Restless legs syndrome

Restless legs syndrome (RLS) is not rare but is rarely diagnosed by clinicians. Articles published in this and recent issues of Neurology reflect the growing interest in this area, fueled by new findings from pharmacologic, electrophysiologic, and neuroimaging studies. Despite a lucid description over 50 years ago by Ekbom,1 there is considerable misconception about RLS. Persons with RLS, even when their symptoms are quite troublesome or disabling, often do not seek medical attention, or the symptoms are wrongly attributed by physicians to nervousness, insomnia, stress, muscle cramps, arthritis, or a simple consequence of aging. Although no medical specialty has claimed rights of ownership to RLS, the correct diagnosis is usually made by neurologists, movement disorder experts, and sleep specialists.

The poor recognition and frequent misdiagnosis have hampered epidemiologic studies in RLS. Estimated prevalence rates vary widely, from 1% to 15%, but the true prevalence is probably close to 5% in the general population and considerably higher in the elderly. One study of 133 patients with typical RLS found the mean age at onset to be 27.2 years and the presence of RLS in at least one first-degree relative in 63% of cases.2 Future epidemiologic studies will be aided by the diagnostic criteria formulated by the International Restless Legs Syndrome Study Group (IRLSSG).3 The minimal criteria include the following: 1) an intense, irresistible urge to move the legs, usually associated with sensory complaints (paresthesia or dysesthesia); 2) motor restlessness; 3) worsening of symptoms at rest and relief with motor activation; and 4) increased severity in the evening or at night. Periodic limb movements in sleep (PLMS), detected by an overnight sleep study and present in at least 80% of patients with RLS, is the only laboratory abnormality typically associated with RLS. There is, however, night-to-night variability, which makes it problematic to use PLMS as a confirmatory finding. Therefore, other laboratory diagnostic criteria are needed. Neurophysiologic changes, such as blink reflex and H-reflex abnormalities, startle reflex and spinal flexor reflex hyperactivity, and magnetic brain stimulation finding of motor cortex disinhibition, have been noted in some patients with RLS, but their sensitivity and specificity have yet to be determined. As there are no physiologic markers for RLS, the severity is best measured by the unified rating scale recently developed, but not yet validated, by the IRLSSG.

Validated diagnostic criteria will allow rigorous genetic studies in RLS. Up to one-third to one-half of RLS cases are transmitted as an autosomal dominant trait. There are no twin studies and linkage has not been established in RLS. In this issue of Neurology, the observation by Gemignani et al.4 of RLS in 10 of 27 (37%) patients with Charcot-Marie-Tooth neuronal type (CMT2) and in 0 of 17 with CMT1 support the conclusion that a disorder of the sensory input may play a role in the pathogenesis of RLS. The findings also suggest that CMT2 associated with RLS represents a discreet genetic subgroup. Autosomal dominant cerebellar ataxia may be another genetic disorder associated with RLS. Schols et al.,5 for example, observed RLS in 45% of 89 patients with familial ataxia, of whom 59 had the SCA3 type. These studies suggest genetic as well as phenotypic heterogeneity in RLS.

Primary (sporadic or genetic) RLS may be difficult to differentiate from RLS associated with neuropathy, uremia, iron deficiency, or other disorders (secondary RLS). One study6 found that 15 of 41 patients with RLS had electrophysiologic evidence of polyneuropathy or radiculopathy, although only 7 of the 15 showed clinical signs of neuropathy. Positive family history of RLS was more common in the primary than in the secondary forms.6 Another study7 involving a consecutive series of patients with polyneuropathy, however, documented only a 5% frequency of RLS. Sural nerve biopsy findings in 7 of 8 patients with primary RLS were consistent with an axonal neuropathy,8 although there may be other explanations. It is not clear why some patients with peripheral neuropathy develop symptoms of RLS whereas others do not.

There have been scattered reports of deficiencies—such as vitamin B₁₂, folate, magnesium, and iron—causing secondary RLS. O’Keeffe et al.9 first reported low iron, measured by ferritin levels, in RLS patients. Iron is needed as a cofactor for tyrosine hydroxylase, the rate limiting enzyme in the synthesis of dopamine; therefore, iron deficiency may impair the normal production of dopamine. Furthermore, D2 receptor is an iron containing protein and, hence, iron deficiency may impair the normal function of D2 receptors. Whether iron deficiency is primarily responsible for causing RLS or whether it simply aggravates or triggers RLS symptoms in patients who are already predisposed to develop RLS, however, remains to be determined.10

Pharmacologic studies have provided indirect evidence of dopaminergic abnormality in RLS. Levodopa,11 for example, nearly always relieves, and dopamine antagonists often worsen, RLS symptoms. Exacerbation of RLS in the evening when dopamine activity is at its lowest level, and exacerbation with iron deficiency, which may interfere with the production of dopamine, are further evidence in support of impaired dopamine transmission in RLS. In this issue of Neurology, Montplaisir et al.,12 using pramipexole, and Wetter et al.,13 using pergolide, confirmed the findings of other recent controlled studies14,15 that dopamine agonists are effective in the treatment of RLS. Montplaisir et al.12 found that, compared with placebo, pramipexole, a D2 and D3 receptor agonist, was associated with a robust reduction in the PLMS index at doses of 0.375 to 0.75 mg/day. Furthermore, pramipexole markedly alleviated leg discomfort at bedtime and during the night as measured by home questionnaires. The study by Wetter et al.13 showed that 0.5 mg of pergolide 2 hours before bedtime significantly reduced PLMS and prolonged total sleep time by 2 hours compared with placebo. In addition to the sleep benefits, pergolide produced a meaningful improvement in quality of life. It is not clear which of the available dopamine agonists is most potent against the symptoms of RLS. Furthermore, longitudinal studies are needed to determine whether the benefits from dopamine agonists are more sustained than those derived from levodopa alone. Opiates, which are beneficial in RLS, may also act through the dopaminergic system, as evidenced by the observation that the beneficial effects of opiates may be blocked by pimozide,16 a dopamine antagonist. The role of opiates in RLS is further supported by the finding that naloxone, an opiate antagonist, reverses the benefits of opiates.17

Recent imaging studies using functional MRI (fMRI) and PET scanning have drawn attention to the possible role of upper brainstem and diencephalon in the pathogenesis of RLS. Turjanski et al.18 provide evidence of striatal dopamine receptor dysfunction in RLS. In this issue of Neurology, they report that patients with RLS have normal ¹⁸F-dopa striatal (putamen) uptake, but D2 binding, determined by ¹¹C raclopride PET scan, is mildly (10%), but significantly, reduced in the putamen. These findings agree with those of Staedt et al.,19 who found reduced striatal D2 binding using ¹²³I-IBZM (Iodobenzamide) SPECT in patients with PLMS. Turjanski et al.18 suggested that reduced putamen D2 binding indicated an increase in endogenous dopamine in RLS. In a study by Montplaisir et al.,20 levodopa-responsive RLS patients showed increased CSF levels of dopamine and its metabolite homovanillic acid, supporting the hypothesis of increased dopamine release or turnover. Turjanski et al.18 further suggested that reduced putamen D2 binding in RLS could be due to decreased central dopaminergic transmission secondary to prefrontal overactivity, which inversely correlates with striatal dopamine transmission. However, absence of prefrontal overactivity by fMRI in RLS argues against this hypothesis. A recent fMRI study21 showed that RLS patients with only sensory symptoms had activation of bilateral cerebellum and contralateral thalamus, whereas those with the additional features of periodic limb movements in wakefulness also showed increased activation of the red nuclei and of brainstem sites in the region of the reticular formation.

The role of impaired central dopaminergic transmission in RLS is further suggested by the overlap between some symptoms of PD and RLS. For example, akathisia—an inner feeling of restlessness frequently associated with complex stereotypic movements—may be seen in patients with PD as well as during chronic treatment with dopamine receptor blockers. In contrast to RLS, akathisia usually does not have diurnal variations and is not typically associated with paresthesias or dysesthesias. It is possible, however, that separate dopaminergic systems are involved in PD and RLS. Whereas the nigrostriatal dopaminergic system is primarily involved in PD, other dopaminergic systems, such as the diencephalic-spinal dopaminergic system, may be involved in RLS.22 In support of this hypothesis is the observation that diencephalic A11 dopaminergic cells project to the spinal cord and that lesions of this midbrain region with 6-hydroxydopamine in a rat produce behavioral features similar to RLS.23

Several electrophysiologic studies have suggested that brainstem or the spinal cord may be the site of generation for PLMS in RLS. Absence of cortical prepotentials on back-averaging,24 normal EEG, and absence of high amplitude cortical potentials in the somatosensory evoked response argue against these movements being of cortical origin. Several studies of patients with RLS or PLMS, however, have found some support for the presence of hyperexcitable brainstem reflexes. Briellmann et al.25 and Wechsler et al.,26 for example, found enhanced excitability of the late component of the blink reflex. In contrast, Bucher et al.27 found no abnormalities of the blink reflex in such patients. These studies, however, need to be repeated during both the symptomatic and asymptomatic periods. In this issue of Neurology, Tergau et al.28 studied intracortical inhibition in 18 RLS patients and 17 age-matched controls by using paired transcranial magnetic stimulation technique. They found a significant reduction of intracortical inhibition in RLS, suggesting motor cortex disinhibition, possibly as a result of subcortical mechanisms.

The possibility remains that some involuntary movements, including PLMS, may be of spinal or propriospinal origin. The presence of PLMS in patients with spinal cord lesions, including complete thoracic transection,29,30 provides support for the location of a generator for PLMS in the spinal cord. It is well recognized that generators for cyclical motor behaviors (e.g., locomotion) exist in the isolated spinal cord.31 Thus, suprasegmental spinal cord lesions may disinhibit the lumbosacral spinal cord generator to produce PLMS. A recent study using flexor reflex response32 in seven patients with RLS–PLMS and in 10 normal controls provided evidence of enhanced excitability of the spinal cord mechanisms, facilitated by the loss of supraspinal inhibition. In another study,33 the pattern of muscle activation in 18 patients with RLS–PLMS suggested a propriospinal mechanism. Finally, some preliminary studies support the presence of individual limb oscillators that may be capable of producing PLMS.34

Although the studies reviewed here do not put RLS to rest, they signal an exciting era of clinical and basic research that promises to provide new insights into the pathophysiology of RLS–PLMS. Future clinical, epidemiologic, and genetic studies should build on the existing knowledge by designing controlled studies of well-defined populations of patients using modern neurophysiologic, functional, metabolic, biochemical, and neuroimaging techniques. Because RLS patients are peculiarly susceptible to placebo effects and natural remission of RLS symptoms occurs for prolonged periods in some patients, double-blind, placebo-controlled, multicenter, clinical trials are needed to find the best treatments for these patients. Finally, because of the lack of pathologic material and, consequently, of any pathologic–clinical correlates, a mechanism needs to be developed that would facilitate the harvesting, storing, and processing of neural tissues, including brain, spinal cord, and peripheral nerves, from patients with RLS. A detailed morphologic and biochemical examination of such tissue would undoubtedly further our knowledge about this common and mysterious disorder.