Impulse control disorder (ICD) behaviors, including hypersexuality, excessive gambling and shopping, have been recently reported among patients with Parkinson’s disease (PD) treated with dopamine agonists, including pramipexole and ropinirole (Dodd et al. 2005). Although the mechanism of action leading to this toxicity remains unclear, the temporal association of compulsivity with induction of dopaminergic therapy and resolution of these behaviors upon withdrawal of therapy suggests a strong causal relationship. A similar association of ICD with use of pramipexole for treatment of fibromyalgia (FM) may also be present.

With the discovery of the role of deficient hippocampal dopaminergic neurotransmission (Wood 2004, 2006) in patients with FM by fDOPA PET scanning (Wood et al. 2007), and the validation of reduced pain, fatigue and improved function in patients with FM treated with dopamine agonists (Holman 2003a, b, 2004; Holman et al. 2004; Holman and Myers 2005), these medications, especially pramipexole, have become important treatment options for many patients with FM. Use of these agents has been endorsed by recent FM treatment guidelines published by the EUropean League Against Rheumatism (Carville et al. 2008). Unfortunately, six years after the first use of these agents in FM, similar behavioral concerns have been recently observed in some patients. ICD caused by dopamine agonist was initially reported by lay press articles, then in internet reports and finally, by solicitation for litigation. Unfortunately, current understanding of this problem remains limited and underreported.

While this important concern has been reviewed in the setting of PD treatment, this is the first report of ICD related to use of dopamine agonists in patients with FM. Because this behavioral concern is not easily detected, potentially leads to devastating social and financial loss for patients and their families, and is rarely spontaneously reported to monitoring clinicians, dissemination of this potential toxicity is important. And, given the remarkably different pathophysiology of FM compared to PD, and the ability to rapidly withdraw therapy in patients with FM, further insight into this unexpected adverse response may be possible.

Materials and Methods

A comprehensive chart review of all patients treated at a suburban rheumatology practice between 2002 and 2006 was conducted to identify patients with FM. Although an exemption for consent was authorized by Western Institutional Review Board (Olympia, WA, USA) for a chart review, consent was requested from each subject who had developed ICD by the investigator. From patient telephone contact, monitoring visits and occasionally lost patients contacting this office after reading lay press reports, patients were identified as having ICD leading to excessive shopping and gambling with serious social or financial consequences consistent with ICD according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria. Hypersexuality and other compulsive behaviors associated with ICD were not assessed.

All patients were evaluated in the clinic within one week of identifying compulsivity and tapered off the dopamine agonist over 7 days in the course of routine clinical care. No clinical decisions were made for research purposes. Patients, and often their accompanying families, were counseled and asked to fill out questionnaires related to their potential risk of ICD, demographics, FM treatment response and to provide comments. Patients were asked to return for reevaluation within 10–14 days to monitor their response to medication withdrawal and their FM status. No patients were advised to continue pramipexole, even at a lower dosage, following the 7-day withdrawal schedule.

Statistical analysis of data was completed using SPSS 14.0 (SPSS, Chicago, IL.) to describe patient characteristics, including use of Pearson correlation to screen this exploratory data for potential risk factors.

Results

Initial chart screening identified 3006 patients with fibromyalgia, of which, 1356 had taken at least one dose of a dopamine agonist ( >95% pramipexole). Of 21 subjects identified with compulsive gambling or shopping, all had been treated with pramipexole at a mean dose of 4.5 mg po qhs for 40.0 ± 21.4 (range 8–96) months consistent with the most efficacious dose of pramipexole use to treat FM in a randomized, controlled trial (Holman and Myers 2005). All had been advised to taper pramipexole by 0.75 mg qhs until discontinued without authorization to reinitiate this therapy. All 21 patients were monitored by telephone follow-up and 19 patients agreed to further evaluation consisting of interview and examination. Fifteen provided completed detailed questionnaires and provided permission to publish their comments.

Demographics are described in Table 1. Most patients were women of middle-age, well educated, married, employed and described a positive clinical response to treatment of FM with pramipexole. Over 73% of patients described their clinical response to this treatment option as ‘good’ or better, including 40% who designated ‘excellent’. Most patients did not smoke cigarettes or consume alcohol. Compulsive behaviors were split between gambling (33%), shopping (40%) or both (27%). Some patients also noted compulsive eating (13%), cleaning (7%) and crafting (13%). Mean onset of compulsive behavior was 14.4 ± 14.9 (range 1–52) months after initiating pramipexole. Resolution of these behaviors occurred within 3–10 days after discontinuation for 19 of 21 patients and for all within 3 months.

Table 1 Characteristics of patients with compulsive behavior
Full size table

Potential risk factors and similar compulsive behavior prior to initiation of pramipexole were noted by some patients. Prior psychiatric care and a family history (FH) of similar compulsive behaviors were common (Table 2). Statistically significant correlations were unavailable due to lack of a control cohort.

Table 2 Demographic features of patients with impulse control disorder (ICD) behavior and fibromyalgia treated with dopamine agonists
Full size table

Case Examples

Given the unique and unexpected variety of clinical vignettes and the difficulty of identifying compulsivity among patients with FM, case descriptions are provided to assist clinicians.

Patient 6

This 52 year-old, divorced woman noted an ‘excellent’ response to pramipexole, but noted compulsive gambling within 6 months of treatment and loss of >$ 60,000. She had forgotten to renew her prescription and within 3 days of sudden discontinuation, “my compulsion stopped. It was like a light switch that had gone off in my head. While gambling (and working at a casino), I was so incredibly focused on the games that no one and nothing mattered. I was evading questions and even lied about how much money I was spending and where I was getting it. I never lie––totally out of character.” A few years prior, she had moved to a rural area to avoid department stores due to compulsive shopping not attributed to medication. Compulsivity and FM symptoms have not returned.

Patient 13

This 59 year-old woman described a history of compulsive shopping and a FH of compulsivity and bipolar disorder. “About 20 years ago, I used to have all my cupboards full of food and I always (bought) more groceries. I spend my husband’s check before he got it.” She had noted a ‘good’ response to pramipexole, but noted compulsive gambling, without compulsive shopping, after treatment for 12 months. While lost to follow-up, pramipexole was discontinued and restarted three times by a different physician. Compulsive behavior resolved and returned with each discontinuation and reinitiation of pramipexole. However, she concluded “I just think I was always compulsive and it wasn’t the medicine.”

Patient 15

This 53 year-old employed, divorced woman with post-graduate education and an ‘excellent’ response to pramipexole noted a FH of compulsivity (parents and 2 siblings), but not prior excessive gambling or shopping. After one year of treatment with pramipexole, compulsive gambling and shopping began and ultimately involved $150,000. “I just couldn’t stop. It was a very, very unusual feeling. I bought a new house and just kept buying things without any feeling of monetary consequences. Some people close to me laughed it off—others (my children) became very concerned I would gamble away everything I had.” After 18 months, the association of pramipexole and compulsivity was noted, and behavior resolved within one week of discontinuation.

Discussion

ICD behaviors, including excessive gambling, shopping, hypersexuality and binge eating, have been observed in patients with PD treated with dopamine agonists. Initial reports were sporadic after FDA approval of these agents in 1997, but recently, controlled assessment of pathological gambling has been reported in 8% of Scottish PD patients receiving pramipexole (Grosset et al. 2006). Similar, but unquantified reports have described pathological gambling in patients using lower doses these agents for restless legs syndrome (Quickfall and Suchowersky 2007; Evans and Butzkueven 2007). As patients with FM also begin treatment with dopamine agonists, development of ICD may be an important new concern for them as well.

In this retrospective chart review, an accurate prevalence of compulsivity has not been determined. As a likely underestimate, 15 cases in 1358 patients (1.1%) were identified with serious compulsive gambling and shopping. Other ICD behaviors noted in patients taking dopamine agonists for PD, such as binge eating and hypersexuality, were not assessed. A direct link to use of pramipexole was supported by resolution of these behaviors within 1–2 weeks following a mandatory, monitored and gradual discontinuation of pramipexole over one week. This causal relationship has also been reported in patients with PD treated with DAs, although their rate of dose reduction and discontinuation has generally required many months (Mamikonyan et al. 2008).

Understanding of the pathogenesis of compulsive gambling is developing slowly. Dopamine genes, specifically the DRD1-800 T/C allele, may be associated with pathological gambling (da Silva Lobo et al. 2007) Manipulation of dopamine is thought to play an important role in risk/reward behaviors (Ondo and Lai 2008) as well as some psychiatric disorders already associated with excessive compulsivity and pathological gambling: notably bipolar disorder (McIntyre et al. 2007). Ironically, pramipexole has also been reported as an effective treatment for both bipolar and unipolar depression (Aiken 2007). As a risk factor for patients taking dopamine agonists for PD, bipolar disorder has not been reported. But, a FH of both bipolar disorder and compulsivity were common among our patients with FM who excessively gambled or shopped. Controlled study of pathological gambling in PD has identified risk factors, including younger age and larger doses of dopamine agonists in PD, but not in RLS (Gallagher et al. 2007; Singh et al. 2007). Males with PD were over-represented, but are underrepresented in FM cohorts that usually include a female to male ratio of 4:1.

As our understanding of a dopaminergic pathogenesis of FM emerges and clinical trials demonstrate the benefits of treating FM with dopamine agonists, monitoring and preventing unexpected adverse events becomes increasingly important. Patients must be counseled about a broader risk profile for these agents. It is also incumbent upon patients to collaborate and provide their prescribing clinician with feedback about these behaviors.

Monitoring the onset of compulsivity is fraught with logistical hurdles, and most clinicians do not routinely monitor for ICD. Fortunately, at least for a majority of patients with FM, these behaviors appear rapidly reversible and timely appraisal for ICD is possible with tools, such as the Minnesota Impulse Disorder Interview (Grant 2008). These preliminary findings suggest a need for more formal, controlled studies as well as validation of tools used to assess ICD among patients with FM.