Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Dec 1.
Published in final edited form as: J Psychiatr Res. 2023 Oct 6;168:38–44. doi: 10.1016/j.jpsychires.2023.10.008

Early Life Adversity and Substance Use: The Mediating Role of Mood and the Moderating Role of Impulsivity

Mustafa al’Absi a, Briana DeAngelis a, Jacob Borodovsky b, Michael J Sofis b, Mark Fiecas c, Alan Budney b
PMCID: PMC10872790  NIHMSID: NIHMS1941176  PMID: 37883864

Abstract

Introduction.

Early life adversity (ELA) is a risk factor for substance use and misuse, and multiple factors mediate and moderate this association. We examined whether moods mediate the relationships between ELA and nicotine use, cannabis use, and co-use, and whether these mediation effects varied as a function of delay discounting.

Methods.

A total of 2,555 adults completed a delay discounting task and responded to questions related to demographics, ELA, mood, and substance use. Data were analyzed using Pearson correlations and moderated mediation using Hayes’ PROCESS macro (V3.4, Model 15).

Results.

ELA was positively associated with cannabis use, nicotine use, co-use of both substances, depressed and stressed moods, and it was negatively associated with positive mood. While cannabis use was associated negatively with stressed and depressed moods and positively with positive mood, nicotine use was associated negatively with positive mood. Moderated mediation analyses indicated that positive mood mediated the relationship between ELA and cannabis use for those with average and above average delay discounting. Positive mood also mediated the relationship between ELA and co-use among those with above average delay discounting.

Conclusion.

The results suggest that ELA’s associations with cannabis use and cannabis-nicotine co-use may be partially attributable to ELA’s effects on positive mood among those who are predisposed to moderately to highly impulsive decision making.

Keywords: Adversity, cannabis, nicotine, stress, delay discounting, mood

1. Introduction

Early life adversity (ELA), such as verbal, physical, or sexual abuse experienced during childhood, predicts onset and patterns of use for multiple addictive substances, including nicotine and cannabis (Anda et al., 1999; Edwards et al., 2007; El Mhamdi et al., 2017; Lemieux et al., 2016; Bellis et al., 2014). Understanding the mechanisms that drive and factors that moderate the relationship between ELA and substance use is critical for developing targeted interventions. Indeed, multiple factors mediate and moderate the connections between early life stress and risk of substance use (al’Absi et al., 2021). Given that cannabis is among the most popular psychoactive drugs in the United States (Substance Abuse and Mental Health Services Administration, 2017), and in light of increasing rates of co-use of tobacco and other nicotine products with cannabis (Badiani et al., 2015; Conway et al., 2018), this study was conducted to examine the extent to which early life adversity predicts use and co-use of these substances while examining potential mediating effects of mood and potential moderating effects of delay discounting.

In addition to predicting substance use, early life adversity is associated with depressive disorders and symptoms (Norman et al., 2012; Zhang et al., 2020), elevated negative mood (Zhang et al., 2020), and lower positive affect (Etter et al., 2013; Oshio et al., 2013) during adulthood. Previous research also documents higher levels of negative moods among nicotine users (Patton et al., 1996) compared to non-users as well as compared to cannabis-only and cannabis-nicotine co-users (Bonn-Miller et al., 2010; Patton et al., 1996). Studies using ecological momentary assessment have also documented an association between elevated positive and negative affect and subsequent cannabis use (e.g., Buckner et al., 2015; Shrier et al., 2014). Together, these studies suggest that mood may be one mechanism through which ELA impacts substance use. To date, one study examined and found evidence of emotion dysregulation as a mediator of the relationship between ELA and substance use (Wolff et al., 2016), and initial studies indicate that both depression and anxiety may mediate the relationship between early life adversity and substance use (He et al., 2022; Kim et al, 2021); however, no studies have examined positive moods as potential mediators of the relationship between ELA and substance use. Therefore, one aim of this study was to examine the mediating roles of positive, stressed, and depressed moods in the relationship between ELA and substance use.

A second aim of this study was to examine whether delay discounting moderates the aforementioned relationships. Delay discounting refers to the tendency to discount the value of a reward as the time to receiving the reward increases (Kirby & Maraković, 1996). Thus, when presented with a choice between a smaller, more immediate reward and a larger, delayed reward, individuals who tend to have steeper delay discounting are more likely to select the more immediate reward; and this tendency is thought to reflect impulsivity (Kirby & Maraković, 1996; Koffarnus & Bickel, 2014; Moreira & Barbosa, 2019). While some studies have indicated an association between early life adversity and impulsivity (Andrews et al., 2021; Narvaez et al., 2012), the focus here is on impulsivity as a moderator of the relationships between moods and ELA with substance use. A recent meta-analysis conducted on 27 studies showed that greater cannabis use was associated with greater delay discounting (Strickland et al., 2021). Impulsive tendencies, including discounting future rewards, are associated with propensity to initiate and maintain substance use (Amlung et al., 2017; Białaszek et al., 2017). Furthermore, such tendencies have been demonstrated to influence the relationship between affect and substance use. For instance, one study found that positive affect was associated with alcohol use, but this effect was only observed among individuals who were relatively impulsive (Colder & Chassin, 1997). Another study found that the relationship between depressive symptoms and substance use was stronger among those who were more impulsive (Felton et al., 2020). Given these studies, it is possible that moods mediate the relationship between ELA and substance use, but only among those who tend to be impulsive.

In sum, this study aimed to better understand the mechanisms that facilitate and factors that moderate the relationship between ELA and substance use by testing moderated mediation models in which moods (positive, stressed, and depressed) mediate the effects of ELA on nicotine and cannabis use (as well as co-use), with delay discounting as a moderator of these associations (see Figure 1). We predicted that high early life adversity would be associated with increased risk for substance use due to elevated negative moods (stress and depressed mood) and higher (for cannabis use) or lower (for nicotine use) positive mood; and we expected these relationships to be strongest among those who tend to select sooner rewards (i.e., exhibit steeper delay discounting).

Figure 1.

Figure 1.

Open in a new tab

The predicted moderated mediation models in which positive, stressed, and depressed moods mediate the relationship between early life adversity (ELA) and substance use (nicotine-only use, cannabis-only use, or nicotine-cannabis co-use) and in which delay discounting moderates this mediation through its interaction with positive, stressed, and depressed moods as well as its interaction with ELA.

2. Methods

2.1. Participants and recruitment

Two cross-sectional, online surveys were created to assess ELA, cannabis and nicotine use, mood, delay discounting, and demographics, among other variables. Following previously published guidance about using Facebook for research (Borodovsky et al., 2018), survey URLs were distributed using targeted Facebook advertisements that were run in late November and early December of 2018 (Survey #1) and July through September of 2019 (Survey #2). Ads targeted residents of the United States who were aged 18+ (key words, such as marijuana and cigarette, were used to narrow the target audience) and used messages such as “Share your experience with life stress, pain, tobacco, and marijuana in an online research survey!” and “Adults wanted for an online research survey on life stress, marijuana, and tobacco! To participate, click below or copy & paste this link!” The survey landing page contained a consent form approved by the University of Minnesota Institutional Review Board; participation in the surveys was completely anonymous and no compensation was offered. The ads reached 220,638 Facebook users. Of the 9,118 Facebook users who clicked on the advertised survey URLs, 2,555 (~28%) consented and confirmed that they were residents of the USA and 18 years or older. The median amount of time that respondents spent completing a survey was 13 minutes (interquartile range, IQR, = 9).

To minimize unintentional non-responses, we used the “Forced Response” (for items with “I prefer not to answer” as a response option) and “Request Response” (for items without “I prefer not to answer” response options) features in Qualtrics © (Provo, UT). Following published guidance for conducting on-line survey research (Borodovsky et al., 2018), we implemented a captcha verification to prevent responses from internet bots.

2.2. Measures

2.2.1. Demographics

In both surveys, participants responded to questions regarding their age, sex, ethnicity, residence (state and urban classification), education, and employment status.

2.2.2. Early Life Adversity

A 10-item Adverse Childhood Experiences questionnaire with dichotomous response options (Dube et al., 2002; Felitti et al., 1998; Sciolla et al., 2019) was used in both surveys to assess 10 domains of ELA: emotional abuse, parental separation, emotional neglect, substance abuse in household, mental illness in household, physical abuse, sexual abuse, mother treated violently, physical neglect, and incarceration of household member. Responses to the items were summed to create a total score, which served as the primary predictor in our models.

2.2.3. Mood

In both surveys, participants reported how often they felt happy or cheerful, upbeat or chipper, and stressed or overwhelmed over the last 2 weeks using response scales that ranged from 0 (not at all) to 3 (nearly every day). These items were written for this project, but similar items have been used in studies that demonstrate expected changes in these items in response to acute stress (e.g., DeAngelis & al’Absi, 2020). Responses to the first 2 items (Pearson r = .79) were averaged to create an index of positive mood and the third item was used to measure stressed mood. Participants also completed the 9-item Patient Health Questionnaire (PHQ-9; Kroenke et al., 2001), which assesses depressive affect over the last 2 weeks (Cronbach’s α = .89); scores can range from 0–27.

2.2.4. Delay Discounting

A 5-trial adjusting delay discounting task (Koffarnus & Bickel, 2014), with $1000 specified as the larger later amount, was used to assess discounting rates (k) in both surveys. In this task, respondents are presented with a series of 5 choices in which they must select between a larger amount of money ($1000) after a specified time delay and a smaller amount of money ($500) after no time delay. Depending on the respondent’s choice, the task adjusts (up or down) the time delay paired with the larger reward in the subsequent trial presented to the respondent. A discounting rate for each participant, k, was calculated as follows: If V represents the value of a given reward, then V = A/(1+kD), where A is the magnitude of the reward, k is the discounting rate, and D is the delay. As can be noted, the larger the k, the larger the impact of delay (D) in reducing the value of the reward.

2.2.5. Substance Use

Respondents were classified as regular nicotine users if they self-identified as daily or almost daily nicotine users (Survey #1) or if they indicated using nicotine “18–22 days (~5 days per week)” or more during the last 30 days (Survey #2). The cut-off used for Survey #2 was selected to reflect a cut-point comparable to that which was available in Survey #1, which identified nicotine users as those who identified as “daily or almost daily” nicotine users. Respondents reported their age when they first tried nicotine, how soon they use nicotine after they first wake up, and the types of tobacco or nicotine that they use. Respondents were classified as regular cannabis users if they indicated using cannabis on 10+ of the last 30 days (al’Absi et al., 2022). Cannabis users also reported their typical frequency of use on days that they use (11 response options: 1, 2…10, more than 10); and they reported how intoxicated they get when they use (1 = light buzz to 10 = so high that you vomit/throw up) (Borodovsky et al., 2020).

2.3. Analytic Approach

Prior to analysis, we created three dichotomous substance use variables that served as our primary dependent variables wherein regular users of the substance(s) were coded as 1 and others were coded as 0: (a) nicotine-only use during the last 30 days (vs. those who did not regularly use nicotine nor cannabis), (b) cannabis-only use during the last 30 days, (vs. those who did not regularly use nicotine nor cannabis), and (c) co-use during the last 30 days (vs. those who did not regularly use nicotine nor cannabis). This dichotomization was done to make the measures on nicotine-use comparable across surveys, and to ensure that results across substances were comparable. The distribution of discounting rates, k, deviated substantially from normal (skew = 8.83, SEskew = .05; kurtosis = 79.32, SEkurtosis = .10), so discounting rates were natural log transformed to approximate a normal distribution prior to analyses.

Demographics and descriptive information related to cannabis and nicotine use were summarized and examined using means, medians, standard deviations, counts, and percentages. Relationships among the primary predictor (ELA), mediators (positive, stressed, depressed moods), moderator, and dependent variables (dichotomous substance use variables) were examined using Pearson correlations before testing the proposed moderated mediation models. Note that the frequencies reported for some subject characteristics occasionally sum to less than the full sample due to occasional respondents who chose not to answer certain items.

Three separate moderated mediation models were specified and tested (one for each primary dependent variable) using the PROCESS macro (V3.4, Model 15; Hayes, 2017) within SPSS (V24), which implements a logistic regression-based approach (with listwise deletion) for dichotomous outcomes. Positive mood, stressed mood, and depressed mood were specified as mediators; delay discounting was specified as a moderator (see Figures 1–2). We requested 5000 bootstrap samples, Davidson and Mackinnon’s (1993) approximated HC3 estimator, and 95% confidence intervals. Interactions with p-values < .05 were probed at the mean and +/− 1 SD from the mean; and inferences regarding indirect effects were informed by bootstrap confidence intervals. All independent variables were mean centered for relevant cases prior to model testing.

Figure 2.

Figure 2.

Open in a new tab

This figure depicts the coefficients estimated in the moderated mediation models that were tested and that are presented in Table 3.

Note: ELA = early life adversity; Discounting = delay discounting; Substance use = either nicotine-only use, cannabis-only use, or nicotine-cannabis co-use

3. Results

3.1. Subject Characteristics

Table 1 summarizes demographics and descriptive measures for those who reported information regarding their substance use. The sample included 940 males and 1605 females who were predominantly White (92.2%), with 3.8% identifying as Hispanic. Respondents ranged from 18 to 94 years of age (M = 40.7, SD = 15.7). Respondents were residents in each of the 50 United States and in the District of Columbia, with 69.6% residing in urban areas or urban clusters. The sample was relatively well educated, with 63.1% reporting education beyond high school. Approximately 48.5% worked full- or part-time jobs, 12.3% were unemployed, 8.2% were retired, 7.7% were students, and 22.3% were on disability.

Table 1.

Demographics and descriptive measures by substance use during the last 30 days

Not Regular Users of Nicotine Nor Cannabis Nicotine-Only Users Cannabis-Only Users Co-Users of Cannabis & Nicotine
n = 380 n = 498 n = 585 n = 1027
Age in years (M, SD) 42.8 (17.0) 42.2 (15.2) 41.6 (16.1) 38.8 (14.9)
Male (n, %) 108 (28.4%) 167 (33.5%) 215 (36.8%) 414 (40.3%)
Caucasian (n, %) 347 (91.3%) 462 (92.8%) 529 (90.4%) 961 (93.6%)
Education beyond H.S. (n, %) 286 (75.3%) 321 (64.5%) 409 (69.9%) 571 (55.6%)
Work full- or part-time job (n, %) 187 (49.2%) 253 (50.8%) 261 (44.6%) 507 (49.4%)
Reside in urban areaa (n, %) 114 (30.0%) 145 (29.1%) 175 (29.9%) 298 (29.0%)
Reside in urban clusterb (n, %) 166 (43.7%) 201 (40.4%) 231 (39.5%) 412 (40.1%)
Reside in rural areac (n, %) 91 (23.9%) 147 (29.5%) 174 (29.7%) 311 (30.3%)
Cannabis, # times/use day (Mdn, IQR) - - 3.00 (3.00) 3.00 (3.00)
Cannabis intoxication (M, SD) - - 4.74 (2.00) 5.04 (1.97)
Nicotine use before 19 yrs old (n, %) - 429 (86.1%) - 918 (89.4%)
Nicotine in 1st hr of waking (n, %) - 458 (92.0%) - 928 (90.4%)
ELA (M, SD) 3.14 (2.64) 3.58 (2.67) 3.52 (2.66) 3.84 (2.57)
Positive mood (M, SD) 1.38 (0.89) 1.18 (0.83) 1.69 (0.94) 1.43 (0.88)
Stressed mood (M, SD) 1.75 (0.99) 1.85 (0.96) 1.47 (1.00) 1.73 (1.02)
Depressed mood (M, SD) 9.69 (6.65) 10.56 (6.61) 7.93 (6.44) 9.39 (6.89)
Raw delay discounting rate (Mdn, IQR) .007 (0.01) .009 (0.04) .007 (0.02) .013 (0.03)
Open in a new tab

Note: Percentages represent percent of total n for the substance use category (i.e., percent of 380, 498, 585, 1027, respectively). M = mean; Mdn = median; SD = standard deviation; IQR = interquartile range (75%-25%); H.S. = high school; ELA = early life adversity; yrs = years; hr = hour; Supplementary Table 3 presents a summary of statistically significant differences for demographics and descriptive variables presented in Table 1.

a

Urban area (population of 50,000+)

b

Urban cluster (population between 2,500 to 50,000)

c

Rural area (population less than 2,500)

Overall, 1619 respondents reported regular cannabis use (regardless of nicotine use), among whom the median frequency of use was 3 times per day (IQR = 3.00, M = 4.22, SD = 2.77); and they reported typically achieving a median intoxication level of 5 (M = 4.92, SD = 1.98). A total of 1561 respondents reported regular nicotine use (regardless of cannabis use), most of whom (n = 1356, 90.3% of responses) tried nicotine prior to age 19, and a majority of whom reported using nicotine within the first hour of waking (n = 1396, 90.9% of responses). Among these regular nicotine users, 1554 reported the types of nicotine that they typically use; cigarettes were the most frequently reported method of nicotine use (n = 1263, 81.3%), followed by e-cigarettes (n = 467, 30.1%), chewing tobacco (n = 111, 7.1%), cigars (n = 87, 5.6%), cigar-cigarettes (n = 80, 5.1%), and hookah or shisha (n = 28, 1.8%).

3.2. ELA, Delay Discounting, and Mood

Overall, 1,952 respondents reported experiencing one or more ELAs and 306 respondents reported experiencing no ELA (M = 3.6, SD = 2.6). The sample reported moderately positive moods (M = 1.44, SD = 0.90, n = 2425); moderately stressed moods (M = 1.70, SD = 1.01, n = 2425); and mildly depressed moods (M = 9.34, SD = 6.75, n = 2366) during the last two weeks. Correlations among the primary predictor, mediators, moderator, and dependent variables are depicted in Table 2. As shown, ELA was associated positively with depressed and stressed moods, but it was associated negatively with positive mood. ELA was also weakly associated with greater delay discounting (k), as well as with cannabis use, nicotine use, and co-use of both substances.

Table 2.

Pearson Correlations (n)

ELA Positive Mood Stressed Mood Depressed Mood Delay Discountinga
ELA
Positive Mood −.16** (2257)
Stressed Mood .22** (2257) −.45** (2424)
Depressed Mood .26** (2247) −.54** (2365) .66** (2365)
Delay Discounting a .08** (2238) −.15** (2279) .09** (2279) .14** (2267)
CB .07* (878) .16** (935) −.14** (934) −.13** (910) .08* (882)
NIC .08* (805) −.11** (854) .05 (853) .07 (838) .15** (812)
COUSE .12** (1265) .03 (1361) −.01 (1361) −.02 (1333) .15** (1279)
Open in a new tab

Note: ELA = early life adversity; CB = cannabis-only user (coded 1) vs. not regular users of nicotine nor cannabis (coded 0); NIC = nicotine-only user (coded 1) vs. not regular users of nicotine nor cannabis (coded 0); COUSE = regular user of both cannabis and nicotine (coded 1) vs. not regular users of nicotine nor cannabis (coded 0)

a

natural log transformed delay discounting rate

*

= significant at p < .05;

**

= significant at p < .01

3.3. Moderated Mediation Models

Model coefficients are presented in Figure 2 and Table 3, model summary results are presented in Supplementary Table 1, and indices of moderated mediation are presented in Table 4. Consistent with our predictions, ELA was a significant predictor of cannabis-only as well as co-use (Table 3, c3’). However, ELA was not a significant predictor of nicotine-only use (Table 3, c3’). In fact, delay discounting was the only significant predictor of nicotine-only use (Table 3, c1’). As can be seen in Table 3, the coefficients linking ELA with the three mediators: positive mood, stressed mood, and depressed mood (a1, a2, a3) were significant and in the directions that were predicted for all three models. Positive mood was the only mediator whose direct coefficient (Table 3, b1) was significant in predicting substance use (cannabis use), though there was evidence of an interaction between positive mood and delay discounting (Table 3, b4) in the models for cannabis use as well as co-use. Probing this interaction indicated that positive mood was associated with increased risk of using cannabis-only and of co-using nicotine and cannabis; however, this increased risk was only significant for individuals with moderate (cannabis-only use) to high (cannabis-only and nicotine-cannabis co-use) delay discounting (Supplementary Table 2). Moreover, as indicated by the indices of moderated mediation (Table 4; Hayes, 2015), there was evidence of moderated mediation, through positive mood, in both of these models. In the model predicting cannabis-only use, the conditional indirect effect of ELA through positive mood was consistently negative, though it was only significant at the mean (−.018, BootSE = .007, OR = .98) and at 1 SD above the mean (−.032, BootSE = .012, OR = .97) of delay discounting, indicating that for those with average or above average discounting rates, higher ELA was associated with lower likelihood of cannabis use due to ELA’s negative relationship with positive mood. In the model predicting co-use, the conditional indirect effect of ELA through positive mood was significant only at 1 SD above the mean of delay discounting (−.019, BootSE = .009, OR = .98), indicating that for those with above average discounting, higher ELA was associated with decreased likelihood of co-use due to ELA’s negative relationship with positive mood. There was no evidence of conditional direct nor conditional indirect effects in the model predicting nicotine-only use.

Table 3.

Model coefficients (and standard errors)

Relationship (Coefficient) Nicotine-Only Users Cannabis-Only Users Co-Users of Cannabis & Nicotine
ELA -> Positive Mood (a1) −.079*** (.011) −.059*** (.012) −.052*** (.010)
ELA -> Stressed Mood (a2) .093*** (.013) .104*** (.012) .086*** (.011)
ELA -> Depressed Mood (a3) .781*** (.088) .738*** (.086) .713*** (.071)
Positive Mood -> Substance Use (b1) −.199 (.105) .305** (.100) .134 (.091)
Stressed Mood -> Substance Use (b2) −.020 (.100) −.136 (.101) −.009 (.088)
Depressed Mood -> Substance Use (b3) −.007 (.016) −.023 (.016) −.017 (.014)
Positive Mood*Discounting -> Substance Use (b4) .049 (.047) .112* (.046) .100* (.040)
Stressed Mood*Discounting -> Substance Use (b5) .062 (.045) .043 (.048) .066 (.041)
Depressed Mood*Discounting -> Substance Use (b6) .006 (.007) .010 (.008) .004 (.007)
Discounting -> Substance Use (c1’) .126*** (.034) .107** (.035) .164*** (.032)
ELA*Discounting -> Substance Use (c2’) −.005 (.013) .001 (.013) −.016 (.013)
ELA -> Substance Use (c3’) .053 (.030) .097*** (.029) .122*** (.027)
Constant (in models of substance use) .277*** (.075) .453*** (.074) 1.043*** (.068)
Open in a new tab

Note: ELA = Early Life Adversity. Discounting = Natural log of delay discounting. Model coefficients displayed in this table correspond to the relationships labeled in Figure 2. All coefficients labeled as b# or c#’ are in log-odds metric and can be exponentiated to obtain odds ratios.

***

indicates p <.001;

**

indicates p <.01;

*

indicates p <.05

Table 4.

Indices of Moderated Mediation (bootstrap standard error)

Mediator Nicotine-Only Users Cannabis-Only Users Co-Users of Cannabis & Nicotine
Positive Mood −.004 (.004) −.007* (.003) −.005* (.003)
Stressed Mood .006 (.005) .004 (.006) .006 (.004)
Depressed Mood .005 (.007) .007 (.008) .003 (.006)
Open in a new tab
*

95% bootstrap confidence interval does not include 0.

Running these moderated mediation models with sex (1 = male) as a covariate had minimal impact on the significant effects reported above. The only relationship that changed in significance was in the model predicting nicotine-only use, wherein the addition of sex as a covariate changed the significance of ELA in predicting nicotine-only use so that it became statistically significant (whereas it was not statistically significant in the model that did not include sex as a covariate). Moreover, the moderated mediation models had no changes in significance nor direction of effects when the models were run on a subset of the sample that excluded respondents who took less than 5 minutes or more than 3 hours between starting and ending the survey.

4. Discussion

Consistent with previous research (Anda et al., 1999; Etter et al., 2013; Zhang et al., 2020), findings from this study demonstrated that ELA was negatively associated with positive mood and it was positively associated with depressed and stressed moods as well as with nicotine-only, cannabis-only, and nicotine-cannabis co-use. Results also showed that nicotine-only use was negatively related to positive mood; and cannabis-only use was associated positively with positive mood and negatively with stressed and depressed moods. Results from the moderated mediation analyses showed that ELA was a significant predictor of cannabis-only and nicotine-cannabis co-use; however, ELA was not a significant predictor of nicotine-only use. Moreover, for those with average or above average discounting rates, positive mood was a significant mediator of the relationship between ELA and cannabis use, though this effect was not sizable. Positive mood also mediated the relationship between ELA and cannabis-nicotine co-use, though only for those with above average delay discounting.

Consistent with our predictions, ELA and delay discounting were associated with increased risk of substance use. In line with previous research, we found that positive mood was a significant predictor of cannabis use; moreover, we found that delay discounting moderated the indirect effect of ELA on substance use through positive mood. Specifically, our moderated mediation results suggest that even though the direct effect of ELA is positive, for individuals who tend to be relatively impulsive, experiencing greater ELA is indirectly associated with slightly reduced risk of substance use due to adversity’s negative relationship with positive mood. Though we didn’t find evidence that delay discounting moderated the direct effect of ELA on substance use, our finding that discounting moderated the indirect effect suggests that individual difference factors, such as impulsivity, may be important to consider when examining mechanisms that account for the positive association between ELA and substance use and when designing potential interventions that aim to reduce substance use. Our findings are consistent with previous studies that found mood predicts substance use, but only among those who tend to be more impulsive (Colder & Chassin, 1997; Felton et al., 2020). In combination with those studies, our findings suggest that decisions made by individuals who tend to be relatively impulsive may be more subject to emotion-related motives than decisions made by those who tend to be less impulsive.

Mechanisms for these connections likely involve specific central processes and substrates that are supported by previous clinical and preclinical research. For example, neuroimaging research indicates that high delay discounting is related to attenuated activation in the dorsolateral prefrontal cortex (DLPFC; Steinbeis et al., 2012), an area of the prefrontal cortex that is involved in executive functions, including working memory, decision making, and cognitive flexibility (Friedman & Robbins, 2022). It is possible that DLPFC functions explain the moderating effects of delay discounting on risk for substance use. This is further supported by the growing literature demonstrating the role of delay discounting as a risk factor for addictive behaviors (Amlung et al., 2017), including smoking (Audrain-McGovern et al., 2009; Baker et al., 2003; Fields et al., 2009; Johnson et al., 2007). Thus, interventions aimed at reducing impulsivity (including delay discounting) or at enhancing emotion regulation may help reduce risk of substance use. We do note, however, links of our results with these potential mechanisms are speculative at this time.

Stressed and depressed moods did not emerge as significant mediators of the ELA-substance use relationships in our study, suggesting that, when examined within the context of early life adversity, cannabis use and nicotine-cannabis co-use may be “positive mood risk behaviors” in that they may be more likely to occur following or during positive moods. If this is the case, then individuals who are relatively high in positive urgency (i.e., those who have a tendency to act impulsively when feeling positive) (Cyders et al., 2007) may be particularly at risk of using these substances. Future longitudinal research should examine this hypothesis further.

It is worth noting that in this study, social media was used to recruit a national sample that included a large number of cannabis and nicotine users. Consistent with past studies (Borodovsky et al., 2018; Thornton et al., 2016), Facebook was an economic, streamlined platform from which we were able to gain relevant participants from relatively evasive populations. This study used an uncompensated survey and key words to target Facebook advertisements, respondents may not be representative of average cannabis users, limiting the generalizability of the results. Nevertheless, non-probability samples such as those targeted by this internet-based research, can be useful for understanding correlates of substance use, including cannabis, while mitigating challenges associated with probability sampling (Borodovsky, 2022; Borodovsky et al., 2018). The cross-sectional nature of the design also precludes definitive conclusions regarding the causal order of variables in the models; future longitudinal studies are needed. This study did not find evidence of moods as mediators of ELA’s relationship with nicotine-only use, the associations involving delay discounting and ELA had small effect sizes, and ELA continued to have a direct effect on substance use after accounting for its indirect effects in the models predicting cannabis-only and nicotine-cannabis co-use. Therefore, future longitudinal research should continue to explore other potential mediators of the ELA-substance use relationship using more diverse and comprehensive assessment approaches. Future research might also examine mediators of substance use as continuous (rather than dichotomous) outcomes to better understand mediators and risk factors for particularly problematic or heavy substance use.

In conclusion, this study found that positive mood was a significant mediator of the relationships between ELA and cannabis-only use as well as co-use of nicotine and cannabis, but only among those who exhibited relatively moderate to high delay discounting. Our findings support previous research suggesting that bolstering emotion regulation or reducing delay discounting may help reduce risk of substance use.

Supplementary Material

1

Role of Funding Sources

This research was supported, in part, by National Institutes of Health (NIH) grants R01DA016351 and R01DA027232. NIH had no role in: designing the study; collecting, analyzing, and interpreting the data; writing and submitting this manuscript.

Footnotes

Conflict of Interest

None.

Author Agreement

The manuscript that is being submitted has been seen and approved by all authors. This submission is our original work; it hasn’t been published previously; and it isn’t submitted for publication elsewhere.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. al’Absi M, Ginty AT, & Lovallo WR (2021). Neurobiological mechanisms of early life adversity, blunted stress reactivity and risk for addiction. Neuropharmacology, 188, 108519. 10.1016/j.neuropharm.2021.108519 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. al’Absi M, DeAngelis B, Fiecas M, Budney A, & Allen S (2022). Effects of regular cannabis and nicotine use on acute stress responses: Chronic nicotine, but not cannabis use, is associated with blunted adrenocortical and cardiovascular responses to stress. Psychopharmacology, 239(5), 1551–1561. 10.1007/s00213-022-06087-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Amlung M, Vedelago L, Acker J, Balodis I, & MacKillop J (2017). Steep delay discounting and addictive behavior: A meta-analysis of continuous associations. Addiction, 112(1), 51–62. 10.1111/add.13535 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Anda RF, Croft JB, Felitti VJ, Nordenberg D, Giles WH, Williamson DF, & Giovino GA (1999). Adverse childhood experiences and smoking during adolescence and adulthood. JAMA, 282(17), 1652–1658. 10.1001/jama.282.17.1652 [DOI] [PubMed] [Google Scholar]
  5. Andrews K, Atkinson L, Harris M, & Gonzalez A (2021). Examining the effects of household chaos on child executive functions: A meta-analysis. Psychological Bulletin, 147(1), 16–32. 10.1037/bul0000311 [DOI] [PubMed] [Google Scholar]
  6. Audrain-McGovern J, Rodriguez D, Epstein LH, Cuevas J, Rodgers K, & Wileyto EP (2009). Does delay discounting play an etiological role in smoking or is it a consequence of smoking? Drug and Alcohol Dependence, 103(3), 99–106. 10.1016/j.drugalcdep.2008.12.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Badiani A, Boden JM, De Pirro S, Fergusson DM, Horwood LJ, & Harold GT (2015). Tobacco smoking and cannabis use in a longitudinal birth cohort: Evidence of reciprocal causal relationships. Drug and Alcohol Dependence, 150, 69–76. 10.1016/j.drugalcdep.2015.02.015 [DOI] [PubMed] [Google Scholar]
  8. Baker F, Johnson MW, & Bickel WK (2003). Delay discounting in current and never-before cigarette smokers: Similarities and differences across commodity, sign, and magnitude. Journal of Abnormal Psychology, 112(3), 382–392. 10.1037/0021-843x.112.3.382 [DOI] [PubMed] [Google Scholar]
  9. Bellis MA, Hughes K, Leckenby N, Jones L, Baban A, Kachaeva M, Povilaitis R, Pudule I, Qirjako G, Ulukol B, Raleva M, & Terzic N (2014). Adverse childhood experiences and associations with health-harming behaviours in young adults: Surveys in eight eastern European countries. Bulletin of the World Health Organization, 92(9), 641–655. 10.2471/BLT.13.129247 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Białaszek W, Marcowski P, & Cox DJ (2017). Differences in delay, but not probability discounting, in current smokers, e-cigarette users, and never smokers. Psychological Record, 67(2), 223–230. 10.1007/s40732-017-0244-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Bonn-Miller MO, Zvolensky MJ, & Johnson KA (2010). Uni-morbid and co-occurring marijuana and tobacco use: Examination of concurrent associations with negative mood states. Journal of Addictive Diseases, 29(1), 68–77. 10.1080/10550880903435996 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Borodovsky JT (2022). Generalizability and representativeness: Considerations for internet-based research on substance use behaviors. Experimental and Clinical Psychopharmacology, 30(4), 466–477. 10.1037/pha0000581 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Borodovsky JT, Marsch LA, & Budney AJ (2018). Studying cannabis use behaviors with Facebook and web surveys: Methods and insights. JMIR Public Health and Surveillance, 4(2). 10.2196/publichealth.9408 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Borodovsky JT, Marsch LA, Scherer EA, Grucza RA, Hasin DS, & Budney AJ (2020). Perceived safety of cannabis intoxication predicts frequency of driving while intoxicated. Preventive Medicine, 131, 1–7. 10.1016/j.ypmed.2019.105956 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Buckner JD, Zvolensky MJ, Crosby RD, Wonderlich SA, Ecker AH, & Richter A (2015). Antecedents and consequences of cannabis use among racially diverse cannabis users: An analysis from ecological momentary assessment. Drug and Alcohol Dependence, 147, 20–25. 10.1016/j.drugalcdep.2014.12.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Colder CR, & Chassin L (1997). Affectivity and impulsivity: Temperament risk for adolescent alcohol involvement. Psychology of Addictive Behaviors, 11(2), 83–97. 10.1037/0893-164X.11.2.83 [DOI] [Google Scholar]
  17. Conway KP, Green VR, Kasza KA, Silveira ML, Borek N, Kimmel HL, Sargent JD, Stanton CA, Lambert E, Hilmi N, Reissig CJ, Jackson KJ, Tanski SE, Maklan D, Hyland AJ, & Compton WM (2018). Co-occurrence of tobacco product use, substance use, and mental health problems among youth: Findings from wave 1 (2013–2014) of the population assessment of tobacco and health (PATH) study. Addictive Behaviors, 76, 208–217. 10.1016/j.addbeh.2017.08.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Cyders MA, Smith GT, Spillane NS, Fischer S, Annus AM, & Peterson C (2007). Integration of impulsivity and positive mood to predict risky behavior: Development and validation of a measure of positive urgency. Psychological Assessment, 19(1), 107–118. 10.1037/1040-3590.19.1.107 [DOI] [PubMed] [Google Scholar]
  19. Davidson R, & MacKinnon JG (1993). Estimation and inference in econometrics. Oxford University Press. [Google Scholar]
  20. DeAngelis BN, & al’Absi M (2020). Regular cannabis use is associated with blunted affective, but not cardiovascular, stress responses. Addictive Behaviors, 107, 106411. 10.1016/j.addbeh.2020.106411 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Dube SR, Anda RF, Felitti VJ, Edwards VJ, & Williamson DF (2002). Exposure to abuse, neglect, and household dysfunction among adults who witnessed intimate partner violence as children: Implications for health and social services. Violence and Victims, 17(1), 3–17. 10.1891/vivi.17.1.3.33635 [DOI] [PubMed] [Google Scholar]
  22. Edwards VJ, Anda RF, Gu D, Dube SR, & Felitti VJ (2007). Adverse childhood experiences and smoking persistence in adults with smoking-related symptoms and illness. Permanente Journal, 11(2), 5–13. 10.7812/TPP/06-110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. El Mhamdi S, Lemieux A, Bouanene I, Ben Salah A, Nakajima M, Ben Salem K, & al’Absi M (2017). Gender differences in adverse childhood experiences, collective violence, and the risk for addictive behaviors among university students in Tunisia. Preventive Medicine, 99, 99–104. 10.1016/j.ypmed.2017.02.011 [DOI] [PubMed] [Google Scholar]
  24. Etter DW, Gauthier JR, McDade-Montez E, Cloitre M, & Carlson EB (2013). Positive affect, childhood adversity, and psychopathology in psychiatric inpatients. European Journal of Psychotraumatology, 4(1), 20771. 10.3402/ejpt.v4i0.20771 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards V, Koss MP, & Marks JS (1998). Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. American Journal of Preventive Medicine, 14(4), 245–258. 10.1016/S0749-3797(98)00017-8 [DOI] [PubMed] [Google Scholar]
  26. Felton JW, Shadur JM, Havewala M, Gonçalves S, & Lejuez CW (2020). Impulsivity moderates the relation between depressive symptoms and substance use across adolescence. Journal of Clinical Child & Adolescent Psychology, 49(3), 365–377. 10.1080/15374416.2018.1537189 [DOI] [PubMed] [Google Scholar]
  27. Fields S, Leraas K, Collins C, & Reynolds B (2009). Delay discounting as a mediator of the relationship between perceived stress and cigarette smoking status in adolescents. Behavioural Pharmacology, 20(5–6), 455–460. 10.1097/FBP.0b013e328330dcff [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Friedman NP, & Robbins TW (2022). The role of prefrontal cortex in cognitive control and executive function. Neuropsychopharmacology, 47(1), Article 1. 10.1038/s41386-021-01132-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Hayes AF (2015). An index and test of linear moderated mediation. Multivariate Behavioral Research, 50(1), 1–22. 10.1080/00273171.2014.962683 [DOI] [PubMed] [Google Scholar]
  30. Hayes AF (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Publications. [Google Scholar]
  31. He J, Yan X, Wang R, Zhao J, Liu J, Zhou C, & Zeng Y (2022). Does childhood adversity lead to drug addiction in adulthood? A study of serial mediators based on resilience and depression. Frontiers in Psychiatry, 13. 10.3389/fpsyt.2022.871459 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Johnson MW, Bickel WK, & Baker F (2007). Moderate drug use and delay discounting: A comparison of heavy, light, and never smokers. Experimental and Clinical Psychopharmacology, 15(2), 187–194. 10.1037/1064-1297.15.2.187 [DOI] [PubMed] [Google Scholar]
  33. Kirby KN, & Maraković NN (1996). Delay-discounting probabilistic rewards: Rates decrease as amounts increase. Psychonomic Bulletin & Review, 3(1), 100–104. 10.3758/BF03210748 [DOI] [PubMed] [Google Scholar]
  34. Kim Y, Kim K, Chartier KG, Wike TL, & McDonald SE (2021). Adverse childhood experience patterns, major depressive disorder, and substance use disorder in older adults. Aging & Mental Health, 25(3), 484–491. 10.1080/13607863.2019.1693974 [DOI] [PubMed] [Google Scholar]
  35. Koffarnus MN, & Bickel WK (2014). A 5-trial adjusting delay discounting task: Accurate discount rates in less than one minute. Experimental and Clinical Psychopharmacology, 22(3), 222–228. 10.1037/a0035973 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kroenke K, Spitzer RL, & Williams JBW (2001). The PHQ-9: Validity of a brief depression severity measure. Journal of General Internal Medicine, 16(9), 606–613. 10.1046/j.1525-1497.2001.016009606.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Lemieux A, Olson L, Nakajima M, Schulberg L, & al’Absi M (2016). Life adversity is associated with smoking relapse after a quit attempt. Addictive Behaviors, 60, 71–77. 10.1016/j.addbeh.2016.03.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Moreira D, & Barbosa F (2019). Delay discounting in impulsive behavior: A systematic review. European Psychologist, 24(4), 312–321. 10.1027/1016-9040/a000360 [DOI] [Google Scholar]
  39. Narvaez JCM, Magalhães PVS, Trindade EK, Vieira DC, Kauer-Sant’Anna M, Gama CS, von Diemen L, Kapczinski NS, & Kapczinski F (2012). Childhood trauma, impulsivity, and executive functioning in crack cocaine users. Comprehensive Psychiatry, 53(3), 238–244. 10.1016/j.comppsych.2011.04.058 [DOI] [PubMed] [Google Scholar]
  40. Norman RE, Byambaa M, De R, Butchart A, Scott J, & Vos T (2012). The long-term health consequences of child physical abuse, emotional abuse, and neglect: A systematic review and meta-analysis. PLOS Medicine, 9(11), e1001349. 10.1371/journal.pmed.1001349 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Oshio T, Umeda M, & Kawakami N (2013). Childhood adversity and adulthood subjective well-being: Evidence from Japan. Journal of Happiness Studies, 14(3), 843–860. 10.1007/s10902-012-9358-y [DOI] [Google Scholar]
  42. Patton GC, Hibbert M, Rosier MJ, Carlin JB, Caust J, & Bowes G (1996). Is smoking associated with depression and anxiety in teenagers? American Journal of Public Health, 86(2), 225–230. 10.2105/AJPH.86.2.225 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sciolla AF, Wilkes MS, & Griffin EJ (2019). Adverse Childhood Experiences in medical students: Implications for wellness. Academic Psychiatry, 43(4), 369–374. 10.1007/s40596-019-01047-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Shrier LA, Ross CS, & Blood EA (2014). Momentary positive and negative affect preceding marijuana use events in youth. Journal of Studies on Alcohol and Drugs, 75(5), 781–789. 10.15288/jsad.2014.75.781 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Steinbeis N, Bernhardt BC, & Singer T (2012). Impulse control and underlying functions of the left DLPFC mediate age-related and age-independent individual differences in strategic social behavior. Neuron, 73(5), 1040–1051. 10.1016/j.neuron.2011.12.027 [DOI] [PubMed] [Google Scholar]
  46. Strickland JC, Lee DC, Vandrey R, & Johnson MW (2021). A systematic review and meta-analysis of delay discounting and cannabis use. Experimental and Clinical Psychopharmacology, 29(6), 696–710. 10.1037/pha0000378 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Substance Abuse and Mental Health Services Administration. (2017). Key substance use and mental health indicators in the United States: Results from the 2016 National Survey on Drug Use and Health (SMA 17–5044; NSDUH Series H-52). Substance Abuse and Mental Health Services Administration (SAMHSA). [Google Scholar]
  48. Thornton LK, Harris K, Baker AL, Johnson M, & Kay-Lambkin FJ (2016). Recruiting for addiction research via Facebook. Drug and Alcohol Review, 35(4), 494–502. 10.1111/dar.12305 [DOI] [PubMed] [Google Scholar]
  49. Wolff S, Holl J, Stopsack M, Arens EA, Höcker A, Staben KA, Hiller P, Klein M, Schäfer I, Barnow S, & Group the C. S. (2016). Does emotion dysregulation mediate the relationship between early maltreatment and later substance dependence? Findings of the CANSAS study. European Addiction Research, 22(6), 292–300. 10.1159/000447397 [DOI] [PubMed] [Google Scholar]
  50. Zhang L, Mersky JP, & Topitzes J (2020). Adverse childhood experiences and psychological well-being in a rural sample of Chinese young adults. Child Abuse & Neglect, 108, 104658. 10.1016/j.chiabu.2020.104658 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1941176-supplement-1.pdf (147.9KB, pdf)

RESOURCES