ICBs occurred in 7.1% (10/140) of RLS patients treated with either a DA (8/10) or levodopa monotherapy (2/10). Eight of ten (80%) subjects had more than one comorbid ICB. ICBs in RLS patients were associated with higher levodopa-equivalent DA dose [23
], young age of RLS onset, history of experimental drug use, female gender and a family history of gambling disorders. These factors also accounted for 52% of the variance and predicted ICBs at 42.9%.
Study strengths include the use of validated or previously utilized screening questionnaires and established instruments to assess associated factors, as well as the use of stringent diagnostic criteria and extensive interviews with a psychologist or a psychiatrist. Based on the identification of ICBs in only 1/32 randomly selected non-responders (3.1%) we caution that the occurrence of ICBs in RLS patients found in this study may be over-estimated, as the experience of behavioural changes may have motivated the subjects to fill out the questionnaires and thus bias the results of the study.
One limitation of the study is the lack of a control group. However, all RLS patients with ICBs were on dopaminergic therapy and none of the RLS patients with non-dopaminergic medication was identified with an ICB. In accordance with previous findings [24
] this is supportive of an association between ICBs and dopaminergic medication in RLS patients. We can also confirm the previous association between the occurrence of ICBs in RLS and treatment with higher dosages of dopaminergic agents compared to RLS patients who did not develop ICBs [24
]. As depression may present an individual susceptibility factor for abnormal behaviours in PD patients [28
] and neuropsychological testing revealed preserved executive functions in a previous study on PD patients with pathological gambling [29
], the lack of further psychiatric and neuropsychological evaluation in our cohort further limits the study presented here. Comparative assessment of psychiatric comorbidities and cognitive profiles should be addressed in further studies.
Compulsive eating was identified in 6/140 (4.3%), compulsive shopping in 5/140 (3.6%), pathological gambling in 3/140 (2.1%), punding in 3/140 (2.1%), hypersexuality in 2/140 (1.4%) and compulsive medication use in none of the subjects. These frequencies are relatively similar to RLS studies conducted in North America [24
] although the phenotypes of ICB are different compared to US patient cohorts. One study compared 100 RLS treated patients, 275 controls with obstructive sleep apnoea and 52 RLS untreated patients. RLS patients on dopaminergic medications had greater pathologic gambling (5%), compulsive shopping (9%) and punding (7%) than the sleep apnoea control group along with greater compulsive shopping relative to the RLS untreated group [24
]. Our study uniquely focuses on ICBs in RLS patients in a European rather than a North American cohort and further investigates associated factors. The frequency of ICBs is greater in Parkinson's disease compared to RLS and has been suggested to be related to dose effects [24
]. In a large multicentre North American study focusing on Parkinson's disease, pathological gambling is reported in 2.9% (with problem gambling in an additional 2.3%), compulsive shopping in 6.0%, hypersexuality in 3.5%, and binge eating in 3.5% [10
]. Punding has been reported in 1.5 to 14% [12
] and compulsive medication use in 3 - 4% [16
In the North American general population, binge eating disorder is identified in 3.5% of women and 2% of men [31
] and compulsive shopping in 6% [32
]. The lifetime prevalence of a binge eating disorder in European countries, including Germany, has been reported to be 1.12% [33
]. The lifetime prevalence of pathological gambling in North America is 1.5% [34
] whereas the 2006 German Epidemiological Survey of Substance Abuse (n = 7817, aged 18-64 years old), that was controlled for mania, found a much lower lifetime prevalence of DSM IV-diagnosis of pathological gambling at 0.09% [35
]. Pathological gambling in Germany may be lower due to a negative cultural attitude towards gambling as well as a less widespread access to casinos. Intriguingly, the rate of pathological gambling in RLS is similar to the rate reported in PD in North America, but access to gambling in our study in Germany was through the Internet. The prevalence rates of binge eating and compulsive shopping in RLS patients may reflect a different pattern of presentation in females. We did not find RLS patients with compulsive medication use, which may be due to the lower frequency and dosage of levodopa use in Germany.
Associated factors such as a history of experimental drug use and a family history of gambling disorders are consistent to that observed in the literature on ICBs in Parkinson's disease [10
]. Factors such as cigarette and alcohol use and novelty seeking may reach significance with a larger sample size. This suggests that a similar biological predisposition may underlie these behaviours rather than Parkinson's disease itself being an absolutely necessary factor. The young age of RLS onset suggests that genetically mediated forms of RLS may have a different susceptibility to dopaminergic medications. In Parkinson's disease, male gender is considered a risk factor for specific ICBs such as hypersexuality [36
]. In our study only women were affected by ICBs. Further studies in large cohorts are necessary to clarify whether this is due to the greater overall prevalence of RLS in women - as reflected in our study cohort.
None of the RLS patients identified with ICBs in this study attributed their ICB symptoms to dopaminergic treatment. Although all affected study subjects were advised to have their treatment regimen changed, none felt sufficiently impaired by their behaviours to stop dopaminergic medications completely, but agreed to a reduced dosage or combination therapy with non-dopaminergic agents.