Gambling has become increasingly accessible and socially acceptable over the past two decades, with an increasing number of venues and opportunities through casinos, video lottery terminals, sports betting venues, and online poker and other gambling sites. Although most people participate in gambling activities recreationally, some experience gambling problems, including the most severe form, gambling disorder. Gambling disorder has been associated with significant financial debt, family tension, divorce, and criminal activity such as fraud and embezzlement. Extreme cases have involved staged kidnappings and serious child neglect leading to death, murder, and suicide.
Gambling disorder is defined as persistent and recurrent maladaptive gambling behavior that jeopardizes personal, occupational, or social functioning. In the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), gambling disorder is classified as the first non–substance-based disorder in a new addiction category.
The Psychiatric Nosology of Gambling Disorder
Gambling disorder has been conceptualized as a disorder falling within an obsessive-compulsive spectrum and as a “behavioral addiction.” Studies of impulse control disorders describe clinical elements including an urge to engage in a typically enjoyable yet, in the long term, counterproductive or harmful behavior, a mounting tension until the behavior is completed, a temporary abatement of tension following completion of the behavior, and a return of tension or appetitive urge following varying amounts of time. Impulse control disorders have been described as having elements of impulsivity and compulsivity. Although the underlying motive of gambling disorder is initially pleasure, with increasing frequency individuals may feel out of control and their urges may become unpleasant or ego-dystonic. Although some compulsive aspects to gambling are evident, the co-occurrence of obsessive compulsive disorder and gambling disorder is not that common, while comorbidity with substance dependence occurs frequently. The diagnostic criteria of gambling disorder, listed in the DSM-5, share similarities with those for substance dependence. Individuals with gambling disorder can demonstrate tolerance and withdrawal symptoms as they gamble with increasing amounts of money in order to achieve the same hedonic experience, and they may become irritable or restless when attempting to cut down or quit their gambling. Like individuals with drug addictions, those with gambling disorder demonstrate impaired control over their behavior and may hide the extent of their involvement from loved ones or commit forgery or fraud to sustain their gambling. The term problem gambling has been at times used to describe less severe patterns of gambling than exhibited in gambling disorder. This category is conceptually similar to that of substance abuse, although no formal criteria exist for problem gambling. In addition, the term has been used at times inclusive and at other times exclusive of gambling disorder. The most commonly used screening instrument for gambling disorder is the South Oaks Gambling Screen, and this screen queries the types and frequencies of gambling behaviors as well as gambling-related impact on life functioning, particularly with respect to borrowing money for gambling. The South Oaks Gambling Screen is valid and reliable, and a score ≥5 signifies probable gambling disorder.
A frequently acknowledged criterion of gambling disorder is the “chasing” of losses, whereby gamblers attempt to regain accumulated losses by returning to a gambling venue shortly following sustaining gambling losses. Nearly winning (e.g., receiving identical symbols on 2 of the 3 reels on an electronic gambling machine) has been suggested to contribute to gambling behaviors. Individuals with gambling disorder, as well as recre-ational gamblers, may report other cognitive distortions, such as overestimating their chances of winning and their sense of control: “I know what it takes to win this game.” In dice gambling and certain other forms, individuals may keep track of previous numbers in order to inform their subsequent bets with the thought that certain numbers will either appear more frequently because they have been observed previously (“hot numbers”) or not (“numbers that are due”). Such a gambler’s fallacy ignores laws of probability: that each role of the dice functions independently of the last. Superstitious behaviors (“I only play at nights”) and attributional biases (“That dealer always makes me lose”) are also expressed in gambling disorder as well as in recreational gambling groups. Cognitive distortions may represent relevant considerations in the maintenance of gambling disorder, although their frequent occurrence in nonpathological gambling samples questions their centrality to the disorder.
Prevalence Estimates and Characteristics
Precise gambling disorder prevalence estimates may be related to assessment measures and other factors. However, most studies report lifetime prevalence estimates ranging from 0.4% to 3% in the general population, representing approximately 2 to 3 million adults in the United States. All types of gambling are also not equally represented in gambling disorder populations; one study suggests that pull-tabs, casino gambling, bingo, cards, lottery and sports betting, in descending order, are most strongly associated with gambling disorder, and another study found the highest proportion of pathological gamblers at off-track compared with other venues. Pathological gamblers may engage in multiple types of gambling. Factors associated with gambling disorder include male sex, adolescent and young adult age, and presence of other psychiatric disorder(s). In addition, minorities and persons with a lower socioeconomic status also appear more likely to gamble and may be at particular risk for gambling disorder. Some studies have found that men and women with gambling disorder show similarities in demographic and clinical features, including time spent gambling, percentage of income lost through gambling, and gambling urge severity. Other studies have identified gender differences in manifestations of gambling behaviors that may have significant implications for prevention and treatment strategies. Although men constitute about two-thirds of the gambling disorder population and often show a longer duration of onset and begin gambling early in life (childhood/adolescence), women appear more likely to develop gambling disorder later in life and demonstrate a more rapid progression between onset and problematic engagement, a phenomenon observed in substance use behaviors and described as “telescoping.” Gender differences also exist in the types of gambling behavior and in gambling triggers. Women may report engaging in fewer forms of gambling, mostly bingo and slot machines, and often cite feeling prompted by negative mood states. In contrast, men are more likely to gamble on cards or sporting events and report a greater saliency of sensory cues, such as sounds or advertisement, in their triggers for gambling. In addition, women, as compared with men with gambling problems, may experience greater psychiatric comorbidity, particularly with mood and anxiety disorders.
Gambling disorder frequently co-occurs with other psychiatric disorders. Some studies estimate that up to three-fourths of individuals with gambling disorder report an alcohol use disorder, over 60% are daily tobacco smokers or nicotine dependent, and up to 40% report other drug abuse. About half of individuals diagnosed with gambling disorder also experience a mood disorder, with a particularly high odds ratio of 8.6 for mania, and roughly 40% are also diagnosed with anxiety disorders.
Estimates of personality disorders range from 29% to 93% in the gambling disorder population, with one study reporting an average of 4.6 personality disorders per person with gambling disorder. Although borderline, histrionic, and antisocial personality disorders are most often cited, these may represent a component of an externalizing syndrome. Personality and temperamental factors may play a role in the maintenance of gambling disorder, as pathological gamblers may show high levels of impulsiveness, novelty-seeking, rigidness, extravagance, and harm avoidance combined with low levels of self-directedness. In particular, impulsivity has been investigated as a key underlying construct, and accordingly, in gambling disorder, severity of gambling behavior and psychological disturbances appear related to this measure. Identification of co-occurring disorders is important as the disorders may guide treatment strategies and influence treatment outcome.
The Biochemistry of Gambling Disorder
Pathological gambling shares similar biochemical features with substance dependence and other disorders characterized by impulsive features. Low central levels of serotonin metabolites are observed in the cerebral spinal fluid samples of individuals with impaired impulse control including those with gambling disorder. However, the precise nature of central serotonin function in gambling disorder is complicated by findings suggesting increased levels in gambling disorder. Low endogenous levels of serotonin in gambling disorder are suggested by blunted prolactin responses following a pharmacological challenge. Pharmacological challenges using the partial agonist metachlorophenylpiperazine produce a euphoric high in pathological gamblers, a response also observed in individuals with other impulsive disorders. Together, these findings suggest a role for serotonin in gambling disorder, although the precise nature of its involvement requires further investigation.
Given a role for the mesocorticolimbic dopamine system in mediating the reinforcing properties of drugs, dopamine has been hypothesized to be involved in gambling behaviors. The maturation of the mesocorticolimbic dopamine and other systems during adolescence may in part explain the high estimates of gambling problems evidenced during this period. Some data suggest that dopamine levels may increase during gambling behaviors. However, ligand-based imaging studies involving pathological gamblers have yet to be published in peer-reviewed journals. Like with serotonin, studies examining dopamine metabolites in gambling disorder populations have generated inconsistent findings. Although one study reported alterations in dopamine metabolites suggesting increased dopamine turnover in gambling disorder, this finding was largely mitigated when controlling for cerebrospinal fluid flow rates.
During gambling activity, dopamine levels increase after longer playtimes in both recreational and pathological gamblers. Consistent with the idea that gambling and stimulants generate similar effects, priming individuals with the prodopaminergic (and pronoradrenergic) drug amphetamine was associated with an increase in the desire to gamble and reduction in the confidence to resist gambling in pathological gamblers, and pleasurable and motivational responses were positively associated with problem gambling severity. However, the dopamine D 2 -like receptor antagonist haloperidol was also found to promote gambling thoughts and behaviors. Hence, a precise role for dopamine in gambling disorder requires further investigation.
Individuals with Parkinson disease, a disorder characterized by dopamine system degeneration, have experienced gambling problems. Dopamine agonists, such as pramipexole, ropinirole, and pergolide, have been associated with impulse control disorders such as gambling disorder in Parkinson disease. Other factors, including levodopa dosage, age at Parkinson disease onset, marital status, family history of gambling problems, family or personal history of alcoholism, high levels of impulsivity, and presence of an impulse control disorder prior to Parkinson’s disease onset have also been associated with impulse control disorders such as gambling disorder in Parkinson disease in systematic, cross-sectional studies. As such, the extent to which gambling disorder in Parkinson disease reflects the pathophysiology of Parkinson disease, its treatment, a combination thereof, or other factors requires additional research.
Norepinephrine, implicated in sensation seeking and arousal, has also been investigated in the neurobiology of gambling disorder. Although healthy individuals demonstrate increased levels of norepinephrine prior to, as well as during, gambling sessions, pathological gamblers show particularly high levels of this neurochemical. The desire to start or continue gambling positively correlated with norepinephrine levels in one study of pathological gamblers. The report of altered catecholaminergic response patterns in gambling disorder subjects suggests that gambling may represent a compensatory behavior to heightened arousal levels.
The Genetics of Gambling Disorder
An elevated frequency of gambling disorder in first-degree relatives of those with the disorder suggests a genetic component to the disorder. Individuals who report gambling problems in their parents are themselves more likely to have higher scores on the South Oaks Gambling Screen; in addition, if their grandparents are also perceived as having gambling problems, these individuals may have a 12-fold higher odds of meeting criteria for gambling disorder. The heritability estimate of a gambling disorder diagnosis from the Vietnam Era Twin registry is 46%, and lifetime prevalence estimates of gambling disorder in identical twins and fraternal twins are 22.6% and 9.8%, respectively. A further analysis of this sample revealed that both identical and fraternal twins with subclinical gambling disorder symptoms were more likely to have a twin with full gambling disorder. These results support a continuity model of gambling disorder, where subclinical gambling and gambling disorder are differentiated by the number rather than the type of contributing factors. Genetic studies provide support for a familial co-aggregation of gambling disorder and other disorders, such as alcohol dependence, antisocial behaviors, and depression, with significant contributions stemming from shared genetic factors.
Molecular genetics have inconsistently implicated allelic variants. In one early study of dopamine-related genes, a D 2 dopamine receptor gene variant associated with substance dependence was found in 51% of pathological gamblers but only in 26% of controls. Individuals with the most severe pathology and comorbid substance use were more likely to carry the D2A1 gene. Altered distributions of other dopamine receptor gene variants (e.g., those encoding the D 1 and D 4 receptors) have been reported in pathological gamblers. However, these early studies have been criticized on methodological grounds, and a more recent study using a better controlled design and more thorough assessments did not replicate these findings. As such, further research is needed to identify precise molecular genetic contributions to gambling disorder.
There are also suggestive data for serotonergic and noradrenergic genetic contributions to gambling disorder. One study found that men with gambling disorder are more likely to have a shorter variant of the gene coding for the serotonin transporter. Other studies have also reported differential distributions of polymorphisms of monoamine oxidase-A–encoding genes in men with gambling disorder. Larger, genome-wide studies are needed to identify more precisely genes implicated in gambling disorder and to investigate gene-by-environment and gene-by-gene interactions.
The Neuropsychology of Gambling Disorder
To date, few studies have examined neuropsychological functioning in pathological gamblers. Initial studies suggest deficits in executive functioning—not accounted for by intellectual differences, as assessed by standard intelligence quotient tests—in gambling disorder that are similar to those evidenced in substance-dependent populations. Consistent with gambling disorder’s classification as an impulse control disorder, pathological gamblers demonstrate impairments on response-inhibition tasks. The Stroop task assesses cognitive control involving attention, conflict monitoring, and response inhibition. Participants are required to name rapidly the ink color of matched (congruent) or mismatched (incongruent) color-word pairs. On congruent trials, the word “red” may be written in the color red while on incongruent trials the word “red” may be written in blue ink and, therefore, requires that the individual responds “blue.” Not surprisingly, incongruent trials present greater difficulty as individuals are required to inhibit the prepotent reading response. Pathological gamblers show impairment on this task by producing more errors (i.e., reading the word, rather than naming the word’s color) and in taking longer to respond. Modified versions of this task, sometimes referred to as Emotional, Drug, or Gambling Stroop Tasks, use emotional, drug-related, or gambling-related words, respectively. Subjects are presented with neutral or theoretically disorder-valenced words in different-colored ink. In affected individuals as compared with healthy controls, the variant Stroop tasks tend to produce further delays and errors in processing. For example, in both recreational and pathological gamblers, when the words are theoretically more emotionally or motivationally salient rather than are neutral words (e.g., “dice” vs. “door”), a more pronounced Stroop effect is observed. Such findings suggest not only an attentional bias for disorder-related stimuli, but also a certain level of automaticity in processing.
The neuropsychological function of pathological gamblers, as compared with other subject groups, has been examined. Four groups, consisting of individuals with gambling disorder, Tourette syndrome, alcohol dependence, or no psychiatric disorder, were compared on tasks assessing executive functioning. On tasks involving response inhibition, including the Stroop Task, the three clinical groups performed significantly worse than did healthy controls, but did not differ from one another. This trend was also observed on the Wisconsin Card Sorting Task, a measure of cognitive flexibility. However, on tasks of planning and time estimation, pathological gamblers and alcohol-dependent individuals showed significantly poorer performance relative to healthy control subjects and those with Tourette syndrome.
The Iowa Gambling Task is a neurocognitive measure assessing risk/reward decision-making, where individuals can choose between different decks of cards with varying schedules of reward. Two disadvantageous decks confer high rewards, but also present even higher penalties, thereby resulting in a net loss for players. Two advantageous decks provide low rewards, but even lower penalties. Therefore, consistent selections from these decks produce an overall gain in money. Pathological gamblers and alcohol-dependent subjects demonstrated disadvantageous performance compared with the healthy control group as well as the Tourette syndrome group. These findings are consistent with prior reports that pathological gamblers show disadvantageous performance on this task. The performance profile of the gambling disorder group also showed that they responded faster, made fewer response shifts following losses, and demonstrated less conceptual knowledge about the task than did healthy controls. 35 These findings suggest an impulsive and perseverative response style, and this profile may relate to loss chasing or altered reward processing in the gambling disorder group. A separate study examining the psychophysiological correlates on the Iowa Gambling Task showed that, unlike healthy controls, pathological gamblers fail to show increases in skin conductance response or heart rate accelerations prior to making a disadvantageous choice. These alterations in psychophysiological responses suggest an impairment in risk assessment related to disadvantageous risk-reward decision-making.
Neurocognitive research findings in gambling disorder should be interpreted cautiously, as many studies do not control for comorbidity, medication status, gambling severity, or gambling type or provide comparison control groups. Gambling motivations may also be important to consider when examining neurocognitive performance in pathological gamblers. Whether individuals gamble in order to heighten arousal or relieve their dysphoric mood may relate to their performance and its underlying biobehavioral substrates. Impaired performance on some neurocognitive tasks assessing inhibition and decision-making may represent phenotypic markers in gambling disorder that have potential in predicting relapse. More research is needed to identify intermediate phenotypic or endophenotypic markers that may be used in the diagnosis and treatment of gambling disorder.
Neuroimaging studies suggest altered functioning in frontal, temporal, and limbic structures in pathological gamblers. The first published study using functional magnetic resonance imaging in gambling disorder utilized happy, sad, and gambling videotapes. While viewing the videos, participants reported the onset of an emotional (e.g., feelings of sadness) or motivational (e.g., gambling urge) response by pressing a button. During the gambling scenarios (but not the happy or sad ones), pathological gamblers showed signal decreases in frontal and orbitofrontal cortical areas, thalamus, and basal ganglia. These brain changes occurred prior to conscious awareness of an emotional/motivational response, that is, preceding the button-press. This activation pattern contrasts with those from symptom provocation studies in obsessive compulsive disorder, in which increased activation of cortical-basal-ganglionic-thalamic circuitry is observed.
During the viewing of the final portion of the gambling scenarios, when the most robust gambling stimuli were presented, pathological gamblers (relative to controls) showed less activation of the ventromedial prefrontal cortex. Subsequent studies using a functional magnetic resonance imaging Stroop task, a decision-making, and a simulated gambling task have also demonstrated relatively diminished activation of the ventromedial prefrontal cortex in association with gambling disorder. The ventromedial prefrontal cortex has been implicated in mood regulation, decision-making, and impulsivity. The ventral striatum, functionally connected to the ventromedial prefrontal cortex, has also been shown to activate less strongly in pathological gamblers. Activation in the ventromedial prefrontal cortex and ventral striatum correlated inversely with gambling severity in pathological gamblers during simulated gambling, further suggesting the relevance of these regions to clinical aspects of gambling disorder. Similar patterns of brain activations, including relatively diminished activation of ventral striatum, have been reported in cocaine-dependent subjects viewing cocaine tapes and gambling disorder subjects viewing gambling tapes, suggesting similar neural contributions to appetitive urge states across disorders. These neurobiological findings support the conceptualization of gambling disorder as a behavioral or nonsubstance addiction.
Although to date fewer than 10 neuroimaging studies examining neural correlates in gambling disorder have been published, studies using healthy controls have investigated intertemporal choice, loss aversion, and other components influencing decision-making. One functional magnetic resonance imaging study examining the neural correlates of loss-chasing behavior demonstrated increased ventromedial prefrontal cortex activation when healthy individuals tried to win back money lost on previous gambles. Loss-chasing, therefore, appears linked to brain areas involved in reward processing and raises the possibility that recreational gamblers may chase losses because they believe that winning is imminent. The extent to which these findings relate to gambling disorder requires further, direct investigation.
Few pharmacological and behavioral therapies targeting gambling disorder have been investigated with respect to their tolerabilities and efficacies. It is estimated that only 75%–12% of pathological gamblers seek formal treatment for gambling disorder. These individuals may seek treatment for various reasons (e.g., threats of spousal divorce, suicide attempts), and thus treatment-seeking pathological gamblers may differ from pathological gamblers in the general population.
Although Gamblers Anonymous is arguably the most widespread intervention for gambling disorder, questions exist regarding its effectiveness. One study reported that most individuals attend only one or two meetings and less than 10% remain in attendance after 1 year. Cognitive therapies have shown promise in the treatment of gambling disorder. One cognitive therapy targets erroneous cognitions, such as illusions of control over random events, and was found to be helpful in an initial, small, wait-list–controlled study. Following this treatment, approximately 86% of individuals no longer met gambling disorder criteria, and individuals reported greater self-efficacy and perception of control over their gambling problem. This type of therapy may also be effective in group format, and therapeutic gains appear to be maintained after 1 year. Cognitive behavioral therapy for gambling disorder identifies gambling triggers and cognitive biases, reinforces nongambling behaviors, teaches coping skills, and addresses finance management and debt settlement. Individuals receiving cognitive behavioral therapy showed greater reductions in gambling problems and time spent gambling than did those attending Gamblers Anonymous. However, both groups demonstrated improvements over time.
Other psychological interventions have been developed, including aversive therapy, imaginal desensitization, motivational enhancement, brief guided therapy, self-help workbooks, and eclectic therapies. The effectiveness of psychological interventions has been complicated by differences in assessments used to evaluate treatment outcome. However, a review of behavioral therapy outcome studies showed that these interventions are associated with significant improvement both posttreatment and after long-term follow-up when compared with no treatment. It should be noted, however, that drop-out rates in many studies approach 50%. Future studies should examine the efficacy of combining different therapies that target different cognitive and motivational aspects of gambling disorder.
Like in behavioral treatments, the evaluation of the efficacies of pharmacological therapies in gambling disorder is complicated by differences in sample sizes, trial durations, dosing strategies, trial designs, and outcome measures.
The findings of low serotonin levels in gambling disorder and blunted ventromedial prefrontal cortex to serotonergic drugs in impulse control disorders suggest that selective serotonin reuptake inhibitors could be useful therapeutic agents for gambling disorder. Several studies have demonstrated that selective serotonin reuptake inhibitors such as fluvoxamine and paroxetine are associated with short-term improvement in pathological gamblers. However, placebo-controlled trials of fluvoxamine and paroxetine have also yielded negative results. Some variability in outcome may relate to heterogeneity of pathological gamblers, and guiding selection of therapies according to presence of co-occurring disorders (e.g., selective serotonin reuptake inhibitors for individuals with co-occurring gambling disorder and anxiety disorders) may help improve treatment outcomes. Consistent with this notion, a study examining lithium in the treatment of individuals with co-occurring gambling disorder and bipolar-spectrum disorders found lithium to be superior to placebo in reducing symptoms of both gambling and mania.
Three separate studies have found opioid antagonists (naltrexone and nalmefene) to be superior to placebo in the treatment of gambling disorder. Individuals with a family history of alcoholism may be particularly responsive to treatment with an opiate antagonist. Medications targeting dopamine receptors directly (e.g., the serotonin/dopamine antagonist olanzapine) have been shown in two placebo-controlled trials not to be superior to placebo in the treatment of gambling disorder.