Psychiatric Comorbidity Associated With Pathological Gambling

Mood disorders

Three general population surveys have investigated the association of disordered gambling and psychiatric comorbidities, including major depression, dysthymia, bipolar, and suicidality. Bland and associates²¹ found higher rates of mood disorders in pathological gamblers than in non-gamblers (33.3% vs 14.2%, respectively). Rates of major depression were also higher among pathological gamblers in the sample reported by Cunningham-Williams and coworkers.⁹ These investigators also found that recreational gamblers are at greater risk for major depression and dysthymia than are those who have never gambled. Neither of these 2 surveys found a significant association between PG and mania.

The most recent of the general population surveys—and the largest to date— reported that 49.6% of persons with PG have a mood disorder.²³ This survey established that mania was the mood disorder most strongly related to PG (odds ratio [OR], 8). The OR for major depression and dysthymia were each 3.3, and 1.8 for hypomania. In the DSM-IV, criterion B for PG states “the gambling behavior is not better accounted for by a manic episode.” Thus, the clinician must determine whether the patient was gambling excessively when manic or whether his or her gambling fit the more chronic pattern of PG.

Mood disorders are also relatively common in the treatment-seeking segment of the PG population. One study found that 12% of 592 PG treatment-seekers also had a current mood disorder.²⁸ Another reported that 60% of PG subjects who had been recruited from classified advertisements met lifetime criteria for a mood disorder.¹⁵

Anxiety disorders

If we look again at our general population surveys, we find a strong association between PG and anxiety disorders. For example, Kessler and colleagues²⁴ found that 60.3% of those in their sample had any anxiety disorder: 52.2% had phobias, 21.9% had panic disorder, 16.6% had generalized anxiety disorder (GAD), and 14.8% had posttraumatic stress disorder. These authors also found that PG is temporally predicted by panic disorder, GAD, and phobia.

Petry and associates²³ reported that panic disorder with and without agoraphobia was most strongly related to PG; the odds of having phobias or GAD were significant, but less so. Cunningham-Williams and coworkers⁹ also found that the highest percentage of participants experienced panic disorder (23.3%); this was followed by phobias (14.6%), GAD (7.7%), and obsessive-compulsive disorder (OCD, 7.7%). Likewise, Bland and coworkers²¹ found that persons with PG had high rates of anxiety disorders. These authors reported lifetime rates of 26.7% for GAD, 17.7% for phobias, 16.7% for OCD, and 3.3% for panic disorder.

In their sample of 43 treatment-seeking outpatients, Ibáñez and colleagues²⁹ found a lifetime GAD rate of 7.2%. This was much lower than the 40% reported by Black and Moyer¹⁵ and the 37.5% by Specker and associates.³⁰ Black and Moyer also found rates for panic disorder and OCD of 10% each, while Specker and colleagues reported rates of 20% and 2.5%, respectively. While samples and rates differ somewhat, there remains little doubt that PG and anxiety disorders are associated.

Some investigators believe that PG falls within the obsessive-compulsive spectrum. They point to similarities between PG and OCD exemplified by persistent thoughts and urges followed by repetitive behaviors. However, there are major differences as well—OCD is unwanted, while gambling is generally perceived as plea-surable. Comorbidity studies suggest that from 2.5% to 20% of persons with PG also have OCD.^30,31 But in 2 family studies of OCD that also looked at PG, there were few data to support the existence of a relationship between these disorders.^32,33

ADHD

PG has a number of attributes in common with attention-deficit/hyper-activity disorder (ADHD), and clinical data suggest substantial overlap. Goldstein and coworkers³⁴ concluded that the electroencephalographic activation patterns to right and left brain tasks seen in 8 men with PG resembled those in unmedicated children with attention deficit disorder. Carlton and Manowitz³⁵ showed that persons with PG and those with alcoholism had excessive and comparable levels of ADHD-related behaviors in their childhood, much more so than did control subjects.

Rugle and Melamed³⁶ compared 33 non–substance-abusing persons with PG with 33 nonaddicted controls on 9 attention measures and childhood behavior questionnaires. Persons with PG performed significantly worse than controls on higher-order attentional measures; those with PG also exhibited more childhood behaviors consistent with ADHD. The authors concluded that attention deficits and the associated behavior problems are long-standing and may be a risk factor for PG.

Specker and colleagues³⁷ reported that 8 of 40 (20%) persons with PG met criteria for ADHD, and another 7 (17.5%) had symptoms that were considered subthreshhold. These authors hypothesized that ADHD may predispose to substance abuse or PG, and that gamblers with attention deficits might choose gambling activities that do not require sustained attention or concentration.

Finally, Kessler and colleagues²⁴ found in their general population survey that 13.4% of persons with PG also had ADHD.

Impulsivity, an important attribute of ADHD, is also reported to be common among persons with PG. Castellani and Rugle³⁸ evaluated 843 persons who had been admitted to an inpatient addictions unit with a primary diagnosis of PG, alcohol dependence, or cocaine abuse. Those with PG scored significantly higher than alcoholics and cocaine abusers on measures of impulsivity, such as coming to quick decisions, moving quickly from impulse to action, and lack of future planning. DeCaria and colleagues³⁹ found higher levels of impulsivity in persons with PG than in cocaine abusers, alcoholics, polysubstance abusers, and depressed patients as measured by the Barrett Impulsiveness Scale.

Impulse control disorders

PG and its comorbidity with other ICDs has not been studied as thoroughly as other Axis I disorders, although rates of ICDs appear higher in persons with PG than in the general population. Investigators have reported rates ranging from 18% to 43% for 1 or more ICD.^15,37,40

Specker and colleagues³⁷ examined rates of ICDs in a treatment-seeking sample and found increased levels of compulsive shopping and sexual behaviors, intermittent explosive disorder, and kleptomania. Black and Moyer¹⁵ found high rates of compulsive buying (23%), compulsive sexual behavior (17%), and intermittent explosive disorder (13%) in their sample. Grant and Kim⁴⁰ showed lower rates of these behaviors in a larger sample: compulsive sexual behavior (9%), compulsive buying (8%), and intermittent explosive disorder (2%).

Individuals with one ICD, such as PG, appear more likely to have a second ICD.^41-43

Personality disorders

When reviewing the literature on prevalence rates of personality disorders (PDs), the type of assessment tool used must be carefully noted. Those studies that employ self-report instruments consistently yield higher rates of PD than those that use semistructured interviews. For example, prevalence rates of PD ranged from 87% to 93% among pathological gamblers assessed with self-report instruments; those rates compared with 25% to 61% among persons who were assessed in a structured or semi-structured interview.⁴⁴

Bagby and colleagues⁴⁴ used both a self-report and a semistructured interview in their study of 204 pathological gamblers. As expected, PD prevalence rates with the self-report measure were high (92%); they were lower with the interview tool (23%). These investigators found that only those with borderline personality disorder (BPD) had consistently high and significant prevalence rates in their non–treatment-seeking samples across both types of measures. Because impulsivity and mood dysregulation are hallmarks of both BPD and PG,⁴⁵ it follows that BPD would distinguish PG from other disorders.

Fernández-Montalvo and Echeburúa46 also used a structured clinical interview in their study of 50 non–treatment-seeking pathological gamblers. BPD was the most prevalent PD at 16%; this was followed by antisocial, paranoid, narcissistic, and nonspecified disorder, which were each observed in 8% of patients. The authors also found that the presence of a PD is associated with greater gambling severity and more severe anxiety, depression, and alcohol abuse.

In a general population survey,²³ a robust association was found between PG and all the PDs studied. The odds of having any PD if one also has PG were 8.3 times greater than for the general population. The OR of having histrionic PD was 6.9; avoidant PD, 6.5l; paranoid, 6.1; antisocial, 6.0; dependent, 5.5; schizoid, 5.0; and obsessive-compulsive PD, 4.6. The borderline, narcissistic, and schizotypal types were not assessed.

Antisocial personality disorder (ASPD)—a condition marked by a pervasive pattern of poor social conformity, deceitfulness, impulsivity, criminality, and lack of remorse beginning before the age of 15—occurs at relatively high rates among those with PG disorder. Slutske and coworkers,⁴⁷ in the largest community-based study to examine this relationship, found that the odds of a lifetime diagnosis of ASPD were 6.4 times greater among persons who had a lifetime history of PG. Fifteen percent of their sample of pathological gamblers also had ASPD, compared with 2% of the comparison sample without PG.

Pietrzak and Petry⁴⁸ compared treatment-seeking pathological gamblers with and without ASPD. Those with ASPD had more severe gambling and more medical and drug-related problems; they scored higher on symptom measures of somatization, paranoid ideation, and phobic anxiety. They were also more likely to be younger, male, less educated, divorced, or separated, and to have had a history of substance abuse treatment than their non-ASPD counterparts.

Dimensional personality traits have also been reported in persons with PG. Nordin and Nylander⁴⁹ used the Temperament and Character Inventory (TCI)⁵⁰ and found that those with PG scored higher on novelty seeking and harm avoidance, and scored lower on self-directedness. This finding indicates that pathological gamblers were less responsible, purposeful, and self-acceptant, and had less impulse control than did their non-PG matches.

Recently, Forbush and colleagues⁵¹ reported TCI findings in a comparison of persons with PG and matched controls. They found high levels of novelty seeking, impulsivity, and harm avoidance, and lower levels of self-directedness and cooperativeness.

Conclusion

In summary, the literature suggests a clear connection between PG; addictive, mood, and anxiety disorders; ADHD; and Axis II disorders (particularly antisocial disorder and BPDs). Other ICDs are also frequently comorbid. While the precise mechanism behind these observations is unclear, there are several possible explanations. An Axis I disorder (eg, major depression) could be primary and lead to the secondary development of traits and behaviors found in PG. PG and the Axis I disorder could be unrelated, but they are common in patient populations and tend to co-occur and influence one another. Finally, PG and certain Axes I and II disorders may have common causes that increase their co-occurrence. For example, Slutske and colleagues⁴⁷ have reported that a genetic relationship exists among PG, substance use disorders, and ASPD.

Future investigations should consider the role of comorbid disorders on course and outcome, illness progression, and treatment. We are currently conducting a family study that will help to tease apart these etiological possibilities, and (we hope) will lead to a better understanding of the relationship among the disorders and their consequences.