Accurate assessments of the exact frequency of ICD in RLS patients taking dopaminergic agents are complicated by the lack of standard criteria and validated assessment instruments used in earlier studies. Looking at specific ICDs, including gambling and excessive sexual activity, one group noted that 7% of patients reported a change in gambling habits and 5% an increase in sexual desire [6]. More recently, studies have drawn upon the experience from ICD in PD, utilizing common criteria to define specific conditions [5]. These not only provide a clearer picture of how often ICDs occur in RLS patients on treatment but also allow for comparisons with ICDs in PD patients. Using these criteria, the frequency for any ICD in RLS is approximately 7–17% depending on the study [7, 8]. It is important to keep in mind however that ICD in RLS has only recently become appreciated whereas ICD has been reported within the PD population since at least the late 1990s [9] on dopaminergic medications. In the largest cross-sectional study involving 3090 patients with PD, ICD frequency was 13.6% [10]. Based on these data, the frequency of ICD in both PD and RLS patients treated with dopaminergic agents are similar. Due to its relatively recent recognition within the RLS population, larger and more longitudinal studies are needed.

Interestingly, there may be differences in the types of patients who develop ICD when comparing RLS and PD. In the latter, most patients are younger at onset of ICD symptoms, have earlier PD onset, and are taking relatively higher doses of dopaminergic medications [6]. In contrast, RLS patients who develop ICD tend to be older and are generally on lower doses of dopaminergic medications [6, 7]. As an illustrative example, in two separate studies the mean dopamine dose for RLS patients with ICDs was 63.7 mg/d levodopa dose equivalents, compared to between 780–1100 for PD patients with similar behaviors [7, 11]. While the contrasting prevalence of impulsive behaviors in these individual studies (7.1% for RLS vs. 13.7% for PD) precludes a direct comparison, it highlights the significant difference in dopaminergic doses seen in patients with RLS or PD who develop ICD. Further, there appear to be no gender differences for ICDs overall in PD [10], whereas female gender is associated with a higher risk of ICD in RLS [7]. It should also be noted that although RLS is roughly twice as common in women, while PD is more likely to affect men, female gender is still associated with a higher rate of ICD in RLS after controlling for this difference [7]. Other factors which may predispose RLS patients to impulse control disorders include a history of experimental drug use and a family history of gambling disorder [7] (Table 16.2). Along with the aforementioned demographic differences between the RLS and PD populations with ICD, these factors suggest that while the underlying mechanism for these behaviors likely involves dopaminergic pathways, there may be additional differences in the underlying pathophysiology of the two disease states which result in the observed epidemiology. In the following section, we discuss the current models describing these phenomena, including recent functional imaging evidence supporting the role of aberrant dopaminergic signaling.

Epidemiological evidence and clinical experience implicate medication-induced aberrant dopaminergic signaling at the center of impulse control disorders in both PD and RLS. In the largest cross-sectional study of ICD in PD to date, 17.1% of patients treated with a dopamine agonist had an ICD of some type in contrast to 6.9% in those not taking a dopamine agonist, with an odds ratio of 2.72 [10]. While there are fewer epidemiological data for RLS, use of dopaminergic medications is similarly associated with ICD. This idea is further supported by case reports which have described resolution of these behaviors upon discontinuation of the dopamine agonist, findings which concur with our clinical experience [12–14].

Another theory is that ICD in both RLS and PD relates to the primary disease process, not to drug treatment [15]. However, the increased risk of developing these behaviors in patients taking dopamine agonists, and their resolution upon discontinuation of these medications, argues against this hypothesis. Furthermore, RLS and PD likely represent distinct neuropathophysiologic processes. Thus, similar impulsive and/or compulsive behaviors seen during treatment are more likely related to medications than the underlying disease.

The idea that dopamine agonist treatment influences these behaviors is consistent with our current understanding of dopaminergic reward pathways. While ICDs encompass a range of different behaviors, they are all reward driven and become repetitive and reinforced over time. Thus, the dopaminergic pathways which govern reward systems are the likely neuro-anatomic substrates for these medication effects. The precise pathways and anatomy have been extensively characterized and are beyond the scope of this chapter. However, the critical ones involve the ventral tegmental area, which contains dopaminergic mesolimbic neurons that project to the nucleus accumbens and prefrontal cortex. The D3 subtype of the dopamine receptor is particularly enriched in the mesolimbic system and thus may be intimately involved in addiction and control of impulsive behaviors [16, 17].

Dopamine agonists may interfere with these pathways in a number of ways resulting in aberrant behavior (Fig. 16.1). One prominent hypothesis relates to the disruption of normal phasic, or periodic, release of dopamine through the mesolimbic projections to the nucleus accumbens [5]. Under physiologic conditions, these events occur when an individual anticipates a reward and when an unexpected reward is received, a phenomenon known as reward prediction error manifests. When a reward is unexpectedly not given, phasic suppression of dopamine release occurs instead. Thus, mesolimbic dopaminergic pathways regulate reward and serve as a feedback teaching signal, modulating habit formation and other learned behaviors.

Interestingly, behaviors which do not consistently result in reward may be reinforced by the same system. This applies to many of the behaviors discussed in this chapter, most notably pathologic gambling. We may understand its physiology in terms of tonic, or sustained, release of dopamine. This occurs when individuals anticipate a reward in the setting of uncertainty, in other words, when the probability of actually obtaining or not obtaining a reward is similar. Thus tonic dopamine release may explain the repetitive nature of these behaviors, where simply anticipating a reward in the context of uncertainty is inherently rewarding.

Disruption of these physiologic pathways, which may occur with exogenous dopamine treatment, can explain the impulse control behaviors seen in RLS (Fig. 16.1c, d). For example, the teaching of feedback signaling may be absent, increasing the susceptibility for repetitive behaviors. In this setting, dopamine agonists or exogenous dopamine may further exacerbate the situation by enhancing the habit-forming properties of reward while having no effect on the feedback teaching that normally occurs with “losses” [18]. By influencing behavior through only positive reward reinforcement, impulsive and repetitive behaviors may be increased.