Pathological gambling has become an increasingly important problem, particularly as gambling opportunities have proliferated (1,2,3). The prevalence of pathological gambling has been estimated to range from .5 percent to 1.5 percent of the U.S. population, and there is some concern that prevalence is increasing among women and young persons. Gambling is legal in some form in all but two states, and in 1996 Americans legally wagered $586.5 billion in gambling activities, including pari-mutuel betting, state lotteries, casino gambling, and video lottery terminals (4).

Although "gambling mania" was described by Kraepelin (as cited by Bleuler [5]) more than 100 years ago, the disorder was not officially recognized until its inclusion in DSM-III in 1980. Pathological gambling continues to be listed as a disorder of impulse control in DSM-IV (6). The disorder is characterized by the failure to resist the impulse to gamble despite serious personal consequences. Pathological gambling is more common among men than women, but the course is thought to progress more rapidly among women than men (2,7).

It is generally acknowledged that pathological gamblers suffer high rates of mood and substance use disorders (1,8), although few investigators have actually assessed rates of these or other psychiatric conditions. For example, Linden and colleagues (9) reported that 72 percent of 25 subjects recruited from a Gamblers Anonymous chapter had experienced an episode of major depression. McCormick and associates (10) found that 76 percent of 50 inpatients admitted for problem gambling had major depression. These investigators also reported excessive rates of hypomania (38 percent).

Alcohol and drug abuse are perhaps the best documented comorbid diagnoses, with lifetime prevalence rates reported for up to 48 percent of pathological gamblers (9,10,11). Other conditions reported among pathological gamblers include panic disorder, attention-deficit hyperactivity disorder, and various impulse control disorders (9,12,13). Little has been published about comorbid personality disorders or traits. Murray (14) has observed that pathological gamblers fit no particular personality profile, but several investigators have reported abnormal personality traits in pathological gamblers based on dimensional assessments such as the Minnesota Multiphasic Personality Inventory or the Eysenck Personality Inventory (15,16). Taber and colleagues (17) reported that 20 percent of 66 inpatients with pathological gambling had personality disorders, but they did not specify which type.

In this paper we describe the findings from interviews with 30 subjects who acknowledged pathological gambling behavior. The subjects took part in structured and semistructured interviews to assess the presence of axis I and axis II comorbidity. To our knowledge, this paper represents one of the first detailed psychiatric studies of individuals reporting pathological gambling behavior who were not specifically seeking treatment.

Subjects were recruited through television and radio advertisements and newspaper press releases between June and August 1997. These advertisements invited persons "with compulsive gambling behavior" to participate in a research study. Sixty-five individuals responded to the ads; 48 subjects (74 percent) were ultimately contacted and screened by telephone using the South Oaks Gambling Screen (SOGS), a questionnaire shown to differentiate pathological gamblers from nonproblem gamblers (18).

Of the 48 respondents scoring greater than 5 points (indicating problematic gambling behavior), 30 subjects (63 percent) were eventually interviewed. All gave written informed consent before study participation according to the regulations of the University of Iowa institutional review board. Subjects received compensation ($10) for their participation.

A computer-interactive version of the National Institute of Mental Health Diagnostic Interview Schedule (DIS) revised for compatibility with DSM-III-R was administered to assess major (axis I) mental disorders (19). Subjects answered questions at a cathode ray tube terminal located within the department of psychiatry at the University of Iowa College of Medicine. Subjects were also given a version of the Minnesota Impulsive Disorders Interview (MIDI) to assess the presence of trichotillomania, pyromania, intermittent explosive disorder, kleptomania, compulsive buying, compulsive sexual behavior, and compulsive exercise (20). Axis II disorders were assessed using the Personality Diagnostic Questionnaire-IV (PDQ-IV) (21).

Other questionnaires given included the Beck Depression Inventory (22), the Maudsley Obsessive-Compulsive Inventory (23), and the Medical Outcomes Study Short Form-36 (24). The latter questionnaires were given to assess mood and obsessive-compulsive symptoms, as well as general functioning. The findings from these questionnaires will be reported elsewhere.

The sociodemographic profile of the 30 pathological gamblers is presented in T1. The mean±SD age of the subjects was 44±14 years (range, 21 to 72 years). Seven subjects (23 percent) were women. Half of the subjects were married. Eight subjects (27 percent) were single, and the rest were divorced or widowed. The majority had attended or completed college, and most were clerical workers, laborers, or technical personnel. The mean±SD reported income was $34,250±$19,691 (range, $6,000 to $80,000).

As T2 shows, the most popular gambling pastime was casino gambling, endorsed by 28 subjects (93 percent). Also popular were slot machines and other gambling machines, card games, lotteries, and horse or dog track betting. Seventeen subjects (57 percent) had gambled between $100 and $999 on any single occasion, ten subjects (33 percent) had gambled between $1,000 and $10,000, and two subjects (7 percent) had gambled more than $10,000. Responses to questions from the SOGS indicated that all subjects gambled more money than intended. Twenty subjects (67 percent) reported gambling as a current problem, and nine (30 percent) reported it as a past problem. Using credit cards, passing bad checks, and borrowing money were the most common methods of acquiring money for gambling.

Seven subjects (23 percent) indicated that gambling was familial; two subjects (7 percent) indicated that both parents had gambled too much, three subjects (10 percent) that their father had gambled too much, and two subjects (7 percent) that their mother had gambled too much.

Comorbid psychiatric diagnoses are shown in T3. Nearly two-thirds of the subjects met DIS criteria for a current disorder. The most common lifetime disorders included mood disorders (particularly major depressive disorder), substance use disorders, and anxiety disorders. Half the sample had a current depressive disorder. Only three subjects had no comorbid lifetime axis I disorder.

T4 shows the results of assessment with the MIDI. Compulsive sexual behavior, compulsive buying disorder, and intermittent explosive disorder were relatively frequent. Five of seven compulsive buyers (71 percent) but only one of five individuals with compulsive sexual behavior (20 percent) were female. Other information collected using the MIDI revealed that eight subjects (27 percent) reported being physically abused during childhood, and three subjects (10 percent) reported sexual abuse.

Results of assessment for a personality disorder via the PDQ-IV are shown in T5. Twenty-six subjects (87 percent) met criteria for at least one personality disorder, the most frequent being obsessive-compulsive, avoidant, and schizoid personality disorders. This instrument also assesses personality types under investigation that are not formal diagnoses. Both the depressive and negative personality disorders were identified and affected 15 percent and 9 percent of subjects, respectively. According to the DIS, the prevalence of lifetime antisocial personality disorder was 33 percent; based on the PDQ-IV, it was 17 percent. A consensus from the two instruments yielded a prevalence for antisocial personality disorder of 13 percent.

Our experience and that of our subjects suggest that the effects of pathological gambling are both important and underappreciated. Research has shown that the disorder is widespread and may be increasing in prevalence. Pathological gambling has become a major health concern, although its impact may not be readily apparent since many of its consequences, such as financial problems and marital discord, remain out of the public's view.

The typical subject of our study was a 44-year-old man with at least some college experience and an average income of more than $34,000. Most subjects reported a strong interest in casino gambling, and the games typically found in casinos such as craps, dice games, card games, and slot machines. State lotteries, which have become a reliable source of revenue for many states, were also popular. Subjects mainly reported betting less than $1,000 on any single day, but two reported betting as much as $10,000.

The subjects suffered substantial psychiatric comorbidity, which confirms earlier reports of a connection between pathological gambling and mood, anxiety, and substance use disorders. Although our study did not include a comparison group, data from the Epidemiologic Catchment Area survey (25) and the National Comorbidity Survey (26), as well as the Iowa community (27,28,29), indicate that the frequency of axis I and axis II comorbidity among these 30 subjects was excessive. Sixty percent of our subjects met lifetime criteria for a mood disorder, 40 percent for an anxiety disorder, and 63 percent for a substance use disorder. Rates of lifetime depression were lower than reported among pathological gamblers by Linden and associates (9) or McCormick and colleagues (10), but this may be due to differences in ascertainment. Linden and associates (9) obtained subjects through a Gamblers Anonymous chapter, and McCormick and colleagues (10) recruited subjects from an inpatient unit at a Veterans Affairs hospital.

Impulse control disorders also appear to be excessive in our sample. The data are consistent with those of Specker and associates (12), who reported an excess of both compulsive buying and compulsive sexual behavior among pathological gamblers compared with control subjects. In our study, five of the seven female gamblers were also compulsive buyers, a disorder found mainly among women (30). The data suggest that considerable overlap exists among the various impulse control disorders, and in fact, 43 percent of our sample had at least one other such disorder. We reported similar results for a group of 36 persons with compulsive sexual behavior who showed increased rates of compulsive buying, kleptomania, and pathological gambling (31).

Personality disorders were also quite frequent and were found in all but four subjects. According to PDQ-IV results, the obsessive-compulsive, avoidant, and schizoid personality disorders were the most common. Although antisocial personality disorder has been singled out as having a close relationship with pathological gambling (32), it appears that other personality types are actually more common; antisocial personality disorder may be characteristic of only a subset of pathological gamblers.

It is not readily apparent why these personality disorders are the most common, although persons with these traits may be preferentially attracted to gambling. Gambling requires a preoccupation with details (characteristic of obsessive-compulsive personality disorder), and at least some forms of gambling, such as slot machines, require little interpersonal contact (which fits avoidant or schizoid personality disorders). On the other hand, the impulsivity, recklessness, and irresponsibility characteristic of gambling may seem attractive to persons with antisocial personality disorder.

Our data are somewhat compatible with the findings of Roy and colleagues (16), who found pathological gamblers to have high neuroticism and psychoticism scores on the Eysenck Personality Inventory, indicating that they rate themselves both as nervous, sensitive, or moody and as manipulative and attention seeking. In any event, our results suggest that personality difficulties are relatively common among pathological gamblers.

The rates of physical abuse appeared to be high in our sample (27 percent), but it is similar to rates seen in outpatient psychiatric clinics (33) and to rates we have previously found among persons with compulsive sexual behavior (22 percent) (31). However, the rate of sexual abuse in the study reported here (10 percent) was lower than in our previous study of subjects with compulsive sexual behavior (31 percent) (31). This difference suggests that a history of sexual abuse may be related to compulsive sexual behavior but not to pathological gambling.

Although no family studies of pathological gambling have been conducted, our preliminary data suggest that it may be familial. Seven subjects (23 percent) reported that one or both parents were problem gamblers. Our figures are somewhat lower than those reported by Linden and associates (9), who found that 32 percent of their sample had a first-degree relative with pathological gambling. Clearly, the data indicate a need for careful family studies of this disorder.

Pathological gambling is classified in DSM-IV as an impulse control disorder, although it has been compared with addictive disorders and, more recently, with obsessive-compulsive disorder. Could pathological gambling be a form of addiction? Many similarities between pathological gambling and addictive disorders exist, including failure to control behavior and continuation of behavior despite substantial harmful consequences. The DSM-IV criteria were in fact patterned after the criteria for addictive disorders. Nonetheless, we believe that pathological gambling ought not to be considered an addiction because no external substance is ingested, nor are there physiological consequences from either the pathological gambling behavior itself or the cessation of the behavior.

Although inclusion of pathological gambling in an obsessive-compulsive spectrum implies a special relationship with obsessive-compulsive disorder (23), little evidence exists to support this association, and the types of cognitions and behaviors in pathological gamblers differ from those reported in patients with obsessive-compulsive disorder. With pathological gambling, repetitive gambling behaviors or gambling preoccupations are usually described as pleasurable and ego-syntonic. Subjects may report that they suffer the consequences of gambling later on, but gambling urges are rarely resisted. With obsessive-compulsive disorder, obsessions and compulsions are regarded as intrusive, senseless, and ego-dystonic; they are invariably resisted (34).

Pathological gamblers do not suffer high rates of obsessive-compulsive disorder, nor is there evidence of a familial relationship between the two disorders. In fact, no problem gamblers were found among the relatives of 32 subjects with obsessive-compulsive disorder in a recent family study (35). However, data from the study reported here suggest a possible link between pathological gambling and other impulse control disorders, arguing for its continued inclusion with impulse control disorders.

Our study was an attempt to understand pathological gambling and its psychiatric comorbidity. The study had several limitations. First, subjects were recruited through advertisements that may have attracted people with high levels of emotional distress. Thus the study group may not be typical of individuals with pathological gambling behavior, and our subjects may have had higher rates of impairment and distress than expected. Also, the number of subjects was relatively small, particularly the number of women. Therefore, caution should be exercised in generalizing our findings.

Pathological gambling is relatively common, yet receives scant attention in the psychiatric literature. Its appropriate classification and relationship to other disorders including obsessive-compulsive disorder, substance use disorders, or other impulse control disorders, remain unclear. Additional work is needed to establish its relationship with the availability of gambling opportunities, and to determine whether its frequency is increasing. More work will help pinpoint its risk factors, psychiatric comorbidity, family history, medical consequences, and natural history. The purpose of such research is to learn more about a troubled group of individuals so that effective treatments may be developed.

The authors are affiliated with the department of psychiatry at the University of Iowa College of Medicine, Iowa City, Iowa 52242 (e-mail, donald-black@uiowa.edu).