Comorbid Conditions

Over the years, a large number of disorders or other conditions have been suggested to be associated with restless legs syndrome (RLS) (see Chapter 16, Box 16-2). The evidence for some is stronger than that for others, but these associations indicate that many different medical specialists need to be concerned about RLS (Table 21-1). Many of these conditions share an association with depleted iron stores—iron deficiency and anemia, kidney failure, pregnancy, rheumatoid arthritis—but others cannot yet be explained on the basis of a single pathophysiologic diathesis. An intriguing question, now that specific genetic variants associated with RLS have been reported (see Chapter 8), is whether these same variants will be associated with increased risk of RLS in these comorbid conditions.

TABLE 21-1 Specialties Important for Comorbid Restless Legs Syndrome

Polyneuropathies and Other Lower Motor Neuron Diseases

A subgroup of patients with RLS is found to have sensorimotor neuropathies, occasionally lumbosacral radiculopathies or motor neuron disease (amyotrophic lateral sclerosis); however, the relation between RLS and these entities remains uncertain. Before addressing this controversy, I would like to make some general comments. The sensory complaints of primary RLS patients resemble mostly the central type of sensory discomfort or central pain syndrome and are characterized by deep, disagreeable, unpleasant, often nonlocalizable, nonradiating discomfort unaccompanied by objective sensory impairment in contrast to sensory manifestations of peripheral neuropathy or radiculopathy, consisting of positive symptoms of tingling, numbness, burning (hyperalgesia), and stabbing, present in the stocking and glove distribution or radiating along root distribution, accompanied by sensory impairment distally or along the root distribution. However, if polyneuropathy or lumbosacral radiculopathy occurs in patients predisposed to developing RLS symptoms sometime in the future, then the sensory symptoms may be similar in both primary and secondary cases. Lumbosacral radiculopathies or polyneuropathies simply trigger the symptoms in such patients. It is notable that RLS-like symptoms have never been described, to my knowledge, in patients with cervical radiculopathy. For differentiating comorbid from primary RLS, particular attention should therefore be paid to the nature of the sensory complaints and the duration of the disease. Sensory symptoms of polyneuropathy are present either intermittently or throughout the day and not necessarily worse in the evening and are generally not relieved by movements. In about 20% to 25% of primary RLS patients, actual pain rather than intense disagreeable feeling is a prominent symptom. Comorbid RLS patients generally seek a physician’s advice within 5 years of onset of symptoms, but patients with idiopathic or primary cases, if questioned carefully, will often divulge that their symptoms have been present for many years (10, 20, 30 years, or longer). Another important point of distinction is the family history which may be present in patients with primary RLS rather than in comorbid cases.

Ondo and Jankovic¹ reported that 15 of 41 patients with RLS had electrodiagnostic evidence of polyneuropathy or radiculopathy, although only 7 of these 15 patients showed clinical signs of neuropathy. In an early systematic study, Rutkove and colleagues² studied a consecutive series of patients with polyneuropathy and documented only a 5% frequency of RLS, which is similar to what is found in the general population. Studies have found a greater frequency of RLS in diabetic neuropathy and polyneuropathy (see Chapter 27). RLS may also be particularly prevalent in certain types of neuropathies, such as cryoglobulinemic or familial amyloid polyneuropathy. Newer studies have more decisively implicated the polyneuropathies and diabetic neuropathy as associated with RLS. Gemignani anc coworkers³ in a series of 44 consecutive patients with Charcot-Marie-Tooth (CMT) disease found RLS in 10 of 27 CMT type 2 patients (37%) with prominent sensory symptoms and in none of the patients with CMT type 1, suggesting a role for sensory (afferent) input in the pathogenesis of RLS. Certain studies found morphological evidence of subtle nerve damage. Sural nerve biopsy⁴ findings in 7 of 8 patients with primary RLS were consistent with an axonal neuropathy, and skin punch biopsy⁵ in the legs documented evidence of subclinical small fiber neuropathy in 8 of 22 patients (36%) with primary RLS. There may be other explanations for these subtle morphological changes, such as age-related changes in the sural and skin nerves and trauma caused by vigorous rubbing of the legs, resulting from an urge to move and rub to get relief from the discomforting urge to do so. It is not clear why some patients with polyneuropathy or lumbosacral radiculopathy develop symptoms of RLS, whereas many patients with polyneuropathies or other lower motor neuron disorders do not develop RLS. Sometimes motor neuron disease has been mistaken for RLS, but the diffuse fasciculations, muscle weakness, wasting, and presence of upper motor neuron manifestations should simplify the differentiation.

Anemia and Restless Legs Syndrome

There have been isolated reports of iron, cobalamin (vitamin B₁₂), and folate deficiency causing secondary RLS. The strongest evidence is for iron deficiency. Ekbom,⁶ in his original description, mentioned iron deficiency with low serum iron in 25% of his cases. Nordlander⁷ in 1953 reported resolution of RLS symptoms in over 90% of subjects after intravenous iron infusion. The iron deficiency theory for RLS was revived after a long hiatus by O’Keefe and colleagues⁸ in 1994, who reported low iron, measured by serum ferritin levels in RLS patients. Subsequent elegant studies by the Johns Hopkins group of investigators^9–15 firmly established the role of iron in its pathogenesis and therapy (see Chapters 9, 10, and 34). Iron is needed as a cofactor for tyrosine hydroxylase, the rate-limiting enzyme in the synthesis of dopamine; therefore, iron deficiency may interfere with the normal production of dopamine. Furthermore, the D₂ receptor is an iron-containing protein and hence, iron deficiency may impair the normal function of D₂ receptors. Magnetic resonance imaging (MRI)¹⁴ study of the brain and limited pathological¹⁵ findings in RLS patients clearly established brain iron acquisition and storage problem deficiencies in the substantia nigra of these patients. The appearance of RLS symptoms after multiple blood donations^16,17 associated with iron deficiency anemia, low serum hemoglobin and ferritin levels, and improvement after iron supplementation also clearly supports the role of iron deficiency in the pathogenesis of RLS. Reports of failure^18,19 to find a significant difference in serum iron indices or hemoglobin between patients with RLS and control subjects with renal failure and pregnancy as well as a single report²⁰ of a randomized, double-blind placebo-controlled trial of iron supplementation failing to show any significant difference in RLS symptoms with treatment point to the incompleteness of the iron deficiency theory. RLS appears to be a heterogeneous syndrome with several subtypes associated with different neurobiological mechanisms.

End-Stage Renal Disease and Hemodialysis

There is an increased prevalence of RLS (20% to 50% or even higher) with end-stage renal disease (ESRD) on dialysis^18,21–24 (see Chapter 25). RLS in ESRD and idiopathic RLS are clinically or therapeutically quite similar. Whether the risk factor is comorbidity with ESRD or dialysis or both is not known, but most of the reports of RLS are in patients on hemodialysis. There are reports of increased mortality and morbidity in RLS patients on dialysis.^21,25–27 No conclusive biochemical risk factors are specifically associated with RLS except for low iron status as a predictor of poor outcome. Report of disappearance of RLS after kidney transplantation²⁸ suggests that some unknown biochemical or other factors are causing the ESRD RLS symptoms.

Pregnancy and Restless Legs Syndrome

Ekbom⁶ in his original description of RLS reported higher prevalence of RLS symptoms in pregnant women. Subsequent epidemiologic studies (not a population survey) found that approximately 19% to 27% of pregnant women have RLS symptoms, most prominently in the last trimester of pregnancy^29–31 (see Chapter 23). In some pregnant women with previously diagnosed RLS symptoms, the symptoms worsened during pregnancy, but in the majority, symptoms appeared for the first time during pregnancy with subsequent resolution of symptoms around the time of delivery. In a small minority, symptoms persisted after delivery suggesting that pregnancy might trigger RLS in those who are genetically predisposed to develop RLS. Long-term follow-up studies of pregnant RLS patients to determine the natural history of this subtype of RLS are not available. Using a structured clinical interview and IRLSSG diagnostic criteria Manconi and colleagues³¹ studied 642 pregnant women at the time of delivery and one, three, and six months after delivery. They noted a 26.6% prevalence of RLS amongst these women with resolution of symptoms after delivery. In a previous study, Lee and coworkers¹⁹ gave a figure of 23% of pregnant women with RLS. These authors found reduced serum ferritin at preconception and low folate levels both during preconception and pregnancy. In a survey of 300 RLS patients, Winkelmann and associates³² noted worsening of the previous RLS symptoms or the onset of RLS symptoms for the first time during pregnancy in both familial and sporadic RLS but more frequently in the familial than sporadic patients. Several factors have been postulated to be responsible for the exacerbation or occurrence of RLS symptoms during pregnancy, such as iron and folate deficiencies, hormonal changes (e.g., increased levels of estrogen, progesterone, prolactin) and mechanical factors causing lower limb vascular congestion.

Rheumatoid Arthritis and Restless Legs Syndrome

There are two reports^33,34 of increased prevalence of RLS in patients with rheumatoid arthritis (25% to 30%). In contrast, Banno and colleagues,³⁵ in their questionnaire survey of 218 patients with a diagnosis of RLS, and Ondo and colleagues,³⁶ measuring rheumatologic serologies in 68 RLS patients, did not find increased prevalence of RLS in rheumatoid arthritis.

Restless Legs Syndrome and Parkinson’s Disease

The prevalence studies of RLS in Parkinson’s disease (PD) are few and somewhat contradictory (see Chapter 26). Earlier reports of RLS in PD may have been cases of akathisia in PD. From pharmacologic and neuroimaging points of view, there is a link between RLS and PD. Both PD and RLS patients benefit from dopaminergic treatment, although RLS patients need much less dopaminergic medication than PD patients. Neuroimaging studies of the brain (positron emission tomography and single-photon emission computed tomography) show a mild striatal dysfunction in RLS^37–39 in contrast to severe nigrostriatal degeneration in PD. Special MRI study¹⁴ of the brain in RLS shows evidence of decreased iron storage in the substantia nigra of RLS patients, whereas PD patients have increased iron. Limited neuropathological studies¹⁵ suggest no dopaminergic neuronal degeneration in RLS and, therefore, neuronal dysfunction or degeneration in diencephalospinal dopamine system may be important in RLS pathophysiology.

Earlier reports by Ekbom,⁴⁰ Strang,⁴¹ and Lang and Johnson⁴² showed no increased prevalence rate of RLS in PD. A study by Tan and coworkers⁴³ confirmed these earlier observations. On the other hand, Ondo and associates⁴⁴ found a prevalence of 20.8% in the PD population in their clinic in Houston, Texas, and Krishnan and associates.⁴⁵ in a case-control study noted increased prevalence of RLS in a PD population in New Delhi, India. It should be noted, however, that there is no good epidemiologic study comparing a large number of age- and sex-matched control subjects and PD patients to show the prevalence of PD in RLS and in RLS patients who later developed PD.

Pulmonary Disease, Restless Legs Syndrome, and Sleep Apnea

Spillane⁴⁶ described RLS in eight patients with chronic obstructive pulmonary disease (COPD). I noted an association between RLS and sleep apnea in three COPD patients.⁴⁷ Schönbrunn and coworkers⁴⁸ also mentioned an association between RLS and sleep apnea. These three studies, however, were conducted before the Restless Legs Syndrome Study Group (IRLSSG) criteria were established for the clinical diagnosis of RLS. In a study by Lakshminarayanan and colleagues⁴⁹ in 60 sequentially studied patients with obstructive sleep apnea syndrome (OSAS) diagnosed by polysomnography (respiratory disturbance index of more than 10), the authors found a prevalence of RLS in 8.3% of their patients. This figure is, however, not dissimilar to the prevalence figure for RLS, although spouses serving as age-matched control subjects showed a prevalence figure of 2.5%. Delgado Rodrigues and coworkers⁵⁰ observed improvements in sleep apnea and RLS symptoms in 17 patients with coexistent RLS and OSAS following nasal continuous positive airway pressure titration. It is possible that chronic hypoxemia caused by COPD and nocturnal hypoxemia caused by OSAS played roles in the pathogenesis of RLS in these patients. The association of RLS, COPD, and OSAS remains uncertain until good epidemiologic studies in the general population are performed.

Drugs and Restless Legs Syndrome

In scattered case reports and anecdotally, several antidepressants (e.g., tricyclic antidepressants and selective serotonin reuptake inhibitors [SSRIs]), neuroleptics, dopamine antagonists, antihistamines (mostly H₁ antagonists), and lithium have been shown to cause or worsen RLS symptoms^51,52 (see also Chapter 28). Berger⁵³ noted a 27% prevalence of RLS in a highly selective population in a single practice who were treated with tricyclic antipressants and SSRIs. He further observed that 75% of RLS patients had a history of regular intake of nonopioid analgesics and he suggested that the intake of analgesics enhanced the risk for RLS. Leutgeb and Martus⁵⁴ previously also reported a prevalence rate of 27% in a selective population with anxiety and affective disorders on antidepressant medications who had previously taken analgesics including nonsteroidal anti-inflammatory agents (NSAIDs). These authors postulated that neither antidepressants nor neuroleptics, but nonopioid analgesics seemed to be a major risk factor of RLS. The role of antidepressant medications causing or exaggerating RLS symptoms, thus, remains controversial. In a report after a retrospective chart review of 200 consecutive patients referred to their sleep disorder center for evaluation of sleep-onset insomnia, Brown and colleagues⁵⁵ concluded that there was no statistical association between RLS and antidepressant medications.

Other Miscellaneous Conditions Associated With Restless Legs Syndrome

There are some scattered case reports of RLS associated with varicose veins, thrombophlebitis, hyperthyroidism,⁵⁶ hypothyroidism,^56,57 hyperparathyroidism,⁵⁸ meralgia paresthetica,⁵⁹ saphenous nerve entrapment,⁶⁰ multiple sclerosis, myelopathies, Isaac syndrome, stiff-man syndrome, and hyperekplexia.⁵¹ Reports of RLS in patients with cancer or following gastric surgery may have been secondary to anemia and/or polyneuropathy.⁵¹