Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Jun 1.
Published in final edited form as: Addict Behav. 2019 Jan 21;93:173–179. doi: 10.1016/j.addbeh.2019.01.024

Alcohol Use, Psychiatric Disorders and Gambling Behaviors: A Multi-Sample Study Testing Causal Relationships via the Co-Twin Control Design

Spencer B Huggett 1,2, Evan Winiger 1,2, Robin P Corley 2, John K Hewitt 1,2, Michael C Stallings 1,2
PMCID: PMC7170180  NIHMSID: NIHMS1520556  PMID: 30716592

Abstract

Human laboratory studies and twin research investigating relationships between alcohol use/pathology and gambling generally have yielded contradictory results, sometimes suggesting causal relationships and common genetic risk factors. 2860 individuals (mean age: 25.60, s.d = 3.21, 50.62% female) from separate clinical (n = 636) and community based (twin) samples (n = 2224) were used to assess associations between past year alcohol use and frequency of past year gambling behaviors (gambling frequency). After adjustment for demographic and psychiatric covariates, individual-level analyses detected that increased alcohol use was associated with more frequent gambling behaviors in twin and clinical samples. Co-twin control models were then used to test potential causal (direct) relationships between alcohol use and gambling frequency. Controlling for all covariates and shared genetic/environmental factors, we found increased alcohol use directly predicted more frequent gambling behaviors (consistent with causality). Our study also suggests shared genetic and/or environmental risk factors contribute to the association between increased alcohol use and frequent gambling behavior, a finding that may be more pronounced in males. The present study helps bridge the gap between twin research and human laboratory studies on gambling and alcohol use and corroborates findings across community and clinical samples. Overall, our findings support both common risk factors between alcohol use and gambling as well as a direct relationship between alcohol use and gambling frequency. Recognizing these dual processes could prove useful for gambling-related prevention/intervention programs.

Keywords: Gambling, Alcohol, Co-Twin Control, Psychiatric Disorders, Alcohol use, Gambling Frequency

Introduction

The relationship between alcohol use and gambling has been a critical topic of research for over five decades (Sjoberg, 1969). Two hypotheses have dictated much of this research: 1) alcohol use causes gambling and/or 2) common risk factors drive the association between alcohol use and gambling.

Experimental laboratory studies generally suggest that alcohol consumption causes increases in a variety of gambling behaviors, including: increased betting behavior, increased time spent gambling, larger and faster accumulation of losses, and problems evaluating risk and gambling odds (Cronce et al., 2010; Ellery et al. 2005; Kyngdon & Dickerson, 1999; Phillips & Ogeil, 2010). Contrarily, other laboratory studies report that alcohol consumption does not directly influence gambling behavior (Breslin et al. 1999; Sagoe et al. 2017), warranting more causally informed research. Although laboratory gambling settings can provide useful insight into causal relationships, they do not always generalize to real world situations (Anderson and Brown, 1984) and typically do not disentangle causal influences over time.

Twin studies are useful for understanding the genetic and environmental etiology of complex traits. Twin research on gambling has predominantly been conducted in middle-aged adult males via the Vietnam era twin registry sample (Henderson et al., 1990) and results suggest strong genetic overlap between pathological gambling and anti-social personality disorder (ASPD; Slutske et al. 2001), major depressive disorder (MDD; Potenza et al., 2005), generalized anxiety disorder (GAD; Giddens et al., 2011) and alcohol dependence (AD; Slutske et al. 2000). However, twin studies on gambling rarely examine young-adult samples and/or subclinical gambling outcomes, such as gambling frequency, which may yield different genetic and environmental etiologies than pathological/disordered gambling (Blanco et al. 2012). Understanding the contributions to subclinical gambling outcomes in younger adults is especially important for intervention/prevention techniques since the average age of onset for pathological gambling occurs between ages 29-34 (Blanco et al. 2006), which is later than average onset ages for numerous psychiatric disorders (Kessler et al. 2008). Studies on psychiatric comorbidities with gambling tend to use either community or clinically ascertained samples and typically focus on pathological gambling outcomes. More research is needed to determine psychiatric associations with subclinical gambling outcomes and the similarities/distinctions in the etiological forces influencing gambling behaviors in clinical and community samples.

Genetic overlap between alcohol and gambling outcomes (as observed by twin studies) would be largely inconsistent with causal relationships implied by various laboratory studies. A popular technique to perform “natural experiments” (in twins) and quantify potentially causal relationships is through co-twin control modeling, which could bridge the gap between these areas of research and complement laboratory-based studies in a real-world setting. To our knowledge no twin study has been done to explicitly assess the potential causal relationship between alcohol and gambling outcomes.

The current study made use of both clinical and community (twin) samples to inspect associations between alcohol use and gambling frequency over a one-year time frame. The primary goal was to test our hypotheses that: 1) alcohol use causes increases in gambling behaviors/frequency and 2) common genetic/environmental risk factors explain the association between alcohol and gambling behaviors. We tested these hypotheses using the co-twin control design in a sample of young adult twins and incorporated piecewise model fitting techniques to evaluate potential confounds of psychiatric traits that are typically associated with gambling outcomes (Lorains et al. 2011).

Methods

Our full sample included 2,860 individuals participating in the Colorado Center on Antisocial Drug Dependence (CADD; PI: J.K. Hewitt), who were of legal drinking and gambling age (age at assessment > 21). The full sample included sibling pairs who grew-up together and were ascertained through either clinical or community settings. Participants were excluded if they had an IQ < (80 – standard error), had a history of psychosis or were not English speakers.

Clinical Sample

Clinical subjects included clinical probands and their biological siblings that were recruited via substance abuse treatment facilities and the criminal justice system. Probands were required to have one or more symptoms of a substance use disorder and one or more symptoms of conduct disorder. Siblings were not required to meet these inclusion criteria but show indirect selection for these behaviors (see Supplementary Table S1). In total, 636 individuals were included in the clinical sample (age-range: 21-47; M = 28.00; s.d. = 3.70).

Community (Twin) Sample

Twins were recruited from birth records (Colorado Longitudinal Twin Sample) and Colorado school districts (Colorado Twin Sample) and were assessed as part of the CADD beginning in 1998 (Rhea et al. 2013). Altogether the twin sample included 2,224 monozygotic (MZ) and same/opposite-sex dizygotic (DZ) twins (age-range: 21-34; M = 24.91; s.d. = 2.70).

Measures

Past year gambling frequency was assessed via the South Oaks Gambling Screen (SOGS; Lesieur & Blume, 1987). The SOGS measure encompasses a wide array of individual gambling behaviors, including: playing games for money (cards, dice, pool, poker, slots etc.), trading stocks, betting for money (horse/animal races, sports, lottery) and online gambling. Participants endorsed the frequency of engaging in these behaviors over the past year using the following scale: not at all, less than once a week, once a week or more. Previous researchers assessing gambling frequency using the SOGS have created a composite score summing/averaging across gambling behaviors (Vachon et al. 2004; Wanner et al. 2009), which was highly skewed in our data (see Supplementary Figure S1). To accommodate for this non-normality and assess the frequency of behavior irrespective for types of gambling, we categorized gambling frequency as an ordinal variable. We coded participants who reported any gambling behavior once or more a week as 2’s (defined as weekly gambling), coded those who endorsed gambling less than once a week as 1’s and coded subjects who did not gamble in the past year as 0’s. Gambling frequency yielded good internal consistency, as assessed by an ordinal Cronbach’s alpha (Zumbo et al. 2007) α = 0.87. Problem gambling was determined from a single SOGS question: “Do you feel you have ever had a problem with gambling?” We coded problem gamblers as 1’s and non-problem gamblers as 0’s. Gambling frequency was highly associated with problem gambling, Log Odds Ratio = 2.46, 95% CI [1.41, 3.50], p < 0.001, as revealed from a logistic mixed-effects analysis.

Past year alcohol use was assessed via a semi-structured interview: the Composite International Diagnostic Interview – Substance Abuse Module (CIDI-SAM; Cottler et al. 1989). Frequency of alcohol use in the past year was assessed with a single question: “How many weeks in the past 12 months did you drink at all?” We transformed responses to past year alcohol use into four ordinal categories (see Supplementary Figure S1). We coded those reporting drinking 48-52 weeks in the past year as 3’s (defined as weekly drinking), coded subjects endorsing drinking 12-47 weeks as 2’s (at least monthly use), coded participants drinking < 12 weeks as 1’s and coded those who did not drink alcohol in the past year as 0’s. To assess the reliability of this metric, we paired it with a supplemental CIDI-SAM question: “How many days did you drink in the past 6 months?” We converted responses to this question into proportional ordinal categories to match past-year alcohol use. Using the irr package in R (Gamer et al. 2012) and guidelines outlined in (Koo & Li, 2016), we computed the intraclass correlation coefficient (ICC) of these alcohol use measures (model = ‘two-way’, type = ‘consistency’) and found moderate reliability, ICC = 0.645, 95% CI [0.62, 0.67], p < 0.001.

Measures of psychiatric traits were quantified with semi-structured interviews via the Diagnostic Interview Schedule (Robins et al. 2000) and the CIDI-SAM (Cottler et al. 1989). Lifetime symptoms of ASPD, MDD, GAD, alcohol abuse (AA) and AD were based on Diagnostic and Statistical Manual of Mental Disorders version 4 criteria. We transformed symptom counts into z scores centered on the community twin sample. Standard and ordinal Cronbach’s alpha reliability tests (Cronbach, 1951; Zumbo et al. 2007) indicated excellent internal consistency for MDD (α = 0.95), GAD (α = 0.94) and AD (α = 0.94), and acceptable levels for AA (α = 0.79) and ASPD (α = 0.78).

Analyses

To account for the nested structure of our data and the ordinal categorizations of gambling frequency, we fit mixed-effects probit regression models via the ordinal package in R (Christensen, 2013). We used the clmm command to estimate cumulative link mixed-effects models, which we specified with “flexible” unstructured thresholds. Using a piecewise model fitting approach, we explored the relationship between past-year alcohol use and gambling frequency by fitting three nested models. Model 1 regressed gambling frequency on alcohol use adjusting for age (within-sample z score transformation) and sex (males coded as 1’s / females as −1’s). Model 2 further explored this relationship by adding covariates for lifetime ASPD, MDD and GAD symptoms. Model 3 extended this model by additionally adjusting for AA and AD.

Previous research has observed sex and age differences among the relationships between alcohol use and psychiatric disorders with gambling outcomes (Desai et al. 2007; Desai & Potenza, 2008; Petry et al. 2005). We tested for sex and age differences in our analyses via interaction terms with age and sex with all predictors/covariates. We performed individual-level models separately in clinical and community (twins) samples and then altogether in the full sample controlling for sample/cohort (clinical coded as 1’s, twins coded as −1’s). We included interaction terms between sample and all predictors/covariates to test possible cohort differences in full sample analyses. Unless otherwise specified, we reported results from individual-level analyses using the full sample.

We then investigated potential causal relationships via co-twin control analyses. Co-twin control analyses are a popular quasi-experimental approach that uses twin-pair comparisons to decompose individual-level associations into within-twin pair and between-twin pair effects (McGue et al. 2010). Twin-pairs are matched on their genetic background (perfectly for MZ twins, partially for DZ twins) and shared environmental experiences and thus control for numerous unmeasured confounds and provide a basis for assessing potentially causal relationships. In our primary analysis (alcohol use predicting gambling frequency) the between-twin effect (βB) for alcohol use would capture what is common among twin-pairs, is estimated from the average alcohol use for each twin-pair and reflects the association attributable to shared genetic and environmental factors (common risk factors). The within-twin effect (βW) estimates the potential causal (direct) effect of alcohol use on gambling and it is derived from the difference of an individual twins’ alcohol use from their twin-pair mean. Within-twin pair prediction of gambling frequency removes confounds from shared genetic and environmental influences and is consistent with a potential causal effect.

Co-twin control analyses included both MZ and DZ twin-pairs to increase sample size and power. To test for potential confounds by zygosity (MZs coded as 0’s / DZs coded as 1’s) we included interaction terms between zygosity and our between and within-twin pair predictors. We also assessed interactions among between and within-twin pair predictors in DZ specific analyses to test for etiological differences among opposite-sex DZ twin pairs. We detected no significant interactions among zygosity, all p > 0.199, indicating consistent effects across MZ and DZ twins and same sex and opposite sex DZ twins. We presented results using the full twin sample and from MZ specific analyses since MZ twins share ~100% of their genetic influences and therefore within-twin pair estimates are free from genetic confounding. Additional confounds not shared among twin-pairs can bias within-twin pair effects (Frisell et al. 2012). To address this, we used the same piecewise model fitting technique used for individual-level analyses (i.e. fitting models 1-3) to assess co-twin control results. We further explored significant predictors/covariates from individual-level twin-sample analyses by assessing their between and within-twin pair effects (aside from age and sex) in co-twin control analyses.

Results

Descriptive Information

Table 1 displays the descriptive statistics for our participants by sample. Males yielded significantly more symptoms of ASPD, AD, AA, and more frequent alcohol use and gambling behaviors, all p < 0.001. Females had higher rates of GAD and MDD symptomology, all p < 0.023. As expected, the clinical sample demonstrated increased symptoms of MDD, GAD, ASPD, AA, AD and higher rates of past year gambling, all p < 0.004. Weekly alcohol use was similar across clinical (27.36% weekly drinkers) and community samples (28.63% weekly drinkers), suggesting alcohol use is not a clinically manifested behavior. However, alcohol use was positively associated with AA and AD, all p < 0.001, and hence, appears to be tapping into similar sources of variation as clinical alcohol symptomology.

Table 1.

Pertinent Sample Information

Descriptive Information by Sample: M (s.d.)
Variable Sample
Full Sample Clinical Twins
n 2860 636 2224
Age 25.60 (3.21) 28.00 (3.70) 24.91 (2.70)
Sex [% male] 50.62 59.75 45.05
Antisocial Personality Disorder (ASPD) Symptoms 1.60 (1.88) 2.84 (2.25) 1.25 (1.59)
Major Depressive Disorder (MDD) Symptoms 1.45 (2.86) 1.88 (3.17) 1.33 (2.76)
Generalized Anxiety Disorder (GAD) Symptoms 0.60 (1.57) 0.92 (1.83) 0.51 (1.46)
Alcohol Abuse (AA) Symptoms 0.81 (1.04) 1.35 (2.04) 0.65 (0.90)
Alcohol Dependence (AD) Symptoms 1.36 (1.79) 2.08 (2.06) 1.15 (1.61)
Past Year Alcohol Use [% weekly] 28.39 27.36 28.63
Past Year Gambling Frequency [% weekly] 7.97 13.37 6.40
Problem Gambling [% ever] 1.71 3.14 1.30
Open in a new tab

Table 2 provides descriptive information for each of the individual gambling behaviors and their associations with alcohol use and problem gambling. The most frequent gambling behaviors were lottery/playing the numbers and online gambling or slot/gambling machines (28.29%-26.05%). Controlling for age, sex, cohort and all their interactions, we identified significant associations between alcohol use and most individual gambling behaviors (all β > 0.15, all Z > 2.01, all p < 0.045). Similarly, after covariate adjustment, higher rates for most individual gambling behaviors significantly predicted increased rates of problem gambling, all Log Odds ≥ 1.22, Z > 2.02, all p < 0.044. Of our full sample, 52.76% reported any gambling and 87.48% reported consuming alcohol in the past year.

Table 2.

Individual Gambling Behaviors in the Past Year

Descriptive Information on Individual Gambling Behaviors and Their Associations with Alcohol use and Problem Gambling
Frequency of Gambling Behavior in the Past Year Associations with β / Log Odds (s.e.)
Individual Gambling Behaviors % < Once a Week % Weekly or More Alcohol Use Problem Gambling
Bet on Lotteries or Played the Numbers 24.83 3.46 0.15 (0.04)*** 1.55 (0.53)**
Played Slot Machines, Poker Machines or Online Gambling Games 24.76 1.29 0.21 (0.04)$*** 1.85 (0.59)***
Played Cards for Money 21.89 2.31 0.20 (0.04)*** 1.87 (0.53)***
Bet on Other Sports 11.08 2.03 0.18 (0.09)* 1.22 (0.58)*
Played Dice Games for Money 8.64 0.87 0.16 (0.09) 1.22 (0.61)*
Played Pool or Other Games of Skill for Money 7.48 1.43 0.55 (0.16)$*** 0.66 (0.69)
Traded Stocks and/or Commodities Online 3.29 0.80 −0.02 (0.12) −1.79 (1.19)
Bet on Horses, Dogs or Other Animals 3.25 0.21 0.31 (0.18) 0.79 (0.44)
Open in a new tab

Individual gambling behaviors are sorted by frequency. Alcohol use (independent variable) was associated with most individual gambling behaviors. Most individual gambling behaviors (independent variables) were associated with problem gambling. To assess associations between individual gambling behaviors and problem gambling, we performed logistic mixed effects models using age, sex, sample/cohort as covariates and adjusting for interactions between all predictors/covariates with sample/cohort.

Note:

$

indicates a significant interaction with sex (p < 0.05)

Individual-Level Models

Table 3 reports parameter estimates from individual-level analyses (models 1-3) predicting past year gambling frequency in the clinical, twin and full samples. Increased past year alcohol use was associated with more frequent gambling behavior across samples and after adjustment for all covariates, indicating that these associations were robust, all β > 0.14, all Z > 3.28, all p ≤ 0.001. Model 1 indicated a significant interaction between sample and alcohol use on gambling, β = −0.07, Z = −2.16, p = 0.031 – raising the possibility of a stronger effect of alcohol on gambling in community-based samples. However, this interaction did not persist after adjustment of psychiatric traits (i.e., models 2 and 3; all β < −0.05, all Z > −1.66, all p > 0.096). We observed only one additional significant interaction at the individual-level (see Table1/model 3), suggesting consistent effects of predictors/covariates across sex, age and/or sample.

Table 3.

Results from Individual-Level Analyses Predicting Gambling Frequency

Individual-Level Associations: Unstandardized Estimates: β (s.e.)
Model Variable Sample
Full Sample Clinical Twins
Model 1 Age 0.03 (0.06) 0.06 (0.10) −0.01 (0.08)
Sex 0.21*** (0.06) 0.21* (0.10) 0.17* (0.08)
Sample 0.18** (0.07) NA NA
Past Year Alcohol Use 0.19&*** (0.03) 0.11* (0.06) 0.27*** (0.04)
Model 2 Age 0.03 (0.06) 0.09 (0.10) −0.02 (0.08)
Sex 0.15* (0.06) 0.13 (0.11) 0.15* (0.08)
Sample 0.11 (0.07) NA NA
Past Year Alcohol Use 0.19*** (0.03) 0.12* (0.06) 0.26*** (0.04)
ASPD Symptoms 0.11*** (0.03) 0.05 (0.05) 0.13*** (0.03)
MDD Symptoms −0.01 (0.03) 0.01 (0.05) −0.02 (0.04)
GAD Symptoms 0.02 (0.03) 0.03 (0.05) 0.01 (0.04)
Model 3 Age 0.02 (0.07) 0.08 (0.11) −0.04 (0.08)
Sex 0.23*** (0.07) 0.28** (0.12) 0.18* (0.08)
Sample 0.00 (0.08) NA NA
Past Year Alcohol Use 0.14*** (0.03) 0.13* (0.06) 0.19*** (0.04)
ASPD Symptoms 0.04 (0.03) 0.01 (0.06) 0.08* (0.04)
MDD Symptoms −0.01 (0.03) 0.04 (0.05) −0.04 (0.04)
GAD Symptoms 0.03 (0.03) 0.03 (0.05) 0.00 (0.04)
Alcohol Abuse Symptoms 0.10** (0.04) 0.08 (0.07) 0.07 (0.04)
Alcohol Dependence Symptoms 0.00@(0.04) −0.01 (0.06) 0.03 (0.04)
Open in a new tab

Note:

&

represents significant interaction with sample (p < 0.05)

@

represents significant interaction with age (p < 0.05)

Model 2 demonstrated a significant association between ASPD symptoms and gambling frequency, β = 0.11, Z = 3.64, p < 0.001, but not MDD, β = −0.01, Z = −0.36, p = 0.723, nor GAD, β = 0.02, Z = 0.80, p = 0.422. The association between ASPD and gambling was most robust in the twin-sample, β = 0.13, Z = 3.954, p < 0.001 and not significant within the clinically ascertained individuals, β = 0.05, Z = 1.05, p = 0.296. After controlling for AA and AD (model 3), ASPD symptoms did not significantly predict gambling frequency, β = 0.04, Z = 1.35, p = 0.176 – suggesting that this association could partially be attributed to clinical alcohol symptomology. Controlling for all covariates, increased symptoms of AA were associated with more frequent gambling behavior, β = 0.10, Z = 2.76, p = 0.006. However, this association was not significant in either the twin sample, β = 0.08 Z = 1.85, p = 0.065, nor clinical sample, β = 0.08 Z = 1.15, p = 0.252, independently. We further explored the nature of the significant twin associations using co-twin control modeling.

Co-twin Control Models

Table 4 shows results from co-twin control models 1-3 predicting past year gambling frequency. Adjusting for all covariates (model 3), the within-twin pair effect of alcohol use significantly predicted gambling frequency, βW = 0.17, Z = 2.20, p = 0.028, indicating that the twin who uses alcohol more frequently than their co-twin also gambles more frequently (consistent with causality). Model 3 also revealed a significant between-twin pair effect between alcohol use and gambling frequency, βB = 0.23, Z = 3.11, p = 0.002, suggesting the presence of shared genetic and/or environmental risk factors. Between-twin pair effects interacted with sex across all models in the full twin-sample, all βB > 0.12, all Z > 2.51, all p < 0.012, suggesting a stronger genetic and/or environmental overlap between alcohol use and gambling frequency in males.

Table 4.

Results from Co-Twin Control Analyses Predicting Gambling Frequency

Co-Twin Control Analyses: Unstandardized Estimates: β (s.e.)
Model Variable Sample
All Twins MZ Twins
Model 1 Age 0.01 (0.10) −0.02 (0.06)
Sex −0.02 (0.10) 0.02 (0.16)
Zygosity 0.07 (0.19) NA
βW [Alcohol Use] 0.22** (0.08) 0.25$** (0.08)
βB [Alcohol Use] 0.35$*** (0.07) 0.39$*** (0.08)
Model 2 Age −0.01 (0.10) −0.03 (0.16)
Sex −0.04 (0.10) −0.02 (0.16)
Zygosity 0.07 (0.20) NA
βW [Alcohol Use] 0.22** (0.08) 0.25** (0.08)
βB [Alcohol Use] 0.33$*** (0.07) 0.38*** (0.08)
βW [ASPD] Symptoms 0.14 (0.07) 0.18$* (0.08)
βB [ASPD] Symptoms 0.13* (0.06) 0.12 (0.07)
MDD Symptoms −0.03 (0.04) −0.04 (0.06)
GAD Symptoms 0.00 (0.04) 0.05 (0.06)
Model 3 Age −0.04 (0.11) −0.03 (0.18)
Sex 0.01 (0.11) 0.07 (0.19)
Zygosity −0.01 (0.21) NA
βW [Alcohol Use] 0.17* (0.08) 0.21* (0.08)
βB [Alcohol Use] 0.23$** (0.08) 0.27** (0.09)
βW [ASPD] Symptoms 0.10 (0.08) 0.15 (0.08)
βB [ASPD] Symptoms 0.09 (0.07) 0.07 (0.08)
MDD Symptoms −0.05 (0.04) −0.06 (0.06)
GAD Symptoms 0.00 (0.04) 0.06 (0.06)
Alcohol Abuse Symptoms 0.07 (0.04) 0.01 (0.07)
Alcohol Dependence Symptoms 0.03 (0.04) 0.07 (0.07)
Open in a new tab

Note:

$

indicates a significant interaction with sex (p < 0.05)

To test the specificity of our co-twin control results with individual gambling outcomes, we assessed the relationships between alcohol use and specific gambling behaviors (see Supplementary Table S2). Controlling for all covariates (i.e., model 3), we detected potentially causal relationships between alcohol use and 1) betting on the lottery/playing the numbers, and 2) playing slots/gambling machines or online gambling, all βW = 0.27, all Z > 2.74, all p < 0.006. However, our analyses revealed that common genetic and/or environmental risk factors also explained part of the associations between alcohol use and 1) playing slots/gambling machines or online gambling, and 2) playing cards for money, all βB = 0.26, all Z > 2.85, all p < 0.004.

Discussion

After adjustment for demographics, psychiatric covariates and shared genetic and environmental influences, alcohol use was significantly associated with more frequent gambling behavior, consistent with causal claims reported in some (Ellery et al. 2005; Kyngdon & Dickerson, 1999) but not all (Ellery & Stewart, 2014) laboratory gambling studies. Our results extend findings from laboratory settings by providing real-world evidence that alcohol use may increase risk for gambling frequently and fit well with longitudinal analyses suggesting alcohol consumption predicts gambling behaviors over time (Barnes et al. 2002). The direct influences of alcohol consumption on gambling behavior may be mediated through various processes, such as enhancing general risk-taking behavior. Evidence suggests alcohol alters meso-cortico-limbic reward/motivation/incentive-salience pathways (Bjork et al., 2014), which could augment reward learning, alter motivational processes and heighten the salient aspects of gambling. More frequent gambling behavior was associated with a higher incidence of problem gambling and may increase risk for pathological gambling outcomes.

We also identified that common genetic and/or environmental risk factors contribute to the association between alcohol use and gambling frequency. These findings are consistent with twin research that observed genetic overlap between substance use and gambling frequency (Vitaro et al. 2014). Further evidence of shared genetic risk is found in genome-wide studies that indicate a genetic association between pathological gambling and AD (Lang et al. 2016). These common genetic risk factors could reflect broader personality traits, such as risk taking, which have demonstrated genetic associations with alcohol and drug use (Linner et al. 2019). Discussion of shared environmental risk factors is also warranted. Peer delinquency demonstrates associations with both alcohol use and gambling frequency (Barnes et al. 1999; Caldeira et al. 2017). Socioeconomic status and/or accessibility may also contribute to the environmental risk for alcohol use (Tobler et al. 2009) and gambling behaviors (Marshall, 2007).

Our study has various strengths. We controlled for numerous covariates, used clinical and community samples and assessed multiple gambling behaviors. Our analyses utilized an understudied age group for genetically informed designs, included both sexes and assessed behavioral and problem gambling outcomes.-We found associations between gambling frequency and externalizing-related traits (ASPD and AA), but no associations with internalizing traits (MDD and GAD). These findings suggest both overlapping and distinct risk factors between problem gambling and gambling frequency.

The present study, however, is not without its limitations. Co-twin control analyses and clinical-specific analyses had relatively small sample sizes. Additionally, we performed many analyses, but our threshold for statistical significance did not adjust for multiple testing. Significant results, therefore, should be interpreted cautiously. Both alcohol and gambling were quantified retrospectively, which limits our ability to unravel within-session alcohol/gambling dynamics. Nonetheless, alcohol use was temporally concurrent with gambling frequency and provided a specific time-window to dissect co-occurrence of these addictive behaviors.

We consider alternative interpretations of the inferences derived from within-twin pair predictions. Twins who use alcohol more frequently than their co-twin also gamble more frequently. Since many gambling institutions (i.e., casinos) typically allow and even encourage alcohol consumption, perhaps the frequent gambler also tends to use alcohol more because these activities are co-located. Thus co-twin control findings that are consistent with causality could be interpreted as consequential (alcohol use results from gambling frequently) and/or coincidental (alcohol use co-occurs with gambling without a causal connection). We found significant within-twin pair prediction of alcohol use on gambling behaviors not typically associated with simultaneous alcohol consumption (i.e., playing the lottery), which suggests our results are not simply coincidental. However, analyses of individual gambling activities are not necessarily exclusive to these behaviors. In our study, only 38/228 weekly gamblers engaged in just one individual gambling behavior (19/38 played the lottery only). Regular gamblers, therefore, typically engage in multiple forms of betting that may coincide with additional opportunities to drink.

In conclusion, our study demonstrates an intersection between human laboratory studies and twin research examining relationships between alcohol and gambling behaviors. We found support for both of our hypotheses that 1) alcohol use causes increases in gambling behaviors and 2) common genetic/environmental risk factors also contribute to the association between alcohol use and gambling frequency. We utilized a young adult sample and an age range that typically precedes onset ages for gambling problems (Petry et al. 2005). Our results, therefore, may be useful for preventative and intervention programs for gambling outcomes.

Supplementary Material

1

Highlights:

  • Increased alcohol use is associated with more frequent gambling behaviors in the past year for community and clinically ascertained samples.

  • Co-twin control results suggest that more frequent alcohol use may cause increases in gambling behaviors.

  • Co-twin control analyses also reveal shared genetic and environmental risk factors driving the association between alcohol use and gambling behaviors.

Acknowledgments

Role of Funding Sources

This work was funded via NIDA grants: P50DA011015 and T32DA017637. NIDA did not play a role in writing the manuscript, the overall study design, data analysis nor the submission of this paper for publication.

Footnotes

Conflict of Interest

The authors declare no conflicts of interest.

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 citable 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. Anderson G, & Brown RIF (1984). Real and laboratory gambling, sensation-seeking and arousal. British Journal of Psychology, 75, 401–410. [DOI] [PubMed] [Google Scholar]
  2. Barnes GM, Welte JW, Hoffman JH, & Dintcheff BA (2002). Effects of Alcohol Misuse on Gambling Patterns in Youth. Journal of Studies on Alcohol, 63, 767–775. [DOI] [PubMed] [Google Scholar]
  3. Barnes GM, Welte JW, Hoffman JH, Dintcheff BA (1999). GAMBLING AND ALCOHOL USE AMONG YOUTH : INFLUENCES OF DEMOGRAPHIC , SOCIALIZATION , AND INDIVIDUAL FACTORS. Addictive Behaviors, 24(6), 749–767. [DOI] [PubMed] [Google Scholar]
  4. Bjork JM. & Gilman JM (2014). Neuropharmacology The effects of acute alcohol administration on the human brain : Insights from neuroimaging. Neuropharmacology, 84, 101–110. 10.1016/j.neuropharm.2013.07.039 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Blanco C, Myers J, & Kendler KS (2012). Gambling, disordered gambling and their association with major depression and substance use: a web-based cohort and twin- sibling study. Psychological Medicine, 42(3), 497–508. 10.1017/S0033291711001401. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Blanco C, Hasin DS, Petry N, Stinson FS, & Grant BF (2006). Sex differences in subclinical and DSM-IV pathological gambling : results from the National Epidemiologic Survey on Alcohol and Related Conditions. Psychological Medicine, 36, 943–953. 10.1017/S0033291706007410 [DOI] [PubMed] [Google Scholar]
  7. Breslin FC, Sobell MB, Cappell H, Vakili S, & C. X. Poulos (1999). The Effects of Alcohol , Gender , and Sensation Seeking on the Gambling Choices of Social Drinkers. Psychology of Addictive Behaviors, 13, 243–252. [Google Scholar]
  8. Caldeira KM, Arria AM, Grady KEO, Vincent KB, Robertson C, & Welsh CJ (2017). Risk factors for gambling and substance use among recent college students. Drug and Alcohol Dependence, 179, 280–290. 10.1016/j.drugalcdep.2017.06.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Christensen RHB (2018). Regression Models for Ordinal Data. Http://www.cran.r-Project.org/package=ordinal/. [Google Scholar]
  10. Cottler LB, Robins LN, Helzer JE (1989). The reliability of the CIDI-SAM: A comprehensive substance abuse interview. British Journal of Addiction, 84, 801–814 [DOI] [PubMed] [Google Scholar]
  11. Cronbach LJ (1951). COEFFICIENT ALPHA AND THE INTERNAL STRUCTURE OF TESTS. Psychometrika, 16, 297–334. [Google Scholar]
  12. Cronce JM, & Corbin WR (2011). Effects of Alcohol and Initial Gambling Outcomes on Within-Session Gambling Behavior. Exp Clin Psychopharmacol, 18, 145–157. 10.1037/a0019114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Desai RA, Desai MM, & Potenza MN (2007). Gambling , Health , and Age : Data from the National Epidemiologic Survey on Alcohol and Related Conditions. Psychology of Addictive Behaviors, 21, 431–440. 10.1037/0893-164X.21.4.431 [DOI] [PubMed] [Google Scholar]
  14. Desai RA, & Potenza ÆMN. (2008). Gender differences in the associations between past-year gambling problems and psychiatric disorders. Soc Psychiatry Psychiatr Epidemiol, 43, 173–183. 10.1007/s00127-007-0283-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ellery M, & Stewart SH (2014). Alcohol Affects Video Lottery Terminal ( VLT ) Gambling Behaviors and Alcohol Affects Video Lottery Terminal ( VLT ) Gambling Behaviors and Cognitions Differently. Psychology of Addictive Behaviors, 28, 206–216. 10.1037/a0035235 [DOI] [PubMed] [Google Scholar]
  16. Ellery M, Stewart SH, & Loba P (2005). Alcohol ‘ s Effects on Video Lottery Terminal ( VLT ) Play Among Probable Pathological and Non-Pathological Gamblers Alcohol ‘ s Effects on Video Lottery Terminal ( VLT ) Play Among Probable Pathological and Non-Pathological Gamblers. Journal of Gambling Studies, 21, 299–324. 10.1007/s10899-005-3101-0 [DOI] [PubMed] [Google Scholar]
  17. Frisell T, Oberg S, Kuja-Halkola R, & Sjolander A (2012). Sibling Comparison Designs. Epidemiology, 23, 713–720. 10.1097/EDE.0b013e31825fa230 [DOI] [PubMed] [Google Scholar]
  18. Gamer M, Lemon J, Fellows I, & Singh P (2012). irr: Various Coefficients of Interrater Reliability and Agreement. https://CRAN.R-project.org/package=irr [Google Scholar]
  19. Giddens JL, Xian H, Scherrer JF, Eisen SA, & Potenza MN (2011). Shared genetic contributions to anxiety disorders and pathological gambling in a male population. Journal of Affective Disorders, 132, 406–412. 10.1016/j.jad.2011.03.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Henderson WG, Goldberg J, True WR, Barnes JE, & Vitek ME (1990). The Vietnam Era Twin Registry : a Resource for Medical Research. Public Health Reports, 105, 368–373. [PMC free article] [PubMed] [Google Scholar]
  21. Kessler RC, Hwang I, Labrie R, Petukhova M, Sampson NA, Winters KC, & Shaffer HJ (2008). The prevalence and correlates of DSM-IV Pathological Gambling in the National Comorbidity Survey Replication Ronald. Psychological Medicine, 38, 1351–1360. 10.1017/S0033291708002900’. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Koo TK, & Li MY (2016). A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. Journal of Chiropractic Medicine, 15, 155–163. 10.1016/j.jcm.2016.02.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kyngdon A, & Dickerson M (1999). An experimental study of the effect of prior alcohol consumption on a simulated gambling activity. Addiction, 94, 697–707. [DOI] [PubMed] [Google Scholar]
  24. Lang M, Lemenager T, Streit F, Fauth-Buhler M, Frank J, Juraeva D, … Mann KF (2016). Genome-wide association study of pathological gambling. European Psychiatry, 36, 38–46. 10.1016/j.eurpsy.2016.04.001 [DOI] [PubMed] [Google Scholar]
  25. Lesieur HR, & Blume SB (n.d.). The South Oaks Gambling Screen (SaGS): A New Instrument for the Identification of Pathological Gamblers. Am J Psychiatry. 144, 1184–1188. [DOI] [PubMed] [Google Scholar]
  26. Lind PA, Zhu G, Montgomery GW, Madden PAF, Heath AC, Martin NG, & Slutske WS (2012). Genome-wide association study of a quantitative disordered gambling trait. Addiction Biology, 18, 511–522. 10.1111/j.1369-1600.2012.00463.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Linner RK, Biroli P, Kong E, Meddens FW, Wedow R, … Beauchamp JP (2019). Genomewide association analyses of risk tolerance and risky behaviors in over 1 million individuals identify hundreds of loci and shared genetic influences. Nature Genetics, 10.1038/s41588-018-0309-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lorains FK, Cowlishaw S, & Thomas SA (2011). Prevalence of comorbid disorders in problem and pathological gambling : systematic review and meta-analysis of population surveys. Addiction. 106 490–498. 10.1111/j.1360-0443.2010.03300.x [DOI] [PubMed] [Google Scholar]
  29. Marshall D (2005). The Gambling Environment and Gambler Behaviour: Evidence from Richmond-Tweed Australia. International Gambling Studies, 5, 63–83. 10.1080/14459790500099471 [DOI] [Google Scholar]
  30. Mcgue M, Osier M, & Christensen K (2010). Causal Inference and Observational Research : The Utility of Twins. Prespectives on Psychological Science, 5, 546–556. 10.1177/1745691610383511 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Phillips JG, & Ogeil RP (2010). Alcohol influences the use of decisional support. Psychopharmacology. 208, 603–611. 10.1007/s00213-009-1762-6 [DOI] [PubMed] [Google Scholar]
  32. Potenza MN, Xian H, Shah K, Scherrer JF, & Eisen SA (2005). Shared Genetic Contributions to Pathological Gambling and Major Depression in Men. Archives of General Psychiatry. 62 1015–1021. [DOI] [PubMed] [Google Scholar]
  33. Rhea S-A, Gross AA, Haberstick BC, & Corley RP (2013). Colorado Twin Registry—An Update. Twin Research and Human Genetics. 16 1–14. 10.1017/thg.2012.93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Robins LN, Cottler LB, Bucholz KK. (2000). Diagnostic Interview Schedule for the DSM-IV (DIS-IV). Compton WM, North CS, and Rourke KM. [Google Scholar]
  35. Sagoe D, Mentzoni RA, Leino T, Molde H, Haga S, & Fredericson M (2017). The effects of alcohol expectancy and intake on slot machine gambling behavior. Journal of Behavioral Addictions. 6 203–211. 10.1556/2006.6.2017.031 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sjoberg L (1969). Alcohol and Gambling. Psychopharmacologia. 14 284–298. [DOI] [PubMed] [Google Scholar]
  37. Slutske WS, Eisen S, Xian H, True WR, Lyons MJ, & Goldberg J (2001). A Twin Study of the Association Between Pathological Gambling and Antisocial Personality Disorder. Journal of Abnormal Psychology. 110 297–308. [DOI] [PubMed] [Google Scholar]
  38. Slutske WS, Ellingson JM, Richmond-rakerd LS, Zhu G, & Martin NG (2014). Shared Genetic Vulnerability for Disordered Gambling and Alcohol Use Disorder in Men and Women: Evidence from a National Community-Based Australian Twin Study. Twin Research and Human Genetics. 16 525–534. 10.1017/thg.2013.11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Slutske WS, Eisen S, True WR, Lyons MJ, Goldberg J, & Tsuang M (2000). Common Genetic Vulnerability for Pathological Gambling and Alcohol Dependence in Men. Archives of General Psychiatry. 57 666–673. [DOI] [PubMed] [Google Scholar]
  40. Stallings MC, Corley RP, Hewitt JK, Krauter KS, Lessem JM, Mikulich SK, … Crowley TJ (2003). A genome-wide search for quantitati v e trait loci influencing substance dependence v ulnerability in adolescence. Drug and Alcohol Dependence. 70 295–307. 10.1016/S0376-8716(03)00031-0 [DOI] [PubMed] [Google Scholar]
  41. Tobler AL, Komro KA, & Maldonado-Molina MM (2009). Relationship Between Neighborhood Context, Family Management Practices and Alcohol Use Among Urban, Multi-ethnic, Young Adolescents. Preventative Science. 10, 313–324. doi: 10.1007/s11121-009-0133-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Vachon J, Vitaro F, Wanner B, & Tremblay RE (2004). Adolescent Gambling : Relationships With Parent Gambling and Parenting Practices. Psychology of Addictive Behaviors, 18 398–401. 10.1037/0893-164X.18.4.398 [DOI] [PubMed] [Google Scholar]
  43. Vitaro F, Hartl AC, Brendgen M, Laursen B, Dionne G, & Boivin M (2014). Genetic and Environmental Influences on Gambling and Substance Use in Early Adolescence. Behavior Genetics, 44, 347–355. 10.1007/s10519-014-9658-6 [DOI] [PubMed] [Google Scholar]
  44. Wanner B, Vitaro F, & Tremblay RE (2009). Cross-Lagged Links Among Gambling , Substance Use , and Delinquency From Midadolescence to Young Adulthood : Additive and Moderating Effects of Common Risk Factors. Psychology of Addictive Behaviors, 23, 91–104. 10.1037/a0013182 [DOI] [PubMed] [Google Scholar]
  45. Zmnbo BD, Gadermann AM, & Zeisser C (2007). Ordinal Versions of Coefficients Alpha and Theta for Likert Rating Scales. Journal of Modern Applied Statistical Methods, 6, 21–29. 10.22237/jmasm/1177992180 [DOI] [Google Scholar]

Associated Data

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

Supplementary Materials

1
NIHMS1520556-supplement-1.docx (81.2KB, docx)

RESOURCES