|Home | About | Journals | Submit | Contact Us | Français|
Impulse control disorders are a psychiatric condition characterized by the failure to resist an impulsive act or behavior that may be harmful to self or others. In movement disorders, impulse control disorders are associated with dopaminergic treatment, notably dopamine agonists (DAs). Impulse control disorders have been studied extensively in Parkinson’s disease, but are also recognized in restless leg syndrome and atypical Parkinsonian syndromes. Epidemiological studies suggest younger age, male sex, greater novelty seeking, impulsivity, depression and premorbid impulse control disorders as the most consistent risk factors. Such patients may warrant special monitoring after starting treatment with a DA. Various individual screening tools are available for people without Parkinson’s disease. The Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease has been developed specifically for Parkinson’s disease. The best treatment for impulse control disorders is prevention. However, after the development of impulse control disorders, the mainstay intervention is to reduce or discontinue the offending anti-Parkinsonian medication. In refractory cases, other pharmacological interventions are available, including neuroleptics, antiepileptics, amantadine, antiandrogens, lithium and opioid antagonists. Unfortunately, their use is only supported by case reports, small case series or open-label clinical studies. Prospective, controlled studies are warranted. Ongoing investigations include naltrexone and nicotine.
The hallmark of impulse control disorders (ICDs) is a ‘failure to resist an impulse, drive, or temptation to perform an act that is harmful to the person or to others’ according to the American Psychiatric Association’s Diagnostic and Statistical Manual – fourth edition text revised (DSM-IVTR) [American Psychiatric Association, 2000]. The most common ICDs are pathological gambling (PG), hypersexuality (HS), compulsive shopping (CS) and compulsive eating (CE). Two large epidemiological studies conducted in North America inform us that PG is a relatively rare disorder with a prevalence of 0.42–1% in the general population [Ferris et al. 1996; Petry et al. 2005] and a nationwide study on CS reported a lifetime prevalence of 5.8% in the adult US population [Koran et al. 2006].
The DSM-IVTR [American Psychiatric Association, 2000] defines PG as a ‘persistent and recurrent maladaptive gambling behavior’ (see Table 1). To date, a formal schema for diagnosis of hypersexuality does not exist. A definition of HS based on DSM format has been proposed in a prevalence study in Parkinson’s disease (PD) [Voon et al. 2006] (Table 1) A definition to be included in the upcoming DSM-V was recently published [Kafka, 2010]. CS is characterized by a frequent preoccupation with shopping or intrusive, irresistible, ‘senseless’ buying impulses that exceed what is needed or can be afforded, and causes distress or significant interference with work or social functioning [McElroy et al. 1994] (Table 1). CE is characterized as uncontrolled binge eating without emesis or laxative abuse in a working definition of the DSM-IVTR. It is often, but not always, associated with obesity symptoms [American Psychiatric Association, 2000].
In movement disorders, ICDs have been studied mainly in PD. The next section will cover the brief literature available on ICDs in other movement disorders.
Initial studies of ICDs in PD [Pontone et al. 2006; Weintraub et al. 2006; Grosset et al. 2011; Isaias et al. 2008; Singh et al. 2007; Voon et al. 2006] reported a prevalence of 2.2–28.0% for PG [Weintraub et al. 2006; Isaias et al. 2008], 2.0–11.0% for HS [Voon et al. 2006; Singh et al. 2007], 0.4–6.4% for CS [Voon et al. 2006; Weintraub et al. 2009] and 1–4.5% for CE [Zahodne et al. 2011; Weintraub et al. 2009]. It is important to recognize that these estimates were obtained with heterogeneous study designs, sample sizes, diagnostic criteria and ascertainment methods. For example, a patient-rated questionnaire on HS documented a lifetime prevalence of 2.4% [Voon et al. 2006], while a clinician-rated survey conducted in the same PD population reported a lifetime prevalence of 8.4% (unpublished data from authors). The prevalence of active HS was similar, with 2.0 % for the patient-rated questionnaire [Voon et al. 2006] and 1.9% for the clinician-rated survey (unpublished data from authors).
The largest epidemiological study on ICDs in PD, the DOMINO study, reported a 6-month prevalence of any ICDs of 13.6%, including PG (5.0%), HS (3.5%), CS (5.7%) and CE (4.3%) [Weintraub et al. 2010a].
Other repetitive behaviors have been recognized as part of the spectrum of ICDs in patients with PD, including walkabouts, that is, excessive, aimless wandering [Giovannoni et al. 2000; Quinn et al. 1983] reported by 3.2% of patients [Weintraub et al. 2009], hoarding [O’Sullivan et al. 2010b], kleptomania [Bienfait et al. 2010], impulsive smoking [Bienfait et al. 2010], reckless generosity [O’Sullivan et al. 2010c] and reckless driving [Avanzi et al. 2008].
Dopamine dysregulation syndrome (DDS) and punding are related behaviors. DDS or hedonic homeostatic dysregulation is defined by a compulsive use of dopaminergic medication with drug hoarding regardless of the patient being ‘on’ and experiencing excessive and significant dyskinesias. It can be associated with a hypomanic state and symptoms consistent with a withdrawal syndrome [Giovannoni et al. 2000]. Estimated prevalence for DDS ranges from 0.6 % to 3.4–4% [Weintraub et al. 2009; Giovannoni et al. 2000; Pezzella et al. 2005]. Levodopa and short-acting DAs are the main culprits and DDS is associated, in at least 88% of the cases, with severe off-period dysphoria, disabling dyskinesias and co-occurrence of punding [Evans et al. 2005]. It is worth emphasizing that the reverse does not happen, with primary punders manifesting less often a DDS [Evans et al. 2005]. Punding corresponds to an intense fascination with meaningless movements or activities that are repeated in an endless fashion. It was first used to coin stereotyped motor behaviors in amphetamine and cocaine addicts [Rylander, 1972]. Based on surveys conducted in tertiary outpatient movement disorders centers, the incidence of punding ranges between 1.4% and 13.8% in PD [Evans et al. 2004; Miyasaki et al. 2007; Weintraub et al. 2009]. Punding behaviors encompass repetitive manipulations of technical equipment, handling and sorting common objects, grooming, cleaning, gardening, writing, artistic drawing or craft-making, singing or playing a musical instrument, playing cards, fishing and excessive computer use [Lim et al. 2008]. The patient typically will have difficulty in completing projects, is aware of creating chaos while punding and will resist to disengage from the activity, continuing punding overnight or manifesting irritability if interrupted [Evans et al. 2004]. Hoarding and walkabouts have been included in a spectrum of punding, though other authors study these phenomena separately [Weintraub et al. 2009]. The first description of punding in PD was associated with L-dopa [Friedman, 1994]. Later, a small case series documented the association of punding with the use of DAs [Fernandez and Friedman, 1999]. Hobbyism is considered a complex form of punding related to a specific activity or hobby (e.g. coin collection, repairing or dismantling things, computer use) that frequently corresponds to a premorbid activity or hobby. Hobbyism is reported to occur in 14.6% of patients with PD [Weintraub et al. 2009] and rarely (3.8 %) co-occurs with punding [Weintraub et al. 2009].
The clinician should be aware that ICDs are reported in other movement disorders using dopaminergic medication for treatment. These include restless legs syndrome (RLS) [Evans and Stegeman, 2009; Quickfall and Suchowersky, 2007], progressive supranuclear palsy (PSP) regardless of the clinical variant (Richardson’s syndrome versus PSP Parkinsonism) [Kim et al. 2008; O’Sullivan et al. 2010a] and multiple system atrophy [Klos et al. 2005; McKeon et al. 2007].
The literature is limited to case reports or case series with the exception of RLS. A case–control study reported that the frequency of any ICDs in a RLS-treated group was 17%, based on a questionnaire followed by a phone interview [Cornelius et al. 2010]. The latency between initiation of treatment and onset of any ICD was 9.5 months [Cornelius et al. 2010]. The frequency of individual ICDs was 11% for CE, 9% for CS, 7% for punding, 5% for PG and 3% for HS [Cornelius et al. 2010]. In another study of ICDs with 140 patients, screening though patient-completed questionnaires and confirmation by psychiatric interview revealed 80% of the patients had multiple ICDs [Voon et al. 2011a]. The prevalence of any ICDs was 11% and included CE (4.3%), CS (3.6%), PG (2.1%), punding (2.1%) and HS (1.4%) [Voon et al. 2011a]. DDS was not reported [Voon et al. 2011a], though it has been recognized in a single patient with RLS [Leu-Semenescu et al. 2009]. In addition, ICDs were associated with a higher DA dose, younger age of onset of RLS, history of experimental drug use, female sex and a family history of gambling disorders [Voon et al. 2011a]. Recently, a case series of 27 patients with RLS undergoing treatment with rotigotine reported a prevalence of 21% for single (n = 4) and multiple (n = 2) ICDs. Regardless of occurrence of ICDs, all patients showed an increase of impulsivity measured by the Zurich Screening Questionnaire for ICD [Schreglmann et al. 2012]. Other studies, controlled for the cofounder of medication, reported that patients with RLS without dopaminergic treatment have an impairment of executive function with worse performance in the Wisconsin Card Sorting Test for patients with frequent symptoms [Fulda et al. 2011], reduced decision-making abilities under ambiguity but not risky conditions [Bayard et al. 2010] and deficits in short-term attention and verbal fluency [Fulda et al. 2010] compared with matched controls. These features can also be associated with secondary manifestations of sleep deprivation or depression in RLS [Bayard et al. 2010], but it leads to a hypothesis in need of further testing that the executive dysfunction found in patients with RLS may predispose them to ICDs.
ICDs were also reported in a patient with secondary RLS due to multiple sclerosis [Evans and Butzkueven, 2007].
The knowledge of risk factors for ICDs in PD enables the clinician to identify patients more prone to develop ICDs. DA treatment is the primary risk factor for ICDs in PD [Weintraub et al. 2010b] and ICDs are usually considered a drug class effect. Nevertheless, it is important to recognize that ICDs can occur on a stable dose of dopaminergic treatment, away from the onset of dopaminergic treatment or a change in dose [Cooper et al. 2009; Kurlan, 2004].
Though exceedingly rare (~2%), cases exist [Gallagher et al. 2007] documenting an association of L-dopa, amantadine and selegiline with ICDs [Fernandez and Friedman, 1999; Shapiro et al. 2006; Weintraub et al. 2010b]. In addition, in a recent case–control study, the use of DAs [adjusted odds ratio 20.3; 95% confidence interval (CI) 2.70–65.01] but also of monoamine oxidase B inhibitors (adjusted odds ratio 3.74; 95 % CI 1.11–12.64) stood out among other dopaminergic drug classes as predictors of occurrence of ICDs [Perez-Lloret et al. 2012].
Though nearly all patients with PD are being treated with dopaminergic drugs, at least 17% develop ICDs. The incidence of subsyndromal forms of ICDs is not known. Robust epidemiological data provide insight regarding individual variables associated with ICDs. Studies analyzing individual characteristics associated with the occurrence of ICDs in PD [Singh et al. 2007; Isaias et al. 2008; Pontone et al. 2006; Voon et al. 2007, 2011b; Weintraub et al. 2010b] commonly report younger age, male sex, greater premorbid novelty seeking, impulsivity and depression as risk factors for ICDs. Efforts to identity the contribution of each factor for specific ICDs disclosed that medication-induced mania and adjunctive DA use (versus monotherapy) are risk factors for PG [Voon et al. 2007] and male sex and early onset of PD onset are more associated with HS [Weintraub et al. 2010b; Uitti et al. 1989]. A history of ICD symptoms prior to PD was found to be a premorbid characteristic of patients with PD who went on to develop PG, HS or PS [Weintraub et al. 2006]. Personal or family history of substance abuse including alcoholism increased risk of ICDs in one cohort [Voon et al. 2006]. A history of repetitive behaviors or obesity was associated with CE and weight gain was associated with CS [Nirenberg and Waters, 2006; Rylander, 1972]. Motor complications are more common in patients with ICDs [Solla et al. 2011]. Dyskinesias have been found to be significantly present in patients with HS or CS, but not PG or CE [Solla et al. 2011]. Multiple ICDs will occur more likely in younger patients with more intense dyskinesias and alcohol use [Voon et al. 2011b].
Related behaviors such as DDS share similar risk factors, including the existence of depressive symptoms, alcohol intake, younger age at PD onset, impulsive sensation seeking and novelty-seeking personality traits [Evans et al. 2005], but are better predicted by a higher dopaminergic drug intake and greater past experimental drug use [Evans et al. 2005]. An increased clinical severity of punding has been associated with higher total daily L-dopa equivalents, younger age, reduced sleep time overnight and average number of daily rescue doses [Evans et al. 2004]. Insomnia, previously treated psychosis and excessive libido were commonly reported comorbidities. Dyskinesias have also been associated with punding behaviors and DDS but not ICDs [Silveira-Moriyama et al. 2006].
Clinical scales have been created to screen and diagnose different forms of ICDs in the general population. The South Oaks Gambling Screen [Lesieur and Blume, 1987], Pathological Gambling Adaptation of the Yale–Brown Obsessive–Compulsive Scale [Pallanti et al. 2005], the Gambling Symptom Assessment Scale [Kim et al. 2009] and the Massachusetts Gambling Screen [Shaffer et al. 1994] are used for PG. Some of these have been used in PD for specific ICDs. A Buying Questionnaire [Christenson et al. 1994] is also found in literature. The Minnesota Impulsive Disorders Interview is a global measurement tool that includes sections for PG, HS and CS, however it is not formatted for diagnosis according to DSM-IVTR and lacks clear instructions for its administration and scoring guidance [Weintraub et al. 2012]. These are not validated in PD, but some of the items were incorporated in the Questionnaire for Impulsive-Compulsive Disorders in Parkinson’s Disease (QUIP) (see below) [Weintraub et al. 2009]. All of the above scales are designed to be patient completed. However, in addiction research, patient-completed tools are not sufficient as screening or diagnostic tools. As discussed below, informed ‘others’ increase their sensitivity.
Initial initiatives to produce clinical rating scales specific for ICDs in PD include a hypersexuality questionnaire [Voon et al. 2006] and a punding questionnaire [Evans et al. 2005]. The first attempt to combine various survey instruments in a clinically usable tool in PD was the QUIP [Weintraub et al. 2009]. It is a self-completed screening questionnaire for ICDs and other compulsive behaviors occurring at ‘anytime’ during PD [Weintraub et al. 2009]. A short form of the QUIP (QUIP-S) has been recommend for routine use since it has similar psychometric properties to the original version and a shorter median completion time of 3 min [Weintraub et al. 2009]. An intrinsic limitation is the requirement of a 12th grade reading level [Weintraub et al. 2009]. For the QUIP-S, the ability to identify an individual with any at least one ICD is 0.89 and the discriminant validity for selected cut-offs in the different ICDs was 0.95 (0.84–1.05) for PG, 0.97 (0.93–0.99) for HS, 0.87 (0.72–1.02) for PB and 0.88 (0.72–1.04) for CE. The QUIP is also valid when completed by a patient’s informant [Papay et al. 2011].
A follow-up study of the QUIP showed a sensitivity of 100% regardless of whether a patient or informant completed the instrument and a lower specificity (~75%) that translated to 40% of patients with a positive QUIP based on either patient or informant reporting not fulfilling criteria for a formal diagnosis of any ICD [Papay et al. 2011]. This suggests that a positive screen with QUIP should be followed by a comprehensive psychiatric interview to determine the range and severity of symptoms. In addition, the application of the QUIP documented the existence of subsyndromal ICD symptoms, which likely represent an important window of opportunity for clinicians to act early and prevent more serious consequences of ICDs [Weintraub et al. 2009]. The natural history of subsyndromal forms is not known. For example, do subsyndromal forms necessarily progress to ICDs and over what time course? Could subsyndromal forms be stable for long periods of time? What separates those patients experiencing subsyndromal forms compared with full ICDs? Patients often acknowledge subsyndromal forms only after a reduction in their dopaminergic medication and complete resolution of their symptoms [Voon et al. 2006]. A new version of the QUIP (QUIP rating scale/QUIP-RS) has been designed to capture a change in the severity of ICD symptoms and thus improve the characterization of subsyndromal forms. The QUIP-RS was found to be a valid and reliable rating scale for ICDs and related disorders in PD, applicable both for diagnosis and monitoring temporal changes in the severity of symptoms [Weintraub et al. 2012]. The QUIP and QUIP-RS have not been validated for other populations at risk for ICD development [Weintraub et al. 2009]. A clinical scale with a comprehensive section in impulse control behaviors, including the most common ICDs and related behaviors in PD (PG, HS, PS, punding and DDS), has been created in French [Ardouin et al. 2009] but requires further clinimetric validation.
Clues for future approaches to identify patients with PD more prone to developing ICDs are given by imaging data such as positron emission tomography that consistently document a distinct functional state of the brain in patients with PD who developed ICDs (see Figure 1) [Cilia et al. 2010; Steeves et al. 2009; van Eimeren et al. 2010]. The predictive value of these changes in a premorbid condition is yet to be established.
The best treatment for ICDs is prevention. Counseling patients and ideally spouses or family members prior to initiating therapy and recognizing premonitory signs of ICDs is mandatory. This includes counseling families to question and report to the physician unexplained absences, secretive behavior, irritability, hiding evidence of the ICDs and monetary consequences of ICDs. The majority of patients with PD with ICDs may be undetected by their physicians [Weintraub et al. 2006; Evans et al. 2004]. The normalization of the behaviors associated with dopaminergic treatment induced ICDs improves patients’ comfort in disclosing these activities (some patients develop the following behaviors as a result of their treatment). The disparity between patient self report and spouse or family member report continues in the literature. This highlights either lack of insight regarding the impact of behavior or guilt and shame resulting in denying behavior. In addition, the mean latency of onset of PG from onset of DA treatment is 23.0 months (range 1–84 months) [Gallagher et al. 2007], which warrants a sustained monitoring of ICDs by the physician. Changes in behavior such as an increasing tendency for insomnia, missing meals, doses of medication and absences from work can be potential red flags for the presence of ICDs (Table 2).
Potential risk factors such as male sex, young age and a history of drug abuse should be taken into consideration. Physicians may wish to counsel these individuals carefully or avoid the use of DAs. In addition, premorbid gambling, drug use histories and impulsive sensation seeking personality traits may be relevant in identifying at-risk individuals [Gallagher et al. 2007].
The following nonpharmacological approaches may be helpful. PG consequences may be mitigated by limiting access to bank accounts, limiting access to travel to casinos, discontinuing credit card accounts and providing a budget for patients. Families and physicians should be aware that increased access to the internet and thus online gambling are particular challenges that facilitate ICDs in subjects with motor disability due to PD [Gallagher et al. 2007]. Blocking computer sites for gambling, shopping or pornography is frequently warranted and patients should avoid visits to casinos or places to purchase lottery tickets. Clinicians should discourage patients from taking nocturnal or ‘rescue’ doses of short-acting dopaminergic agents, although this is not systematically shown to be beneficial in reducing risk. Such practices may reduce the overall levodopa equivalent daily dose and thus reduce the risk of ICDs. In those more prone to develop DDS/punding or with subsyndromal forms, medication refills should be limited to avoid self-regulation of dosage and hoarding of medications [Ferrara and Stacy, 2008]. Engaged family members are crucial to the success of all these measures. Individuals without strong social support are most vulnerable to unchecked ICDs potentially from multiple factors: despair at isolation, detached relationship patterns and depression, in addition to lack of support to withstand medication reduction and behavior modification.
Cognitive behavioral therapy is helpful in non-PD behavioral addictions and may have a role in PD [Petry and Roll, 2001]. Other authors have not found ‘12-step’ programs helpful in ICDs without medication modification. Such programs may be more successful for those with a family history of alcohol, substance abuse or pathologic gambling as these individuals may have an increased susceptibility to behavioral addictions, even in the absence of dopaminergic stimulation. Identification of the ICD alone may result in resolution or ability to control it [Petry and Roll, 2001].
The mainstay of treatment is the reduction or discontinuation of DAs [Klos et al. 2005; Gallagher et al. 2007; Nirenberg and Waters, 2006]. Reduction in L-dopa dose may also be required [Gallagher et al. 2007; Quinn et al. 1983]. One case of spontaneous remission has been reported [Klos et al. 2005]. Little is known about the long-term prognosis of treated ICDs. Two studies with a total of 30 patients and a follow-up of 21–29 months showed that an increase in L-dopa after DA withdrawal sufficient to maintain motor symptom control was not associated with recurrence of ICDs, even with higher total daily L-dopa equivalents [Macphee et al. 2009; Mamikonyan et al. 2008]. Nevertheless, one patient developed DDS [Macphee et al. 2009; Mamikonyan et al. 2008]. Some case reports suggest that switching from a short-acting DA (ropinirole, pramipexole) to a long-acting DA (piribedil) or treatment with a DA with less D3 receptor agonist action can be effective [Gallagher et al. 2007]. Nevertheless, this strategy is rarely used. When changing dopaminergic medication, the following conversion rates can be used: 100 mg L-dopa is equivalent to 1 mg pergolide, pramipexole, lisuride or cabergoline, 6 mg of ropinirole, 10 mg of bromocriptine or apomorphine, 75 mg sustained-release L-dopa, 133 mg L-dopa with entacapone, and 150 mg L-dopa with tolcapone [Tomlinson et al. 2010].
DA dose reduction can be complicated by a highly stereotyped symptom complex, including irritability, anxiety, agoraphobia, dysphoria, depression, drug cravings, fatigue, pain, diaphoresis and orthostatic hypotension. The Dopamine Agonist Withdrawal Syndrome (DAWS) has been characterized in two studies in a total of 390 patients with PD and concurrent use of a DA. The reported prevalence ranges from 15.5% to 19% [Pondal et al. 2011; Rabinak and Nirenberg, 2010]. Risk factors for DAWS are the presence of baseline ICDs [Pondal et al. 2011; Rabinak and Nirenberg, 2010], a higher baseline DA use and cumulative exposure of a DA, and lower scores on the UPDRS motor score, regardless of disease duration and dopaminergic medication use [Rabinak and Nirenberg, 2010]. Higher L-dopa dose at withdrawal and baseline smoking have also been identified to be more common in DAWS [Pondal et al. 2011]. The prognosis is good with full recovery in weeks to months in 61% of patients, though 15% were unable to discontinue the DA [Pondal et al. 2011]. Consequently, a point of practice is that DA withdrawal should be slow (over several months to achieve complete discontinuation) to avoid the occurrence of DAWS and close vigilance by the physician. Here again, spouse or family members need to understand the need to discontinue the DA and to encourage the patient to persist to achieve resolution of ICDs.
It is important to recognize that reduction/discontinuation of a DA may not be successful in all patients with PD and ICDs because of persisting ICDs, poor tolerance due to DAWS or significant motor worsening. In a case series of patients with PD with PG and punding, there was a lack of improvement after reduction of dopaminergic medications [Kurlan, 2004] and symptoms persisted in about 80% of patients in a case series of punders, regardless of changes in dopaminergic medication [Avila et al. 2011]. For those cases, the pharmacological options are few and there is a lack of robust evidence to inform the clinician. In a recent Movement Disorder Society Evidence-Based Medicine Review, no randomized, controlled trials were found for ICDs in PD [Seppi et al. 2011]. Data are extremely limited and most studies are case reports or small case series. Prospective, double-blinded, controlled trials are needed [Galpern and Stacy, 2007]. For ICDs in PD, the need to consider placebo responses when analyzing uncontrolled trials of interventions is further emphasized. The placebo effect is highly significant both in non-PD populations with pathological gambling or compulsive buying and in patients with PD [Lim et al. 2008].
Although the use of neuroleptics reportedly improves ICDs, the use of typical neuroleptics, risperidone or olanzapine will most likely worsen motor function in PD [Gschwandtner et al. 2001; Klos et al. 2005; Quinn et al. 1983]. Kurlan did not find improvement of ICDs with risperidone or olanzapine [Kurlan, 2004]. Quetiapine [Evans et al. 2004; Fasano et al. 2006; Jimenez-Jimenez et al. 2002; Sevincok et al. 2007], and clozapine [Fernandez and Durso, 1998; Kurlan, 2004] were used to treat punding, PG and HS without deleterious effects on motor function. Quetiapine is typically used first because clozapine requires long-term hematological monitoring. This advantage has to be weighed against the observation that motor worsening with quetiapine can happen in 32% of Parkinsonian patients and lead to discontinuation in 9% of those patients [Fernandez et al. 2003]. Anecdotally, quetiapine in doses of 100–200 mg triggered punding in two patients with PD and psychosis, which resolved with dose reduction [Miwa et al. 2004].
A crossover study of 17 patients with PD reported a benefit of amantadine 200 mg/day for the treatment of PG in patients with PD [Thomas et al. 2010] and open-label amantadine 100–300 mg/day was beneficial for punding [Fasano et al. 2011; Kashihara and Imamura, 2008]. The role of amantadine on ICDs remains controversial, as in the post hoc analysis of the largest epidemiological study conducted for ICDs, amantadine was associated with the presence of one or more ICDs, including PG, HS, CE and CS, independent of DA use and L-dopa dose [Weintraub et al. 2010b]. Nevertheless, the use of amantadine for the treatment of PG or punding lacks robust evidence, as concluded in a recent Movement Disorder Society evidence-based medicine review [Seppi et al. 2011].
The treatment of associated psychopathologies such as depression and mania, known risk factors for ICDs, can be useful for ICDs. The current recommendations for the treatment of depression or depressive symptoms in PD are selective serotonin reuptake inhibitors (SSRIs) and the tricyclic antidepressants nortriptyline and desipramine [Seppi et al. 2011].
In addition, antidepressants, namely SSRIs, can help reduce ICDs independently of the presence of depressive symptoms [Klos et al. 2005; Mamikonyan et al. 2008; McElroy et al. 2011]. Clomipramine [Evans et al. 2004; Jimenez-Jimenez et al. 2002], a tricyclic antidepressant, has been used for nocturnal punding. In a series of six patients with PD, treatment with a diverse range of antidepressants such as citalopram, fluvoxamine, paroxetine, sertraline, imipramine was not successful for PG or punding [Kurlan, 2004].
Valproate [Klos et al. 2005; Hicks et al. 2011] and topiramate [Bermejo, 2008] have also been reported to improve ICDs. Valproate can worsen Parkinsonism and cause cognitive deterioration. Both side effects are reversible upon drug discontinuation [Armon et al. 1996]. Zonisamide (25–200 mg/day) has been studied in 15 patients with PG, HS, CE or CS in a noncontrolled and open-label design. The study found a marked reduction in the severity of impulsive behaviors without deterioration of the UPDRS motor score and a good tolerance [Bermejo et al. 2010].
Naltrexone, an opioid antagonist, has been found to be efficacious in three patients with PG refractory to the discontinuation of DA [Bosco et al. 2012]. Donepezil has been used with success in a patient with HS and cognitive deterioration [Ivanco and Bohnen, 2005]. Selegiline discontinuation was necessary to treat punding in one patient [Fernandez and Friedman, 1999] and, in an open-label study, punding and PG were treated with success after the introduction of amantadine [Fasano et al. 2011]. The antiandrogen cyproterone has been used in two refractory cases of HS [Evans et al. 2004] and finasteride, a 5 α-reductase inhibitor used for benign prostatic hypertrophy, was reported to reduce PG in two patients with PD [Bortolato et al. 2012].
Ongoing prospective controlled studies are being conducted for naltrexone [ClinicalTrials.gov identifier: NCT01052831] and nicotine for [ClinicalTrials.gov identifier: NCT01216904] ICDs in PD.
Interventional studies conducted for ICDs in patients without PD open new perspectives in the treatment of refractory ICDs in PD. Memantine [ClinicalTrials.gov identifier: NCT00585169], tolcapone [ClinicalTrials.gov identifier: NCT009 27563], ecocipam, a selective dopamine D1 antagonist [ClinicalTrials.gov identifier: NCT01 215357] and the opioid antagonists naltrexone [ClinicalTrials.gov identifier: NCT01057862] and nalmefene [Grant et al. 2010] are being studied for PG in patients without PD. In addition, nonpharmacological interventions such as cognitive-motivational behavior therapy [ClinicalTrials.gov identifier: NCT01135264], a combination of behavioral therapy and N-acetylcysteine specifically for smokers with PG [ClinicalTrials.gov identifier: NCT00967005], repeated low-frequency transcranial magnetic stimulation, ClinicalTrials.gov identifier: NCT01560351 and a community reinforcement approach and family training focused on significant others of people with PG are being investigated [ClinicalTrials.gov identifier: NCT 01340274]. Acamprosate, a synthetic analogue of γ-aminobutyric acid approved for the treatment of alcohol abuse has been tested in binge-eating disorder in patients without PD in a randomized, placebo-controlled study showing the partial benefit in the frequency of episodes and weight gain [McElroy et al. 2011].
Deep brain stimulation (DBS) has been proposed for ICD treatment. Current understanding is that DBS of the subthalamic nucleus (STN) improves ICDs by allowing a reduction in the levodopa equivalent daily dose or discontinuation of DAs [Ardouin et al. 2006]. In surgical case series, ICDs, DDS and punding worsened and persisted after DBS [Lim et al. 2009]. One case report documented the de novo appearance of PG following bilateral STN DBS without adjustment of medication, which recurred when switching on the neurostimulator [Smeding et al. 2007]. A recent chart review in a surgical center reported de novo occurrence of DDS after either STN DBS or globus pallidus pars interna (GPi) DBS [Moum et al. 2012]. One possible explanation is that improved motor function allowed full expression of the ICD following DBS. Nevertheless, some aspects of DBS in PD can be potentially used to the advantage of a patient with a prior or ongoing history of ICDs proposed for DBS. STN may be a more suitable target compared with GPi as it enables greater dopaminergic medication reduction [Rodriguez-Oroz et al. 2005; Volkmann et al. 2004]. In some cases, major depression can occur following reduction of dopaminergic medication after DBS, which may worsen pre-existing ICDs or cause the development of new ICDs. At this time, DBS should not be considered treatment for ICDs in PD per se. Conversely, ICDs are not a contraindication for DBS, although careful monitoring of patients is required. Psychiatric assessment prior to surgery and following implantation and programming may help avoid worsening or triggering of ICDs.
ICDs and related behaviors such as punding and DDS reduce quality of life in a significant minority of patients with PD. Although prevention and reduction of dopaminergic therapy is the mainstay of management, these disorders may be unsatisfactorily treated after such interventions. A variety of other pharmacological treatments used in addiction research has been studied for ICDs in PD. Further well designed clinical trials are required to allow evidence-based treatment of ICDs. The present review emphasizes the need for further research with prospective, controlled studies that will help the physician better manage ICDs in PD and other movement disorders.
Funding: The present review was not supported by any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement: The authors declare no conflict of interest in preparing this article.
Tiago A. Mestre, Movement Disorders Centre and the Edmond J. Safra Program in Parkinson’s Disease, Toronto Western Hospital, University Health Network, Division of Neurology, University of Toronto, Toronto, Ontario, Canada.
Antonio P. Strafella, Movement Disorders Centre and the Edmond J. Safra Program in Parkinson’s Disease, Toronto Western Hospital, University Health Network, Division of Neurology, University of Toronto, Toronto, Ontario, Canada.
Teri Thomsen, Department of Neurology, Carver College of Medicine, University of Iowa, Iowa City, Iowa, USA.
Valerie Voon, Behavioral and Clinical Neurosciences Institute, University of Cambridge, Cambridge, England.
Janis Miyasaki, Movement Disorder Clinic, Toronto Western Hospital, 399 Bathurst Street, MC-7 402, Toronto, Ontario, Canada M5T 2S8.