for effect sizes, these are both substantial effect sizes; as such, it appears that there is a considerable relationship between levels of cannabis and alcohol dependence in our sample. These findings are consistent with evidence of a positive relationship between cannabis dependence with alcohol problems (36, 37), but to our knowledge are the first to demonstrate an association in a community sample of recreational users in a jurisdiction where both substances are legally available. More research should be done to further investigate the link between cannabis and alcohol use and dependence, and to determine the extent to which public health policy contributes to this association.

While we were able to explain 44.9 and 47.6% of the variance in alcohol and cannabis dependence, we were only able to explain 19.9% of the variance in nicotine dependence. This was noteworthy, as prior research has shown a strong connection between nicotine and alcohol use (10, 38, 39), nicotine and cannabis use (40, 41), and nicotine and anxiety sensitivity (42–44). For this study, we included participants who used cigarettes, e-cigarettes, or both cigarettes and e-cigarettes. It is possible that including only one method of nicotine consumption may have led to different results for nicotine dependence, as studies have shown differing levels of dependence between cigarette and e-cigarette users (45–47). However, follow-up analyses were conducted to explore the difference in variance explained in those who used cigarettes vs. those who used e-cigarettes; the variance explained by each model remained fairly stable (within 2% of variance explained). Thus, there must be other key variables in nicotine dependence that we failed to account for in our analyses.

This study is not without limitations, particularly with respect to methodological considerations. First, participants were recruited via Qualtrics panels, an online survey platform. While online panel data has been shown to be an adequate source for exploratory research, questions remain about the generalizability of research conducted via these platforms, particularly with respect to data quality. However, a recent meta-analysis found no significant differences in effect sizes between data collected via online panels and more traditional methods (48). Additionally, several data quality checks (e.g., speeder checks and screening) were in place to help collect the best quality data possible. Second, to meet eligibility criteria, participants were required to use all three substances; as such, we may have decreased the possibility of explaining connections between only two of the substances (e.g., the relationship between nicotine and alcohol). Similarly, while participants were required to use all three substances, we did not ask about or measure other substance use, nor did we measure what their overall substance of choice was. Inclusion of other substances may have led to a more fulsome explanation of the variance in substance dependence levels, whereas substance of choice may have had potential implications for level of dependence on each substance. Additionally, while we included participants who used each substance as little as once per month to as much as daily, it is important to note that these findings may not generalize to groups who use on a less frequent basis. Also, as the study was conducted in a jurisdiction where all three substances are legal, findings may not be generalizable to jurisdictions with different levels of legalization. Finally, as with all cross-sectional studies, the relationship between variables does not infer causation; the direction of effects between the predictors and outcomes in our models is not measured in this study, only the strength of the relationship.

In conclusion, this study demonstrates a strong relationship between alcohol dependence, cannabis dependence, impulsivity, and, to a lesser extent, nicotine dependence in Canadians who use all three substances. Findings suggest that common risk factors may increase vulnerability for substance misuse in general and for dependence on alcohol and cannabis, in particular.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Dalhousie University Health Sciences Research Ethics Board. The patients/participants provided their written informed consent to participate in this study.

AS: conceptualization, methodology, project administration, statistical analyses, writing—original draft, review and editing, and supervision. ŞD: statistical analyses and writing—original draft. SB: conceptualization, methodology, project administration, writing—review and editing, supervision, and funding acquisition. All authors approved the final version of this manuscript.