A family study of pathological gambling

The cause of pathological gambling (PG) is unknown. The current study was conducted to determine whether PG is familial, and to examine patterns of familial aggregation of psychiatric disorder. To that end, 31 case probands with DSM-IV PG and 31 control probands were recruited and interviewed regarding their first degree relatives (FDRs). Available and willing FDRs were directly interviewed with structured instruments of known reliability, and best estimate final diagnoses were blindly assigned for 193 case and 142 control relatives over age 18 years.

The results were analyzed using logistic regression by the method of generalized estimating equations. The lifetime rates of PG and “any gambling disorder” were significantly greater among the relatives of case probands (8.3% and 12.4%, respectively) than among the control relatives (2.1% and 3.5%, respectively). PG relatives also had significantly higher lifetime rates of alcohol disorders, “any substance use disorder,” antisocial personality disorder (ASPD), and “any mental disorder”. “Any gambling disorder,” alcohol disorder, and “any substance use disorder” remained significant after a conservative Bonferroni correction. Interestingly, PG families were significantly larger than control families.

We conclude that gambling disorders are familial and co-aggregate with substance misuse. The data are also suggestive that PG co-aggregates with ASPD. Further research on the heritability of PG is warranted.

Introduction

Pathological gambling (PG) was first described more than 100 years ago, though it has only been included in the official nomenclature since 1980. PG is defined as persistent and recurrent maladaptive gambling behavior that involves loss of control of gambling, progressive deterioration of the disorder, and continuation despite negative consequences. Currently classified as an impulse control disorder (ICD), PG involves fundamentally pleasurable or enjoyable behaviors that are taken to extremes, contributing ultimately to disturbed marital and family life, social and occupational impairment, legal and financial problems and, in some cases, suicide. The lifetime prevalence of PG is estimated to range from 1% to 2% of adults based on epidemiologic surveys conducted in communities across the United States.

The role of heredity in PG has been relatively unexplored, though long suspected by clinicians. Walters (2001) conducted a meta-analysis of published family history information and concluded that PG runs in families, though the effect was relatively weak. The data he examined was not systematically collected, and mostly based on informal surveys. For example, Lesieur and Klein found that 5% of 182 high school students surveyed reported that one or both parents gambled “too much” and 17% of those students showed signs of problem gambling themselves. In a survey of 702 adolescents, 8.7% reported problem gambling; of these respondents, 80% reported that one or both parents gambled. An inpatient sample of persons with substance abuse or problem gambling, 39% reported that their fathers – and 3% of their mothers – were pathological gamblers.

Other mental disorders as well have been reported to be excessive among the relatives of persons with PG. In one of the few systematic attempts to collect family history information, A morbidity risk of 17% for major mood disorders and 18% for alcohol abuse/dependence among 175 first degree relatives (FDRs) of 25 pathological gamblers. 50% of 51 pathological gamblers had a parent with alcohol abuse. 33% and 24% of first-degree relatives of 24 pathological gamblers had mood disorders or alcohol abuse, respectively. We recently reported results from a study of 14 subjects evaluated on the basis of the Family History Research Diagnostic Criteria (FH-RDC); 31% of 75 PG relatives had a lifetime alcohol disorder, 19% had lifetime major depression, 5% had a lifetime drug use disorder, 8% had a lifetime generalized anxiety disorder, and 5% had an antisocial personality disorder (ASPD). Psychiatric disorders in general were more frequent among PG relatives than those reported in a sample of control relatives.

Twin studies can be very informative regarding the influence of heredity. In a small study of 21 monozygotic (MZ) and 25 dizygotic (DZ) twin pairs, moderate heritability estimates for “high action” games in DZ men but not for “low action” games, or in women for either game type. They concluded that genetic influences could be involved in the expression of gambling but differ for men and women.

Other data suggesting that PG is heritable come from a study of 3359 twin pairs who were interviewed regarding their gambling symptoms and behaviors. Heredity explained from 35% to 54% of the liability to develop the five symptoms of PG. In a subsequent analysis, these authors suggested that PG, substance use disorders, and antisocial personality disorder constituted a genetically linked “externalizing factor.” These investigators estimated that about 50% of the variation in risk for PG was accounted for by genetic factors, and that the other 50% was due to nonfamilial environmental factors unique to the individual. The externalizing factor has a high heritability (0.81) and may represent a common genetic diathesis for these conditions.

In summary, while family history studies have been methodologically weak, data suggest that PG may be associated with an increased frequency of PG, substance use disorders, mood disorders, and antisocial personality disorder in first degree relatives (FDRs). The present study was conducted to extend our knowledge regarding the familial nature of PG and to improve upon those that had preceded it. Probands with PG and controls were recruited from the community, and comprehensive family history information was collected. All willing and available FDRs were interviewed using structured assessments of known reliability, and then blindly evaluated to yield a best estimate diagnosis for each FDR. The study was undertaken to determine 1) whether PG is familial, and 2) patterns of familial aggregation of psychiatric illness in family members of PG subjects. The results are reported herein.

Discussion

The findings are consistent with a growing body of literature suggesting that problematic gambling is familial. Gambling disorders were significantly more frequent among relatives of PG than comparison probands. Our findings suggest that, at least for the purposes of future genetic studies, PG and subclinical PG be combined, a conclusion similar to that of Slutske et al. (2000) based on their analysis of 3359 twin pairs. The difference in rates between case and comparison relatives should be considered from the vantage point of prevalence rates of PG in the general population; the estimated lifetime prevalence in the comparison relatives was consistent with expectations for the general population. The findings are also in accord with our earlier pilot study, based entirely on FH-RDC results. In fact, the estimated λ (i.e., ratio of definite PG in case and control relatives, respectively) of 3.95 is similar to that for OCD, bipolar disorder, and panic disorder, and is suggestive of a heritable component. Nevertheless, the fact that a disorder aggregates in families is not necessarily indicative of a genetic etiology.

The findings also show that substance use disorders were excessive among the relatives of PG probands. Unsystematic family history studies of PG, and at least one controlled interview study suggest that the finding is valid. Consistent with our earlier study, the data are also suggestive that PG co-aggregates with ASPD. The finding is consistent with other literature linking the two disorders.

The familial overlap between PG, substance misuse, and ASPD may be explained in part by the existence of a latent externalizing factor. In further examining the Vietnam era twin registry, PG, substance use disorders, and ASPD comprised a genetically linked factor. An externalizing factor composed of antisocial behavior, disinhibited personality, and substance use disorders has high heritability and may represent a common genetic diathesis for these conditions, though he did not include PG. The concept of externalizing disorders has long been recognized among children, but has also been described in adults, and consists of disorders that place a person into conflict with others (i.e., “acting-out” disorders).

Another finding was the lack of a relationship of PG to proband gender or age at onset of PG. Though PG has long been considered primarily a disorder of men, it appears not to be sex-linked, at least in this analysis. The fact that there was no relationship with age of onset is also interesting, because several other disorders show a definite link to age at onset, and in general a greater genetic burden is associated with younger age at onset (e.g., depression, obsessive–compulsive disorder). At least for PG, this association does not appear to be true.

Interestingly, the PG probands had higher obsessionality scores as measured by the MOS than control probands, indicating a possible link with obsessive–compulsive disorder (OCD). Yet, the findings also show no familial link with OCD, though this conclusion is tentative because it was based on so few cases. Because of the infrequent comorbidity of the two disorders, and the lack of excess PG in relatives of OCD probands in family studies, whether these disorders are related remains an open question. PG probands also had higher levels of dimensionally measured depression, as assessed with the BDI, a finding we and others have previously reported, but like the trait of obsessionality, this was not associated with family risk for mood disorders. And, while PG is considered an impulse control disorder (ICD), and ICDs have been shown to be excessive among persons with PG, ICDs were not excessive among their relatives.

A finding that approached significance was the excess burden of mental disorders in general among PG families. The finding may help to explain why these families are often described as chaotic or dysfunctional. For example, surveys of pathological gamblers and their spouses have documented serious family-related issues including psychological distress, emotional/physical abuse, and maladaptive responses by the spouse. Not unexpectedly, divorce rates tend to be high. It could be that PG families are chaotic because mood and substance use disorders run in. On the other hand, one might hypothesize that having a pathological gambler in the family induces psychiatric illness in others, in response to the gambler's disturbed behavior. Likely, it is the combination of genetically transmitted illness, in addition to the psychological reaction to the pathological gambler's misbehavior, that produces dysfunctional family life.

An unexpected finding was that the families of probands with PG were larger than those of controls. Larger family size has been linked to delinquency, crime, and violence. This association may help to account for the perceived dysfunction within the families of persons with PG.