Keywords

1 Introduction

As has been well documented throughout this book, eating disorders (ED) and substance use disorders (SUD) are frequently comorbid and share many similarities with respect to clinical presentation, neurobiology, and psychosocial risk factors (see Chaps. 11 and 12). Each has features seen in individuals with impulse control disorders (ICD) often referred to as “behavioral addictions” (BA) and process addictions. Given the growing evidence that SUD and at least some behavioral addictions, such as disordered gambling, may be more similar than different (Karim & Chaudhri, 2012; Leeman & Potenza, 2012), it should be no surprise that high rates of ICD are seen in individuals with both ED and SUD. For clinicians, knowing the clinical manifestations and treatments of ICD is essential because the severity and clinical course of ED and SUD are much worse when both disorders are comorbid with BA (Fernández-Aranda, Pinheiro, et al., 2008). In addition, ICD can be chronic, relapsing conditions with high rates of morbidity and mortality if not treated adequately (Grant, Levine, Kim, & Potenza, 2005; Grant, Potenza, Weinstein, & Gorelick, 2010).

In this chapter, the terms ICD and BA will be employed since both are used by mental health clinicians, experts, the literature we cite, and in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5: American Psychiatric Association, 2013) itself to characterize the disorders being reviewed in this chapter (Karim & Chaudhri, 2012). In addition, a growing body of scientific evidence supports considering these terms as similar and overlapping because individuals suffering from these disorders have been shown to exhibit clinically, in neurotransmitter systems (American Psychiatric Association, 2013, p. 478), and on neuropsychological measures both impulsivity and the compulsive behavior characteristic of addictions.

We will first give an overview of the concept of BA and then describe some changes to them in DSM-5. After discussing the epidemiology of BA in the general population and among individuals with psychiatric illness, we will discuss some of the BA most commonly seen in individuals with ED and SUD. We initially examine one syndrome of clinical relevance in ED and for which there is good data on diagnostic validity and/or treatment: kleptomania (Schreiber, Odlaug, & Grant, 2011). Another BA that is frequently observed in ED patients is compulsive buying disorder, which is discussed in detail in Chap. 19. We then focus on two BA frequently seen in and with important implications for treatment of ED but for which there is much more controversy regarding their status as disorders and, as a result, for which there are virtually no evidence-based treatments: hypersexual behavior or sexual promiscuity (Fichter, Quadflieg, & Rief, 1994), and Internet addiction (Weintraub, Dunn, Yager, & Taintor, 2011). We then review gambling disorder, a DSM-5 disorder which is highly comorbid with and very similar to SUD (American Psychiatric Association, 2013). We go on to summarize how BA impact the presentation and clinical course of ED and SUD. Intermittent explosive disorder, which will not be as thoroughly reviewed, is also commonly seen in SUD (American Psychiatric Association, 2013, p. 469). We then discuss how treatments of individuals with comorbid ED, SUD, and ICD need to be tailored to address the severe psychopathology seen in these patients. We conclude with future directions in research and treatment for this population.

Although most of the BA to be reviewed are not DSM-5 disorders, they are of tremendous clinical significance in individuals with ED and/or SUD and have the same type of clinical features seen in individuals with DSM-5 BA. They have also been shown to be very prevalent in other medical disorders such as Parkinson’s disease (Weintraub, Koester, et al., 2010; Weintraub, Siderowf, et al., 2006), and a growing body of research has shown that they may involve the same type of dysregulation in brain dopamine circuits underlying all addictions. As an example, in Parkinson’s disease, treatment with dopamine agonists has been associated with the development of sexual addiction, compulsive buying, and compulsive gambling, all of which are quite common in ED and/or SUD (Weintraub, Siderowf, et al., 2006). Because of this new knowledge gained from the recent increased interest and research in BA, it will likely be possible in the future to offer more evidence-based statements regarding their diagnostic validity, epidemiology, etiology, and treatment as they pertain to the management of ED and SUD.

2 What Are Behavioral Addictions?

A major change in the conceptualization of mental disorders has been ushered in with the release of DSM-5, in which gambling disorder (GD) (formally known as pathological gambling) has been placed into the Substance-Related and Addictive Disorders section (American Psychiatric Association, 2013). This new classification represents the formal validation of the view, long held by many experts in the field and now supported by considerable research data, that compulsive behavior not involving exogenous psychoactive substances can be characterized as an addiction (Clark & Limbrick-Oldfield, 2013). A major implication of including this BA in the same diagnostic category as substance-related disorders is that similar approaches to assessment and treatment may be clinically indicated for both GD and SUD (Fong, Reid, & Parhami, 2012; Goodman, 2008; Karim & Chaudhri, 2012; Potenza, Sofuoglu, Carroll, & Rounsaville, 2011). In this context, a unifying feature of BA is the repeated engagement of a particular behavior that once produced a pleasurable feeling but that has become compulsively irresistible and disrupts social, educational, or vocational functioning (Grant et al., 2010). In most respects, such patterns of behavior are quite similar to those seen in individuals with SUD.

Coincident with this change in DSM-5 are the creation of a peer-reviewed arena for the publication of research in this domain, the Journal of Behavioral Addictions (Demetrovics & Griffiths, 2012), and the redefining of the term “addiction” by the American Society of Addiction Medicine to include non-substance addictions (Smith, 2012). Other major shifts and changes in the classification of a number of disorders in which impulsivity is a prominent feature are described and discussed below.

2.1 Clinical Characteristics of Behavioral Addictions

Poor impulse control occurs in many psychiatric disorders not classified as ICD, including some subtypes of ED and most SUD. Moreover, impulsivity is often an explicit diagnostic criterion in these disorders or is included in the syndromal name as in the hyperactive-impulsive subtype of attention-deficit/hyperactivity disorder (ADHD). Most pointedly, the DSM-IV-TR section on ICD is called “Impulse-Control Disorders Not Elsewhere Classified.” The introduction to that section states “This section includes disorders of impulse control that are not presented as part of the presentation of disorders in other sections of the manual (e.g., Substance-Related Disorders, Paraphilias, Antisocial Personality Disorder, Conduct Disorder, Schizophrenia, and Mood Disorders may have features the involve problems of impulse control)” (American Psychiatric Association, 2000, p. 663). With regard to personality disorders, in both DSM-IV-TR and DSM-5, one of the criteria for borderline personality disorder is “impulsivity in at least two areas that are potentially self-damaging (e.g., spending, sex, substance abuse, reckless driving, binge eating).” Similarly, one of the criteria for antisocial personality is “impulsivity or failure to plan ahead.”

There is clearly a continuum of impulsivity in psychiatric disorders. In individuals with disordered eating, for example, impulsivity is much greater in those meeting criteria for disorders involving binging and/or purging (i.e., anorexia nervosa binge-purge subtype [ANBP], bulimia nervosa [BN], and binge eating disorder [BED]) than for the restrictor subtype of anorexia nervosa (ANR) (Brewerton, Hand, & Bishop, 1993; Cassin & von Ranson, 2005; Fernández-Aranda, Pinheiro, et al., 2008).

The ICD or BA can be conceptualized as disorders in which impulsivity is the predominant psychopathology in affected individuals. As defined in the DSM-IV-TR, “The essential feature of Impulse-Control Disorders is the failure to resist an impulse, drive, or temptation to perform an act that is harmful to the person or others.” In addition, “the individual feels an increasing sense of tension or arousal before committing the act and then experiences pleasure, gratification, or relief at the time of committing the act” (American Psychiatric Association, 2000, p. 663).

2.2 Changes in DSM-5

In the DSM-5, a new section is introduced, “Disruptive, Impulse-Control, and Conduct Disorders” (American Psychiatric Association, 2013). Although individuals suffering from them all exhibit deficits in controlling their emotions and behavior, what distinguishes these from other DSM-5 disorders in which there are problems with impulsivity is that “they are manifested in behaviors that violate the rights of others (e.g., aggression, destruction of property) and/or bring the individual into significant conflict with societal norms or authority figures” (American Psychiatric Association, 2013, p. 461). They include most of the disorders in the DSM-IV-TR category “Impulse Control Disorders Not Elsewhere Classified (NOS)”: intermittent explosive disorder, kleptomania, and pyromania. Other disorders in this new category include antisocial personality disorder and the following three DSM-IV-TR childhood disorders: oppositional defiant disorder, conduct disorder, and disruptive behavior (NOS). The other two disorders from the DSM-IV-TR impulse control NOS section have been moved elsewhere, the new DSM-5 gambling disorder to the section on SUD, as mentioned above, and trichotillomania to “Obsessive-Compulsive and Related Disorders.” But as a further illustration of the complexities and challenges involved in categorizing disorders, a decision was made to place ADHD, one of whose most prominent features is impulsivity, into a new category called “Neurodevelopmental Disorders.”

2.3 Epidemiology and Comorbidity of Behavioral Addictions

ICD are very common in the general population. In the National Comorbidity Survey Replication (NCS-R), 12-month and lifetime prevalence of any disorder with prominent impulsivity were 9 % and 12 %, respectively (Kessler, Berglund, et al., 2005; Kessler, Chiu, Demler, & Walters, 2005). A much larger and more recent study of 35,000 Americans found that 17 % of the sample was impulsive (Chamorro et al., 2012). This impulsive subgroup had significantly higher rates of psychopathology, functional impairment, dangerous behavior, and a history of trauma. About 80 % had a psychiatric disorder and were more likely than non-impulsive subjects to have attempted suicide, engaged in domestic violence, started fights, or driven recklessly.

In addition, multiple studies in the USA and abroad have shown high rates of ICD in individuals with psychiatric illness. Prevalence rates range from 20 to 30 % with the BA commonly seen in ED and SUD among the most common (Grant et al., 2005; Müller et al., 2011). These include compulsive buying (6–10 %), kleptomania (8 %), pathological gambling (7 %), pathological Internet use (3–5 %), and nonparaphilic compulsive sexual behavior (2–3 %). These studies also found that despite being frequently comorbid, BA are often underdiagnosed by clinicians.

2.4 Prevalence and Clinical Significance of Behavioral Addictions in Eating Disorders

In the DSM-5, BED has been officially recognized as a distinct major ED, along with AN and BN. For each of these conditions, different patterns of comorbidity with BA are seen. In general, individuals with ANBP, BN, and BED have higher rates of not only SUD but also impulsivity and ICD compared to those with the ANR (Brewerton et al., 1993; Cassin & von Ranson, 2005; Fernández-Aranda, Pinheiro, et al., 2008).

Multiple studies have shown high rates of ICD in ED. The earlier cited NCS-R found all the major ED to be highly comorbid with ICD with rates of 64 %, 44 %, and 31 % for BN, BED, and AN, respectively (Hudson, Hiripi, Pope, & Kessler, 2007). In the past decade, two large international studies have found that individuals with ED have high rates of ICD (Fernández-Aranda, Jiménez-Murcia, et al., 2006; Fernández-Aranda, Pinheiro, et al., 2008). The prevalence of ICD in the BN subgroup in the 2006 study ranged between 20 and 25 %, and in the 2008 study, one out of six subjects with any ED met criteria for an ICD. In the 2006 study, the BN + ICD group had significantly higher rates of psychopathology, including SUD, than the BN-only group and a comparison cohort with pathological gambling. In the 2008 study, the ED + ICD group, compared to the ED-only group, engaged in significantly greater use of laxatives, diuretics, and fasting to control weight. On psychological assessment, this comorbid group had significantly greater body image disturbance, cognitive impulsivity, and other comorbid psychiatric disorders, including borderline personality. The authors concluded that if untreated, comorbid ICD could lead to a more difficult clinical course with less likelihood of recovery in individuals with ED (Fernández-Aranda, Pinheiro, et al., 2008).

Of particular concern, high levels of impulsivity are associated with increased risk for self-harm in individuals with ED, which have rates of attempted suicide comparable to those seen in major depressive disorder (Bulik, Sullivan, & Joyce, 1999). Multiple studies have found that this already elevated risk for suicide attempts and self-harm is increased further in ED patients with high levels of impulsivity (Corcos et al., 2002; Stein, Lilenfeld, Wildman, & Marcus, 2004).

3 Behavioral Addictions Frequently Comorbid with Eating Disorders

3.1 Kleptomania

Shoplifting is an extremely common behavior with US prevalence rates of about 14 % in men and 8 % in women (Hoertel, Dubertret, Schuster, & Le Start, 2012). Some experts report that about 5 % of shoplifters suffer from kleptomania, a DSM ICD characterized by recurrent inability to resist impulses to steal (Dell’Osso, Altamura, Allen, Marazziti, & Hollander, 2006), while more recent data indicates the range of shoplifters with kleptomania to be 4–24 % (American Psychiatric Association, 2013). Individuals with kleptomania are not driven by any type of financial incentive when engaging in this behavior and often feel ashamed of what they have done. Affected individuals experience enormous tension before engaging in stealing and a sense of relief after it is over.

Although there is limited data on kleptomania’s prevalence and clinical course, it is believed to be rare with prevalence estimates below 1 % (Dell’Osso et al., 2006; Schreiber et al., 2011), and it is thought that some patients experience problems only intermittently, whereas others show a chronic course (American Psychiatric Association, 2013). About two-thirds to three-fourths of affected individuals are female (American Psychiatric Association, 2013; Grant, Odlaug, Davis, & Kim, 2009). There are high rates of comorbidity with mood, anxiety, and SUD, which are also common in first-degree relatives (Schreiber et al., 2011). It is also commonly comorbid with ED (American Psychiatric Association, 2013; McElroy, Pope, Hudson, Keck, & White, 1991).

Affected individuals can suffer significant functional impairment including being arrested and spending time in jail (Grant & Kim, 2005). One study found that about two-thirds of individuals with kleptomania had been arrested and 20 % incarcerated over a 6-year period (Grant, Odlaug, et al., 2009).

No evidence-based pharmacological or psychosocial treatments for kleptomania are currently available due to a paucity of research in this area (Grant, Kim, & Odlaug, 2009). The first double-blind controlled medication trial to demonstrate efficacy found that administration of naltrexone resulted in significant reductions in stealing urges and behavior (Grant, Kim, et al., 2009).

3.2 Hypersexual Behavior (Sexual Promiscuity)

Although not felt to justify inclusion as a new, stand-alone disorder in the DSM-5, hypersexual behavior is common in the general population and associated with significant distress and functional impairment. Present in many psychiatric disorders, sexual promiscuity is also commonly seen in individuals with the binging and purging subtypes of ED. In a Japanese study of more than 230 ED patients, 22 % of those with BN, and 16 % of those with ANBP reported having sex with a person they did not know well compared to 5 % of a comparison control group. Individuals with ANR reported the lowest rates of any group (2 %) (Nagata, Kawarda, Kiriike, & Iketani, 2000).

Sexual promiscuity in patients with BN often denotes more severe psychopathology. Two lines of distinct but overlapping research have examined this issue, one involving a highly impulsive subgroup of bulimics called “multi-impulsive bulimics” and the other exploring the role of comorbid borderline personality as a potential mediator or contributor to sexual promiscuity in women with BN (Wonderlich, Myers, Norton, & Crosby, 2002).

Multi-impulsive bulimia (MIB), named by Lacey (1993), refers to patients with severe impulsivity in many domains, more severe psychopathology, and poorer response to treatment. According to one of the most widely used diagnostic systems for this syndrome, affected individuals must have exhibited at least three of the following six impulsive behaviors: (1) suicide attempts; (2) severe self-harm; (3) shoplifting, excluding food; (4) alcohol abuse; (5) drug abuse; and/or (6) sexual promiscuity (Fichter et al., 1994). A more recent study validated these criteria in a study of 125 women with BN (Myers et al., 2006). The MIB group had a higher incidence of anxiety disorders, child abuse, and daily self-damaging behavior (including sexual promiscuity) than the non-MIB group. A high proportion of the sample, 44 %, met criteria for MIB.

Borderline personality disorder (BPD), in which there is frequently impulsive sexual behavior and many other features of MIB, is common in individuals with ED, particularly those with ANBP and BN. In a meta-analysis of studies examining the relationship between personality and ED, the rate of comorbid BPD in ED was at least 20 % and associated with greater overall psychopathology and poorer response to treatment although not increased severity of ED pathology (Cassin & von Ranson, 2005). MIB and BPD are also frequently comorbid and appear to overlap. In an earlier cited study, ED patients with comorbid ICD and other evidence of impulsivity were significantly more likely than those without ICD to have BPD (23 % versus 11 %) (Fernandez-Aranda et al., 2008). Conversely, other research has shown that comorbid BPD increases the risk for MIB (Nagata et al., 2000).

3.3 Internet Addiction

Although not a stand-alone diagnosis in the DSM-5, problematic and/or excessive Internet use is a major clinical problem worldwide that causes enormous suffering, functional impairment, and lost productivity (Weintraub et al., 2011). The lack of empirical support for this syndrome as a disorder in its own right has resulted in the creation of a subtype of Internet addiction called “Internet gaming disorder” and listed in Section III of the DSM-5 as a condition requiring further study.

What is indisputable, however, is that the Internet serves as a vehicle for many psychiatrically ill patients to engage in unhealthy behaviors that exacerbate their disease. For patients with ED, the availability of the Internet has had mixed effects. A concerning example is that in contrast to websites that provide positive help and psychoeducation about ED (“pro-recovery sites”), a larger number of pro-eating disorder websites (“pro-ANA”) actually support eating disordered lifestyles and values and provide information and group discussion on how to maintain unhealthy low weight and pathological behaviors (Wilson, Peebles, Hardy, & Litt, 2006) through the use of virtual communities.

Wilson and colleagues (2006) surveyed 700 females with ED (ages 10–22) and their parents about their knowledge and use of ED-related websites. Approximately 50 % of parents were unaware of pro-ANA websites and that their children were visiting them (Wilson et al., 2006). More than one-third of the patients visited pro-ANA sites, and about 40 % visited both pro-ANA and pro-recovery sites. Of note, 96 % of those who visited pro-ANA sites reported learning new techniques for losing weight or purging, and, even more astonishing, almost 50 % of those who visited pro-recovery sites indicated that they also learned new weight loss techniques. Visiting these websites resulted in other adverse consequences: spending less time on schoolwork, longer duration of illness, more frequent hospitalizations, and spending significantly more time on the Internet each day.

Another study of 60 females with ED found positive associations between compulsive Internet use and compulsive Internet buying, both of which showed significant positive correlations with emotional lability and measures of impulsivity (Claes et al., 2012). Conversely, individuals with Internet addiction may have elevated rates of disordered eating. In a survey of more than 2,000 Chinese college students, those with Internet addiction reported significantly elevated binge eating, concerns about weight, and weight change compared to those without problematic Internet use (Tao, 2013).

4 Behavioral Addictions That Are Often Comorbid with Substance Use Disorders

SUD and BA are united by the core feature of an individual’s recurring failures to resist the temptations, impulses, or drive to engage in particular behaviors known to have harmful consequences (Grant et al., 2010). Similarly, both SUD and BA tend to have similar ages of onset, typically emerging during adolescence or early adulthood, and higher rates of both classes of disorders among younger people (Goodman, 2008). These conditions tend to be chronic with high rates of remittance and relapse, as well as relatively high rates of recovery without formal treatment (Grant et al., 2010). While well known in SUD, those suffering from BA also tend to report physiological symptoms including craving; marked tolerance with repeated exposure; an increase in the behavior is necessary to sustain pleasurable effects; and symptoms of discomfort occurring with abrupt cessation of the behavior (Goodman, 2008). Finally, Goodman (2008) has noted that as individuals begin to effectively manage symptoms of SUD or BA, novel or dormant addictive behaviors tend to materialize or intensify.

These overlaps between SUD and BA are likely due to shared neurobiological underpinnings that are strongly associated with reward and reinforcement. Olsen (2011) notes that stimuli that are inherently reinforcing, such as seeing delicious food on a plate, or sexual behavior, activate the same neural regions as gambling, shopping, and playing video games. These regions are believed to include the dopaminergic rich mesocorticolimbic system and extended amygdala (Volkow, Fowler, & Wang, 2004). Further, these neural areas are also activated by drugs of abuse, in particular, cocaine, amphetamine, alcohol, nicotine, and tetrahydrocannabinol (Koob, 1999). A predominant theory regarding the mechanisms driving drug addiction entails the manner in which particular substances that users initially take voluntarily because of their reinforcing properties transition into substances that ultimately drive compulsory behaviors commonly seen in substance dependence. Everitt and Robbins (2005) argue that the switch of a substance from having reinforcing properties to one that users must habitually consume indicates a corresponding neuroplasticity of cortical and striatal structures, specifically, loss of prefrontal cortical control of the behavior to striatal command. It is now believed that BA induce similar neuroadaptations of the same structures (Olsen, 2011), in particular pathological gambling (Potenza, 2013). These observations suggest the utility of considering SUD and BA together.

4.1 Gambling Disorder

With the first major revision of the DSM in nearly 20 years, the disorder formerly known as pathological gambling (PG) has undergone several changes for its new appearance in DSM-5 (American Psychiatric Association, 2013). In addition to its being moved into the section containing SUD, its name has been changed to gambling disorder (GD) as the term “pathological” was seen to carry a negative valence and has become antiquated (Petry et al., 2013). In this chapter, PG and GD will be used interchangeably. Conceptually, we will follow the DSM-5 criteria for the condition and refer to it as GD; however, the vast majority of the research in this area refers to the condition as PG. More importantly, the diagnostic criteria of GD have changed; chief among them is the removal of the criterion regarding the committing of illegal acts, such as fraud, theft, or embezzlement in order to finance gambling. Individuals whose gambling habits require illegal acts to fund their gambling rarely fail to endorse other diagnostic criteria of the condition (Zimmerman, Chelminski, & Young, 2006). It is unlikely that removing this criterion will result in a loss of diagnostic sensitivity (Petry et al., 2013). The threshold for diagnosis of this disorder has also changed (American Psychiatric Association, 2013). Out of the nine criteria, a diagnosis is permissible if four are endorsed, instead of the DSM-IV-TR threshold of five items. Several studies have demonstrated superior diagnostic sensitivity when four items are used (Jimenez-Murcia et al., 2009; Stinchfield, Govoni, & Frisch, 2005). Obviously, increased detection of GD will allow more sufferers to get into treatment.

The DSM-5 criteria for GD include two major categories: one regarding recurring maladaptive gambling and a second noting that such behavior cannot be exclusively accounted for during a manic episode (American Psychiatric Association, 2013). Several of these criteria are very similar to SUD, including a preoccupation with the behavior, the need to invest larger and larger amounts of money to achieve the same thrill, repeated and unsuccessful attempts to reduce or abstain from the behavior, feelings of irritability or restlessness during attempts to cut back, gambling during times of stress or as a means to escape problems, and attempts to conceal the extent of the behavior from others (American Psychiatric Association, 2013). The remaining features include significant disruption in social, vocational, or educational functioning because of gambling, a tendency to spend more money than intended to compensate for losses, and/or the reliance on a monetary bailout from others because of a financial catastrophe created by gambling (American Psychiatric Association, 2013). It should also be noted that the term “problem gambling” is frequently seen in the literature. This label tends to be applied to gambling behavior that causes problems for an individual, but which is below diagnostic threshold (Petry et al., 2013). Typically, any individual that endorses a single criterion listed in the DSM could be exhibiting problem gambling.

Unlike patients suffering from SUD, who often present with physical manifestations of their addictions, those suffering from GD have no outward signs of their addiction. For example, individuals intoxicated by alcohol show slurred speech and ataxia; inhibitions are lowered; behavior may become erratic. Individuals under the influence of methamphetamine may show akathisia. Intravenous drug users have scar tissue or distinct skin lesions. Metabolites in the urine and/or serum blood levels of most drugs of abuse can be detected. Gambling disorder, however, has no outward signs and has been called the “hidden addiction,” making it harder to detect than SUD (Phillips, 2005). And rarely do these patients present to treatment; one study found that only roughly one in ten with the disorder will seek help (Slutske, 2006).

Interestingly, almost half of the respondents with PG in the National Comorbidity Survey Replication reported being under the care of a mental health professional at some point for a mental health disorder, yet none reported receiving treatment for the gambling problem (Kessler et al., 2008). Detection of GD may also be difficult since nonproblem gambling has such high base rates in the USA. One national survey of almost 5,000 individuals found that 75 % participated in various activities where wagers were placed on an unknown outcome (Welte, Barnes, Tidwell, & Hoffman, 2011). Such bets ranged from church raffles and card games in basements to wagers placed on sporting events and attending casinos (Welte et al., 2011). Even though mere endorsement of intravenous drug use by a patient may signal to a clinician the presence of a possible opioid addictive disorder, the same is not true with respect to GD for individuals presenting for treatment who report betting money on the outcome of the Super Bowl, for example.

While identification of GD is likely to improve with the DSM-5 criteria, it is not thought that the prevalence of the condition will rise (Petry et al., 2013). Research regarding prevalence rates are based on the earlier DSM-IV-TR PG criteria are variable. Kessler et al. (2008) estimated the adult 12-month PG prevalence at 0.3 % and lifetime prevalence at 0.6 %. Blinn-Pike, Worthy, and Jonkman (2007), however, determined a nearly 8 % lifetime prevalence when studying college students. Barnes, Welte, Hoffman, and Tidwell (2009) found that 10 % of adolescents and young adults reported having three or more gambling problems in the past year. Ultimately, a thoughtful meta-analysis by Sussman, Lisha, and Griffiths (2011) estimates the adult 12-month prevalence of PG to be 2 % in the USA. Generally, more men suffer from PG than women (Desai & Potenza, 2008; Petry, Stinson, & Grant, 2005).

The finding of higher rates of PG in younger adults is a consistent observation. This is not surprising as some studies report that as many as 90 % of adolescents have engaged in gambling (Gupta & Derevensky, 2000), with an average age of first wager being 12 years old (Wilber & Potenza, 2006). In one study of more than 1,600 adolescents in Connecticut, 69 % reported gambling before their 12th birthday (Rahman et al., 2012). Children engage in a wide range of gambling behaviors, including buying lottery tickets, playing dice and card games, betting on sporting events, as well as entering casinos despite being under age (Rahman et al., 2012; Wilber & Potenza, 2006). Given that gambling is quite prevalent, is more socially acceptable than substance use in minors, and does not cause intoxication, it is no surprise that most children will have an experience with gambling before they initially use nicotine, alcohol, or other drugs of abuse (Wilber & Potenza, 2006). Experiences with gambling earlier in life may account for the finding that a majority of individuals develop PG earlier in their lives than other addictions. Jimenez-Murcia and her colleagues (2009) found that 45 % of their 904 PG patients reported age of onset of this disorder before the end of their 20s. Kessler et al. (2008) found similar results, with a median age of onset of the disorder at age 23, also noting that individuals with PG tended to start gambling at significantly earlier ages than those who do not have PG.

Blaszczynski and Nower (2002) constructed a pathways model of problem and pathological gambling that is useful when conceptualizing PG. Their model captures the complexity of this disorder and acknowledges its heterogeneous and multidimensional nature, proposing that there are clinically distinct subgroups of sufferers. While all share core symptoms, Blaszczynski and Nower (2002) argue that Pathway 1 of PG consists of “behaviorally conditioned gamblers” who engage in pathological gambling, but are otherwise free from premorbid psychopathology and SUD and are easily treatable. This group tends to have the least severe gambling habits and may even manage reasonable gambling behavior following treatment. Pathway 2 is a second group of “emotionally vulnerable” individuals, characterized by those with preexisting mood disorders and who possess limited problem-solving abilities and poor coping strategies, often coming from disordered families. This group is believed to engage in gambling as a means of emotional escape and is more difficult to treat, as premorbid conditions must also be addressed. Finally, Pathway 3 is the group with the most severe form of PG. It includes “antisocial impulsivist” problem gamblers, who have high rates of impulsivity, drug and alcohol use, criminal behavior, and poor interpersonal abilities. They tend to be the most difficult to treat, it is suspected, because of underlying neurological dysfunction. In summary, this model suggests that some gamblers develop PG due to exogenous factors (Pathway 1), some as a maladaptive coping strategy (Pathway 2), and some, a more severe group, due to endogenous factors (Pathway 3).

Regardless of the pathway, PG can be quite impairing. Individuals with PG have higher rates than nonproblem gamblers of bankruptcy, contacts with law enforcement, divorce, incarceration, and a lower quality of life (Gerstein et al., 1999; Grant & Kim, 2005). Those suffering from PG have been known to have high rates of attempted suicide (Kausch, 2003; Potenza, Kosten, & Rounsaville, 2001), although several authors caution against a causal link between the two (e.g., see Newman & Thompson, 2003). Problem gambling may be a risk factor for intimate partner violence (Muelleman, DenOtter, Wadman, Tran, & Anderson, 2002). Further, those with problem gambling are far more likely to have deleterious health conditions, particularly cardiac and hepatic conditions, than those without a gambling addiction (Morasco et al., 2006). Additionally, its rate of comorbidity with SUD is high. Lorains, Cowlishaw, and Thomas (2011) recently completed a meta-analysis based on general population studies of conditions that are frequently comorbid with GD and found the highest rates to be for nicotine dependence, followed by alcohol use disorder, and then other SUD. Additionally, GD is also highly comorbid with major depression, anxiety disorder, and antisocial personality disorder (Lorains et al., 2011).

5 Gambling Disorder and Specific Comorbid Substance Use Disorders

5.1 Nicotine Dependence

Nicotine dependence co-occurs with PG in very high rates, ranging from 76 % (Cunningham-Williams, Cottler, Compton, & Spitznagel, 1998) in a US sample to about 35 % in a South Korean study (Park et al. 2010). Grant, Desai, and Potenza (2009) speculate that high rates of nicotine dependence and PG co-occur for a variety of reasons, including a genetic predisposition for impulsivity, a tendency for nicotine to facilitate the reinforcing effects of dopamine in the presence of stimuli eliciting such effects, the cognitive-enhancing effects of smoking enhancing experiences involving wagering, and social factors involved in gambling and smoking in groups. Potenza and his colleagues (2004) note that problem gamblers reporting daily tobacco use are more likely to endorse depression and suicidal ideation because of their gambling habit, as well as higher rates of gambling-related arrests, SUD, and engagement in mental health treatment. Grant, Kim, et al. (2009), however, note that even those who are not dependent on nicotine, but who have problematic gambling, still show high rates of SUD and other psychopathology.

5.2 Alcohol Use Disorders

Alcohol abuse and dependence also co-occur with PG at very high rates. Petry et al. (2005) examined a sample in excess of 43,000 people living in the USA and found that of those meeting criteria for PG, an astonishing 73 % also met criteria for alcohol abuse or alcohol dependence (collectively known as “any alcohol use disorder”). Other groups have found lower but still very elevated rates of 44 % (Cunningham-Williams et al., 1998) and 18 % (Welte, Barnes, Wieczorek, Tidwell, & Parker, 2001). As a result, clinicians treating GD are obligated to also assess for alcohol use disorders (AUD). This may be particularly true for patients from higher socioeconomic statuses, as Welte et al. (2001) found that sufferers of PG who fell into higher SES strata were more likely to be dependent on alcohol. It should also be of concern to those working with GD patients in clinical settings that comorbid AUD and PG have been associated with increased rates of intimate partner violence (Brasfield et al., 2012).

Given the concerning frequent co-occurrence of problematic alcohol use and gambling, it is desirable to understand what links these conditions. However, testable models regarding why GD and AUD co-occur remain elusive (Grant, Kushner, & Kim, 2002). Potenza (2008) makes a convincing argument for overlapping cortical systems that predispose individuals to develop gambling problems and excessive alcohol use. Additionally, Fischer and Smith (2008) note that a specific personality type may represent the expression of possible overlapping brain systems; individuals who are high in “trait urgency” (a form of impulsivity) tend to act rashly when distressed. This tendency is strongly associated with problem drinking in both genders, as well as problem gambling in men and binge eating in women.

5.3 Illicit Drug Use

As with AUD, GD is frequently comorbid with illicit drug use. Petry et al. (2005) found that 38 % of those with PG also endorsed drug abuse or dependence. Very high rates of nicotine dependence, AUD, and cannabis use occur in those meeting criteria for GD. PG and problem gambling also occur with relatively high frequency in samples of those with drug use disorders, ranging from 8 to 15 % of those using cocaine; similar rates are found in opioid-dependent individuals engaged in methadone maintenance (Petry & Champine, 2012). Interestingly, in a study measuring PG in treatment-seeking substance abusers, those scoring highest on a pathological gambling measure were those abusing cannabis (Toneatto & Brennan, 2002); 24 % of cannabis abusers scored in the critical range on a gambling screen, a far higher rate than seen in those seeking treatment for alcohol (4 %), cocaine (12 %), or opioid abuse (5 %). Leeman and Potenza (2012) argue that commonalities between those with PG and SUD on neuropsychological tasks (particularly those involving impulsivity, compulsivity, and risk/reward decision making) suggest that both groups share dysfunction in the ventromedial prefrontal cortex and striatum, as well as in monoamine, glutamate, and opioid neurotransmitter systems.

Those with both GD and a SUD are likely to have more severe symptoms of disordered gambling, with onset of problem gambling earlier, and tend to be resistant to treatment. As GD is an “invisible addiction” and is not typically the presenting problem when seeking treatment, clinicians are advised to screen for GD in all of their SUD patients. Patients who are first known to have GD are in turn best served by also being screened for co-occurring drug and alcohol problems.

6 Gambling Disorder and Comorbid Eating Disorders

PG and ED are rarely comorbid, despite both being highly comorbid with SUDs. In a 2013 study, Jiménez-Murcia et al. examined 1,681 ED patients and found only 25 had co-occurring gambling disorder. In an investigation of 94 individuals with BED, only a single patient met criteria for both an ED and PG (Yip, White, Grilo, & Potenza, 2011). However, it should be noted that individuals who report symptoms of GD and who do not meet full clinical criteria may have increased lifetime rates for ED and disordered eating (von Ranson, Wallace, Holub, & Hodgins, 2013).

7 Treatment of Gambling Disorder

Many PG patients recover without formal intervention (Slutske, 2006). Those with both PG and a SUD, however, have worse outcomes (Ladd & Petry, 2003). Treating GD is similar to treating other addictions, and many therapies for PG have been adapted from SUD treatment. For example, the most common treatment for PG is Gamblers Anonymous (GA), closely modeled on the abstinence model of Alcoholics Anonymous (AA) (Petry, 2007) (see Chap. 27). CBT has been shown to be effective in treating PG (Okuda, Balán, Petry, Oquendo, & Blanco, 2009; Petry et al., 2006), although the best results seem to result when CBT is combined with GA (Petry, 2005). Brief therapy, such as motivational enhancement therapy (MET), based on existing treatments for alcohol abuse, has also been shown to be effective (Hodgins, Currie, Currie, & Fick, 2009) (see Chap. 22).

There has been much research on drug treatments of GD, but, as of the writing of this book, none have an FDA indication. A 2007 meta-analysis concluded that antidepressants, mood stabilizers, and opiate antagonists are superior to placebo (Pallesen et al., 2007). Additionally, N-acetylcysteine (NAC), a glutamate modulator, has been shown to lessen the symptoms of PG in a small (n = 12) sample (Grant, Kim, & Odlaug, 2007). In a case report, disulfiram was shown to curb gambling urges in a person with PG (Mutschler et al., 2010). Outcomes with the atypical antipsychotic olanzapine have been equivocal (Fong, Kalechstein, Bernhard, Rosenthal, & Rugle, 2008). Pharmacotherapy is a viable option for GD co-occurring with a mood disorder, although no agent has emerged as a frontline treatment (Hodgins, Stea, & Grant, 2011).

8 Treatment of Comorbid Eating Disorders, Substance Use Disorders, and Behavioral Addictions

Little empirical data exists to inform treatment planning in individuals with ED, SUD, and BA. In fact, scant published literature exists on the treatment of individuals presenting with only the “dual diagnoses” of ED and SUD, with little agreement on or guidelines for “best practices” with this population (Brandt, Crawford, & Halmi, 2011). Reasons for this lack of research relate to the fact that until recently most experts have argued that treatments for these disorders should be sequential rather than concurrent (Brandt et al., 2011; Courbasson, Nishikawa, & Dixon, 2012). In addition, controlled treatment studies of ED and SUD typically exclude patients showing comorbidity with the other disorder (Courbasson et al., 2012).

Remarkably, no pharmacologic treatment studies of individuals with both ED and SUD have been published, and only one psychosocial treatment study of patients with both disorders has been published (Courbasson et al., 2012). In this small outpatient study, 25 women with ED and SUD were randomized to dialectical behavioral therapy (DBT) or treatment as usual (TAU) (Courbasson et al., 2012). The treatment trial lasted 1 year with subsequent follow-ups at 3 and 6 months. The DBT group had a much higher rate of retention posttreatment (80 % to 20 %) and at follow-up (60 % versus 20 %). DBT resulted in positive changes in multiple domains, including improvements in abnormal behaviors and attitudes related to the ED, substance use severity, regulation of negative mood, and depressive symptoms. Subjects’ perceived ability to manage and cope with negative emotions was significantly associated with reductions in emotional eating and increases in confidence to resist cravings for substances. For ethical reasons, the TAU arm had to be terminated prematurely because of worsening symptoms, preventing a head-to-head comparison of the two treatment groups. The authors noted that the huge dropout rate in the TAU arm confirms that retention of this patient group in treatment is very challenging.

Currently, expert clinical experience and the extant limited treatment literature have converged on recommending that comorbid ED and SUD should be treated concurrently (see Chap. 21). Based on what we currently know about these disorders and BA, a recommended treatment program for individuals with all three disorders should be multimodal using a chronic disease model in a patient population with a high degree of morbidity, mortality, relapse, and recurrence. Given the multiple comorbidities and needs of these patients, frequent contacts and coordination of care with multiple providers with expertise in each of the specific comorbid disorders are often necessary to ensure optimal clinical outcomes.

The initial component of successful intervention requires thorough assessment that includes systematically asking about specific BA including gambling disorder, kleptomania, problematic Internet use, and excessive/inappropriate sexual behavior. These conditions often go unrecognized because clinicians working with patients with ED and/or SUD do not think to ask about them.

Since all these conditions are associated with unsafe and potentially dangerous behaviors that cause enormous functional impairment and place individuals at much higher risk for self-harm and harm to others, assessment of safety and level of functional impairment is particularly important in this population. For GD patients, financial and legal problems and their impact on family relationships need to be investigated. Has the person declared bankruptcy or been incarcerated? For the person with hypersexual behavior, assessment for sexually transmitted diseases and unwanted pregnancies is crucial. For the individual with problematic Internet use, the impact of these activities on vocational functioning and relationships must be assessed to determine if excessive use has been associated with academic failure or poor work performance. In addition, has the person used the Internet to fuel another addiction such as going to pro-eating disorder web sites, gambling on-line, or accessing pornography?

The elevated rates of trauma in the backgrounds of patients with ED, SUD, and BA mandate an inquiry about a history of childhood physical, sexual, and emotional abuse (See Chap. 17). The assessment should also include questions about exposure to traumatic experiences in adulthood such as rape and domestic violence. Finally, there should be evaluation for the psychiatric sequelae of trauma, such as PTSD and borderline personality.

A flexible biopsychosocial approach to case formulation and treatment is optimal to address the myriad of problems these patients present. Treatment selection should be guided by the recognition that individuals with these comorbid disorders have problems with global dysregulation including affect regulation difficulties, impulsivity, addiction, and compulsive and habitual self-destructive behaviors, which some experts argue should be the major focus of intervention rather than a specific syndrome (Dennis & Sansone, 1997). Therefore, combinations of treatments shown to work for these deficits should be considered. For example, DBT and CBT are potentially helpful for these disorders. In addition, preliminary studies suggest that SSRIs and naltrexone may show promise as potential pharmacological treatments for BA, providing indirect support that these illnesses may overlap in part with mood disorders, anxiety disorders, obsessive–compulsive illnesses, and alcohol and drug use disorders.

In addition to interventions effective in treating all BA, consideration should be given to treatments specific for particular disorders such as Gamblers Anonymous for those with GD. Treating the major comorbid mood, anxiety, and personality disorders seen in ED and SUD must also be factored into treatment plans for addressing these complex and challenging conditions.

Conclusions

In summary, illness severity is greater and treatment response worse for ED and SUD when comorbid with BA at least in part because there are currently no evidence-based treatments for individuals with both ED and SUD with and without the accompanying BA discussed in this chapter. Moreover, with the exception of GD, there continues to be a lack of consensus regarding the validity of the other BA reviewed despite their obvious clinical importance. Accordingly, it is difficult to devise well-designed treatment studies for conditions whose existence and boundaries continue to be debated.

Nevertheless, future research directions for understanding the etiology of these comorbid conditions and developing novel interventions for them have been suggested by our growing understanding of the natural history of ED, SUD, and BA. For example, there is increasing evidence that impulsivity may be a risk factor for the development of many of the syndromes reviewed in this chapter. It has been shown, for instance, that impulsivity in adolescence actually predicts, often within less than a year, the development of signs and symptoms of ED (Wonderlich, Connolly, & Stice, 2004). In addition, in one of the largest cross-sectional studies of adult ED and ICD cited earlier, in more than 60 % of the subjects ICD preceded BA, and in almost half the sample, the onset of both illnesses was within a 3-year period (Fernández-Aranda, Pinheiro, et al., 2008).

As Fernández-Aranda, Pinheiro, et al. (2008) point out, these findings raise questions about the etiology of this comorbidity. Is the development of ED symptoms after the onset of ICD one of many manifestations of a core problem with impulsivity as seen in multi-impulsive bulimia (Lacey, 1993), or do those with comorbid ICD have more severe variants of ED related to genetic or environmental factors (Fernández-Aranda, Pinheiro, et al., 2008)? Better understanding of the nature of this comorbid relationship is essential in order to devise effective treatments for patients with ED, SUD, and ICD.

Certainly, one possible environmental factor that may be a leading candidate for the development of these disorders of impulse control is trauma, such as sexual abuse, which, with the advances of neuroimaging, has been shown to cause structural brain changes (Heim, Mayberg, Mletzko, Nemeroff, & Pruessner, 2013). In this context, psychosocial adversity may also increase the risk for the development of comorbid ED and ICD. Multiple studies have shown an association between a history of childhood abuse and multi-impulsive bulimia (Fichter et al., 1994; Wonderlich et al., 2001). In the earlier cited Myers et al. study (2006), women with MIB were significantly more likely than those with BN alone to have had a history of physical, sexual, and emotional abuse. They also were significantly more likely to suffer from post-traumatic stress disorder (PTSD) (34 % versus 14 %).

And, as cited earlier, the characteristic of “trait urgency,” which may represent deficits in specific brain systems, is associated with BA but sometimes with gender differences, suggesting that different phenotypes of BA may be related to the same underlying pathophysiology (Fischer & Smith, 2008). In other words, as the authors point out, gambling disorder in men and binge eating in women may actually be related to the same core deficits.

As we await further research, and a better understanding of the nature of non-DSM BA like Internet addiction and hypersexual behavior, additional controlled studies of treatments for comorbid ED and SUD should be conducted, as discovery of evidence-based treatments for these two disorders with prominent impulsivity may also be useful as interventions in BA. More controlled trials of both psychosocial and drug treatments of BA should also be performed given the preliminary promising results of studies cited in this chapter. The key to reducing the devastating effects of these comorbid disorders is for a greater collaboration among experts and researchers in ED, SUD, and BA in developing more effective treatments that alleviate the tremendous suffering of affected individuals.