Should compulsive sexual behavior be considered an addiction? (2016)

Printer-friendly version

COMMENTS: This paper was published under the "Debate" category in the journal 'Addiction'. Its main weakness is that it purports to address compulsive sexual behavior (CSB), an umbrella term that covers everything sexual. For example, "CSB" can encompass hypersexuality or "sex addiction" and may include such behaviors as serial infidelity or acting out with prostitutes. Yet many compulsive porn users do not act out sexually, and limit their compulsive behavior to internet porn use. "Sex addiction," and the research on it, need to be considered separately from Internet porn addiction. The latter is a subtype of Internet addiction. See -

What's most frustrating about this paper is that the "Statement of the problem" and "Defining CSB" sections are about "hypersexuality," while the studies supporting the neurobiological basis of CSB are almost all on internet porn users. This kind of ambiguity creates more confusion than clarity, because it necessitates unnecessarily cautious language with respect to the research on internet porn users, thus slowing the recognition of the strong (and growing) evidence that Internet addictions are unquestionably genuine and that Internet porn addiction is a subtype.

Shane W. Kraus1,2,*, Valerie Voon3 and Marc N. Potenza2,4

Article first published online: 18 FEB 2016

Journal: Addiction

DOI: 10.1111/add.13297


Aims: To review the evidence base for classifying compulsive sexual behavior (CSB) as a non-substance or ‘behavioral’ addiction.

Methods: Data from multiple domains (e.g. epidemiological, phenomenological, clinical, biological) are reviewed and considered with respect to data from substance and gambling addictions.

Results: Overlapping features exist between CSB and substance use disorders. Common neurotransmitter systems may contribute to CSB and substance use disorders, and recent neuroimaging studies highlight similarities relating to craving and attentional biases. Similar pharmacological and psychotherapeutic treatments may be applicable to CSB and substance addictions, although considerable gaps in knowledge currently exist.

Conclusions: Despite the growing body of research linking compulsive sexual behavior (CSB) to substance addictions, significant gaps in understanding continue to complicate classification of CSB as an addiction.

KEY WORDS: Addiction, behavioral addictions, compulsive sexual behavior, hypersexuality, neurobiology, psychiatric disorder, sexual behavior, sexual compulsivity


The release of the Diagnostic and Statistical Manual (DSM-5) [1] altered addiction classifications. For the first time, the DSM-5 grouped a disorder not involving substance use (gambling disorder) together with substance use disorders in a new category entitled: ‘Substance-Related and Addictive Disorders’. Although researchers had advocated previously for its classification as an addiction [2–4], the re-classification has sparked debate, and it is not clear whether a similar classification will occur in the 11th edition of the International Classification of Diseases (ICD-11) [5]. In addition to considering gambling disorder as a non-substance-related addiction, DSM-5 committee members considered whether other conditions such as Internet gaming disorder should be characterized as ‘behavioral’ addictions [6]. Although Internet gaming disorder was not included in DSM-5, it was added to Section 3 for further study. Other disorders were considered, but not included in DSM-5. Specifically, proposed criteria for hypersexual disorder [7] were excluded, generating questions about the diagnostic future of problematic/excessive sexual behaviors. Multiple reasons probably contributed to these decisions, with insufficient data in important domains likely contributing [8].

In the current paper, compulsive sexual behavior (CSB), defined as difficulties in controlling inappropriate or excessive sexual fantasies, urges/cravings or behaviors that generate subjective distress or impairment in one’s daily functioning, will be considered, as will its possible relationships to gambling and substance addictions. In CSB, intense and repetitive sexual fantasies, urges/cravings or behaviors may increase over time and have been linked to health, psychosocial and interpersonal impairments [7,9]. Although prior studies have drawn similarities between sexual addiction, problematic hypersexuality/hypersexual disorder and sexual compulsivity, we will use the term CSB to reflect a broader category of problematic/excessive sexual behaviors that subsumes all the above terms.

The current paper considers classification of CSB by reviewing data from multiple domains (e.g. epidemiological, phenomenological, clinical, biological) and addressing some of the diagnostic and classification issues that remain unanswered. Centrally, should CSB (including excessive casual sex, viewing of pornography and/or masturbation) be considered a diagnosable disorder and, if so, should it be classified as a behavioral addiction? Given the current research gaps on the study of CSB, we conclude with recommendations for future research and ways in which research can inform better diagnostic assessment and treatments efforts for people seeing professional help for CSB.


During the last several decades, publications referencing the study of CSB have increased (Fig. 1). Despite the growing body of research, little consensus exists among researchers and clinicians about the definition and presentation of CSB [10]. Some view problematic/excessive engagement in sexual behaviors as a feature of hypersexual disorder [7], a non-paraphilic CSB [11], a mood disorder such as bipolar disorder [12] or as a ‘behavioral’ addiction [13,14]. CSB is also being considered as a diagnostic entity within the category of impulsecontrol disorders in ICD-11 work [5].

Within the last decade, researchers and clinicians have begun conceptualizing CSB within the framework of problematic hypersexuality. In 2010, Martin Kafka proposed a new psychiatric disorder called ‘hypersexual disorder’ for DSM-5 consideration [7]. Despite a field trial supporting the reliability and validity of criteria for hypersexual disorder [15], the American Psychiatric Association excluded hypersexual disorder from DSM-5. Concerns were raised about the lack of research including anatomical and functional imaging, molecular genetics, pathophysiology, epidemiology and neuropsychological testing [8]. Others expressed concerns that hypersexual disorder could lead to forensic abuse or produce false positive diagnoses, given the absence of clear distinctions between normal range and pathological levels of sexual desires and behaviors [16–18].

Multiple criteria for hypersexual disorder share similarities with those for substance use disorders (Table 1) [14]. Both include criteria relating to impaired control (i.e. unsuccessful attempts to moderate or quit) and risky use (i.e. use/behavior leads to hazardous situations). Criteria differ for social impairment between hypersexual and substance use disorders. Substance use disorder criteria also include two items assessing physiological dependence (i.e. tolerance and withdrawal), and criteria for hypersexual disorder do not. Unique to hypersexual disorder (with respect to substance use disorders) are two criteria relating to dysphoric mood states. These criteria suggest hypersexual disorder’s origins might reflect maladaptive coping strategies, rather than a means of warding off withdrawal symptoms (e.g. anxiety associated with withdrawal from substances). Whether a person experiences withdrawal or tolerance related to a specific sexual behavior is debated, although it has been suggested that dysphoric mood states may reflect withdrawal symptoms for individuals with CSB who have recently cut back or quit engagement in problematic sexual behaviors [19]. A final difference between hypersexual disorder and substance use disorders involves diagnostic thresholding. Specifically, substance use disorders require a minimum of two criteria, whereas hypersexual disorder requires four of five of the ‘A’ criteria to be met. Currently, additional research is needed to determine the most appropriate diagnostic threshold for CSB [20].

Clinical characteristics of CSB

Insufficient data exist regarding CSB’s prevalence. Large scale community data regarding prevalence estimates CSB are lacking, making the true prevalence of CSB unknown. Researchers estimate rates ranging from 3 to 6% [7] with adult males comprising the majority (80% or higher) of affected individuals [15]. A large study of US university students found estimates of CSB to be 3% for men and 1% for women [21]. Among US male military combat veterans, prevalence was estimated to be closer to 17% [22]. Using data from the US National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), the life-time prevalence rates of sexual impulsivity, a possible dimension of CSB, was found to be higher for men (18.9%) than women (10.9%) [23]. Although important, we emphasize that similar gaps in knowledge did not prevent the introduction of pathological gambling into DSM-III in 1980 or the inclusion of Internet gaming disorder into Section 3 of DSM-5 (see wide prevalence estimates ranging from approximately 1 to 50%, depending on how problematic Internet use is defined and thresholded [6]).

CSB appears more frequent among men compared to women [7]. Samples of university-aged [21, 24] and community members [15, 25, 26] suggest that men, compared to women, are more likely to seek professional treatment for CSB [27]. Among CSB men, the most reported clinically distressing behaviors are compulsive masturbation, pornography use, casual/anonymous sex with strangers, multiple sexual partners and paid sex [15, 28, 29]. Among women, high masturbation frequency, number of sexual partners and pornography use are associated with CSB [30].

In a field trial for hypersexual disorder, 54% of patients reported experiencing dysregulated sexual fantasies, urges and behaviors prior to adulthood, suggesting an early onset. Eighty-two per cent of patients reported experiencing a gradual progression of hypersexual disorder symptoms over months or years [15]. Progression of sexual urges over time is associated with personal distress and functional impairment across important life domains (e.g. occupational, familial, social and financial) [31]. Hypersexual individuals may have propensities to experience more negative than positive emotions, and self-critical affect (e.g. shame, self-hostility) may contribute to the maintenance of CSB [32]. Given limited studies and mixed results, it is unclear whether CSB is associated with deficits in impaired decision-making/ executive functioning [33–36].

In DSM-5, ‘craving’ was added as a diagnostic criterion for substance use disorders [1]. Similarly, craving appears relevant to the assessment and treatment of CSB. Among young adult men, craving for pornography correlated positively with psychological/psychiatric symptoms, sexual compulsivity and severity of cybersex addiction [37–41]. A potential role for craving in predicting relapse or clinical outcomes.

In treatment-seeking patients, university students, and community members, CSB appears more common among European/white individuals compared to others (e.g. African American, Latino, Asian Americans] [15, 21]. Limited data suggest that individuals seeking treatment for CSB may be of higher socio-economic status compared to those with other psychiatric disorders [15, 42], although this finding might reflect greater access to treatment (including private-pay treatment given limitations in insurance coverage) for individuals with higher incomes. CSB has also been found among men who have sex with men [28, 43, 44], and is associated with HIV risk-taking behaviors (e.g. condomless anal intercourse) [44, 45]. CSB is associated with elevated rates of sexual risk taking in both heterosexual and non-heterosexual individuals, reflected in high rates of HIV and other sexually transmitted.

Psychopathology and CSB

CSB occurs frequently with other psychiatric disorders. About half of hypersexual individuals meet criteria for at least one DSM-IV mood, anxiety, substance use, impulse control or personality disorder [22,28,29,46]. In 103 men seeking treatment for compulsive pornography use and/or casual sexual behaviors, 71% met criteria for a mood disorder, 40% for an anxiety disorder, 41% for a substance use disorder and 24% for an impulse control disorder [47]. Estimated rates of co-occurring CSB and gambling disorder range from 4 to 20% [25, 26, 47, 48]. Sexual impulsivity is associated with multiple psychiatric disorders across sexes, particularly for women. Among women compared to men, sexual impulsivity was associated more strongly with social phobia, alcohol use disorder and paranoid, schizotypal, antisocial, borderline, narcissistic, avoidant and obsessive–compulsive personality disorders [23].


Understanding whether CSB shares neurobiological similarities with (or differences from) substance use and gambling disorders would help to inform ICD-11-related efforts and treatment interventions. Dopaminergic and serotonergic pathways may contribute to the development and maintenance of CSB, although this research is arguably in its infancy [49]. Positive findings for citalopram in a double-blind placebo controlled study of CSB among a sample of men suggests possible serotonergic dysfunction [50]. Naltrexone, an opioid antagonist, may be effective at reducing both the urges and behaviors associated with CSB, consistent with roles in substance and gambling addictions and consistent with proposed mechanisms of opioid-related modulation of dopaminergic activity in mesolimbic pathways [51–53].

The most compelling evidence between dopamine and CSB relates to Parkinson’s disease. Dopamine replacement therapies (e.g. levodopa and dopamine agonists such as pramipexole, ropinirole) have been associated with impulse control behaviors/disorders (including CSB) among individuals with Parkinson’s disease [54–57]. Among 3090 Parkinson’s disease patients, dopamine agonist use was associated with a 2.6-fold increase odds of having CSB [57]. CSB among Parkinson’s disease patients has also been reported to remit once the medication has been discontinued [54]. Levodopa has also been associated with CSB and other impulse control disorders in Parkinson’s disease, as have multiple other factors (e.g. geographic location, marital status) [57].

The pathophysiology of CSB, currently poorly understood, is being researched actively. Dysregulated hypothalamic–pituitary–adrenal axis function has been linked to addictions and was identified recently in CSB. CSB men were more likely than non-CSB men to be dexamethasone suppression-test non-suppressors and have higher adrenocorticotrophic hormone levels. The hyperactive hypothalamic–pituitary–adrenal axis in CSB men may underlie craving and CSB behaviors related to battling dysphoric emotional states [58].

Existing neuroimaging studies have focused primarily on cue-induced reactivity. Cue-reactivity is clinically relevant to drug addictions, contributing to cravings, urges and relapses [59]. A recent meta-analysis reported overlap between tobacco, cocaine and alcohol cue-reactivity in the ventral striatum, anterior cingulate cortex (AC) and amygdala related to drug cue-reactivity and self-reported craving, suggesting that these brain regions may constitute a core circuit of drug craving across addictions [60]. The incentive motivation theory of addictions posits that addiction is related to the enhanced incentive salience to drug-associated stimuli resulting in greater attentional capture, approach behaviors, expectancy and pathological motivation (or ‘wanting’) for drugs. [61, 62]. This theory has also been applied to CSB [63].

In college female students [64], individual differences in human reward-related brain activity in the nucleus accumbens in response to food and sexual images related prospectively to weight gain and sexual activity 6 months later. Heightened reward responsivity in the brain to food or sexual cues was associated with overeating and increased sexual activity, suggesting a common neural mechanism associated with appetitive behaviors. During functional magnetic resonance imaging (fMRI), exposure to pornographic video cues compared to non-sexual exciting videos in CSB men relative to non-CSB men was associated with greater activation in the dorsal anterior cingulate, ventral striatum and amygdala, regions implicated in drug-cue reactivity studies in drug addictions [63]. Functional connectivity of these regions was associated with subjective sexual desire to the cues, but not liking, among men with CSB. Here, desire was taken as an index of ‘wanting’ compared to ‘liking’. The men with CSB versus those without also reported heightened sexual desire and demonstrated greater anterior cingulate and striatal activation in response to pornographic images [65].

CSB men compared to those without also showed greater attentional biases to sexually explicit cues, suggesting a role for early attentional orienting responses towards pornographic cues [66]. CSB men also demonstrated greater choice preference for cues conditioned to both sexual and monetary stimuli compared to men without CSB [67]. The greater early attentional bias to sexual cues was associated with greater approach behaviors towards conditioned sexual cues, thus supporting incentive motivation theories of addiction. CSB subjects also showed a preference for novel sexual images and greater dorsal cingulate habituation to repeated exposure to sexual pictures, with the degree of habituation correlating with enhanced preference for sexual novelty [67]. The access to novel sexual stimuli may be specific to online availability of novel materials.

Among Parkinson’s disease subjects, exposure to sexual cues increased sexual desire in those with CSB compared to those without [68]; enhanced activity in limbic, paralimbic, temporal, occipital, somatosensory and prefrontal regions implicated in emotional, cognitive, autonomic, visual and motivational processes was also observed. CSB patients’ increased sexual desire correlated with increased activations in the ventral striatum and cingulate and orbitofrontal cortices [68].These findings resonate with those in drug addictions in which increased activation of these reward-related regions is seen in response to cues related to the specific addiction, in contrast to blunted responses to general or monetary rewards [69, 70]. Other studies have also implicated prefrontal regions; in a small diffusion tensor imaging study, CSB versus non-CSB men showed higher superior frontal mean diffusivity [71].

In contrast, other studies focusing on individuals without CSB have emphasized a role for habituation. In non-CSB men, a longer history of pornography viewing was correlated with lower left putaminal responses to pornographic photos, suggesting potential desensitization [72]. Similarly, in an event-related potential study with men andwomen without CSB, those reporting problematic use of pornography had a lower late positive potential to pornographic photos relative to those not reporting problematic use. The late positive potential is elevated commonly in response to drug cues in addiction studies [73]. These findings contrast to, but are not incompatible with, the report of enhanced activity in the fMRI studies in CSB subjects; the studies differ in stimuli type, modality of measure and the population under study. The CSB study used infrequently shown videos compared to repeated photos; the degree of activation has been shown to differ to videos versus photos and habituation may differ depending on the stimuli. Furthermore, in those reporting problematic use in the event-related potential study, the number of hours of use was relatively low [problem: 3.8, standard deviation (SD) = 1.3 versus control: 0.6, SD = 1.5 hours/week] compared to the CSB fMRI study (CSB: 13.21, SD = 9.85 versus control: 1.75, SD = 3.36 hours/week). Thus, habituation may relate to general use, with severe use potentially associated with enhanced cue-reactivity. Further larger studies are required to examine these differences.

Genetics of CSB

Genetic data relating to CSB are sparse. No genome-wide association study of CSB has been performed. A study of 88 married couples with CSB found high frequencies of first degree relatives with substance use disorders (40%), eating disorders (30%) or pathological gambling (7%) [74]. A twin study suggested genetic contributions accounted for 77% of the variance relating to problematic masturbatory behaviors, whereas 13% was attributable to non-shared environmental factors [75]. Substantial genetic contributions also exist for substance and gambling addictions [76, 77]. Using twin data [78], the estimated proportion of variation in liability for gambling disorder due to genetic influences is approximately 50%, with higher proportions seen for more severe problems. Inherited factors associated with impulsivity may represent a vulnerability marker for the development of substance use disorders [79]; however, whether these factors increase odds of developing CSB has not yet been explored.


During the last decade, research on the diagnosis and treatment of CSB has increased [80]. Various researchers have proposed diagnostic criteria [13] and developed assessment tools [81] to aid clinicians in the treatment of CSB; however, the reliability, validity and utility of many of these scales remain largely unexplored. Few measures have been validated, limiting their generalizability for clinical practice.

Treatment interventions for CSB require additional research. Few studies have evaluated the efficacies and tolerabilities of specific pharmacological [53, 82–86] and psychotherapeutic [87–91] treatments for CSB. Evidence based psychotherapies such as cognitive–behavioral  therapy and acceptance-and-commitment therapy appear helpful for CSB [89,91,92]. Similarly, serotonergic reuptake inhibitors (e.g. fluoxetine, sertraline and citalopram) and opioid antagonists (e.g. naltrexone) have demonstrated preliminary efficacy in reducing CSB symptoms and behaviors, although large-scale randomized controlled trials are lacking. Existing medication studies have typically been case studies. Only one study [50] used a double-bind, placebo-controlled design when evaluating the efficacy and tolerability of a drug (citalopram) in the treatment of CSB.

No large randomized controlled trials exist examining the efficacy of psychotherapies in treating CSB. Methodological issues limit the generalizability of existing clinical outcomes studies, as most studies employ weak methodological designs, differ on inclusion/exclusion criteria, fail to use random assignment for treatment conditions and do not include control groups necessary to conclude that the treatment worked [80]. Large, randomized controlled trials are needed to evaluate the efficacies and tolerabilities of medications and psychotherapies in treating CSB.

Alternative perspectives

The proposal of hypersexual disorder as a psychiatric disorder has not been embraced uniformly. Concerns have been raised that the label of ‘disorder’ pathologizes normal variants of healthy sexual behavior [93], or that excessive/problematic sexual behavior may be better explained as an extension of a pre-existing mental health disorder or poor coping strategies used to regulate negative affect states rather than a distinct psychiatric disorder [16,18]. Other researchers expressed concern that some individuals labeled with CSB may merely have high levels of sexual desire [18], with suggestions that difficulty controlling sexual urges and high frequencies of sexual behaviors and consequences associated with those behaviors may be explained more clearly as a non-pathological variation of high sexual desire [94].

In a large sample of Croatian adults, cluster analysis identified two meaningful clusters, one representing problematic sexuality
and another reflecting high sexual desire and frequent sexual activity. Individuals in the problematic cluster reported more psychopathology compared to individuals in the high-desire/frequent-activity cluster [95]. This suggests that CSB may be organized more along a continuum of increasing sexual frequency and preoccupation, in which clinical cases are more
likely to occur at the upper end of the continuum or dimension [96]. Given the likelihood that there is considerable overlap between CSB and high sexual desire, additional research is needed to identify features associated most specifically with clinically distressing sexual behaviors.


With the release of DSM-5, gambling disorder was reclassified with substance use disorders. This change challenged beliefs that addiction occurred only by ingesting of mind-altering substances and has significant implications for policy, prevention and treatment strategies [97]. Data suggest that excessive engagement in other behaviors (e.g. gaming, sex, compulsive shopping) may share clinical, genetic, neurobiological and phenomenological parallels with substance addictions [2,14]. Despite the increasing number of publications on CSB, multiple gaps in knowledge exist that would help to determine more conclusively whether excessive engagement in sexual behaviors might best be classified as an addiction. In Table 2, we list areas where additional research is needed to increase understanding of CSB. Such insufficient data complicate classification, prevention and treatment efforts. While neuroimaging data suggest similarities between substance addictions and CSB, data are limited by small sample sizes, solely male heterosexual samples and cross-sectional designs. Additional research is needed to understand CSB in women, underprivileged and racial/ethnic minority groups, gay, lesbian, bisexual and transgendered people, individuals with physical and intellectual disabilities and other groups.

Another area needing more research involves considering how technological changes may be influencing human sexual behaviors. Given that data suggest that sexual behaviors are facilitated through Internet and smartphone applications [98–100], additional research should consider how digital technologies relate to CSB (e.g. compulsive masturbation to Internet pornography or sex chatrooms) and engagement in risky sexual behaviors (e.g. condomless sex, multiple sexual partners on one occasion). For example, whether increased access to Internet pornography and the use of websites and smartphone applications (e.g. Grindr, FindFred, Scruff, Tinder, Pure, etc.) designed to facilitate casual sex between consenting adults is associated with an increased reports of hypersexual behaviors awaits future research. As such data are collected, acquired knowledge should be translated into improved policy, prevention and treatment strategies


This study was funded by support from the Department of Veterans Affairs, VISN 1 Mental Illness Research Education and Clinical Center, the National Center for Responsible Gaming, and CASAColumbia. The content of this manuscript does not necessarily reflect the views of the funding agencies and reflect the views of the authors. The authors report that they have no financial conflicts of interest with respect to the content of this manuscript.

Declaration of interests

The authors report that they have no financial conflicts of interest with respect to the content of this manuscript. M. N.P. has received financial support or compensation for the following: has consulted for and advised Lundbeck, Ironwood, Shire, INSYS and RiverMend Health; has received research support (to Yale) from the National Institutes of Health, Mohegan Sun Casino, the National Center for Responsible Gaming and Pfizer pharmaceuticals; has participated in surveys, mailings or telephone consultations related to drug addiction, impulse control disorders or other health topics; has consulted for gambling and legal entities on issues related to impulse control; provides clinical care in the Connecticut Department of Mental Health and Addiction Services Problem Gambling Services Program; has performed grant reviews for the National Institutes of Health and other agencies; has edited or guest-edited journals or journal sections; has given academic lectures in grand rounds, CME events and other clinical or scientific venues; and has generated books or book chapters for publishers of mental health texts.