Restless Legs Syndrome in End-Stage Renal Disease

The high prevalence and severe nature of RLS in ESRD potentially provide some tantalizing clues as to the pathogenesis of RLS. ESRD is generally defined as the irreversible loss of renal function, to a creatinine clearance of less than 15 ml/min per 1.73 m². It is present in roughly 400,000 adults in the United States, with an incidence of 260 cases per million per year in whites and 840 per million per year in African Americans. The incidence of ESRD is increasing by 6% per year, probably as a result of the increasing prevalence of diabetes, which accounts for nearly half of all cases of ESRD.² Unfortunately, the long-term prognosis for ESRD is poor, with mortality rates of roughly 20% per year. Over 80% of individuals with ESRD are on hemodialysis, although some people use peritoneal dialysis and others receive transplanted kidneys, often after a period of receiving one of the types of dialysis.

Roughly 60% to 80% of patients with ESRD describe difficulty with nocturnal sleep quality and quantity and/or excessive daytime sleepiness.³ Multiple causes for sleep disturbance are present in those with ESRD. Depressive and anxiety disorders, poor sleep hygiene (e.g., daytime napping), pain, pruritis, symptoms of heart disease (nocturnal dyspnea or chest pain), and symptoms of diabetes (neuropathic pain, fluctuations in blood sugar) are all common in ESRD, and all may contribute to difficulties with nocturnal sleep. Among primary sleep disorders, central and obstructive sleep apneas and RLS are commonly observed at higher rates than in the general population.³

Prevalence and Characteristics of Restless Legs Syndrome in End-Stage Renal Disease

In 1966, Callaghan⁴ first described five patients with classic RLS in ESRD, which he thought was related to subclinical neuropathy. He also described RLS in the upper limbs and leg movements during sleep in two of these patients, both of which are common associated symptoms of RLS. A number of studies have reported the association of ESRD and RLS and have further clarified the prevalence, characteristics, and consequences of the disorder in this population.

Estimates of the prevalence of RLS in ESRD vary between 15% and 70% in the published studies (Table 25-1). A number of potential factors contribute to this variability, including the method of patient recruitment, method of diagnosing RLS (questionnaire versus interview), the ethnicity of the population, adequacy of the anemia treatment, use of intravenous iron in treatments, and the severity of RLS required for a diagnosis. There does not appear to be any difference between rates of RLS in those receiving hemodialysis from those getting peritoneal dialysis.⁵ There are no data on individuals with acute renal failure or on rates of RLS before dialysis dependence.

The RLS phenotype in ESRD appears to be very similar to that of primary RLS. There is some suggestion, from studies including both ESRD-related and primary RLS, that the daytime symptoms and nocturnal sleep disruption are more severe in the former and in particular the leg movements in sleep may be more pronounced.⁶ Nevertheless, diagnosis of RLS in ESRD is complicated by comorbid disease commonly observed in those with ESRD, particularly peripheral neuropathy. For instance, one Italian study demonstrated that a questionnaire similar to that usually used as a screening tool for RLS had modest specificity (81.6%) with low sensitivity (44.7%), giving a low positive predictive value of 47.4% in identifying ESRD patients with RLS.⁷ Less than half of the patients identified as having RLS by the questionnaire actually had the disorder. This raises some uncertainty about the true prevalence of RLS in dialysis, but even with these limits the best estimates indicate a very high RLS prevalence.

Dramatic ethnic variation has been reported for prevalence of ESRD-related RLS, with northern Europeans generally demonstrating higher rates than those reported in India,⁸ Japan,⁹ and Hungary.¹⁰ However, the wide variability in prevalence rates found across all ethnicities in Table 25-1 casts doubt on the validity of the often reported regional or ethnic differences in RLS prevalence. Similarly, it is unclear whether African American patients with ESRD have lower¹¹ or the same¹² rates as whites living in the same areas.

Etiology of Restless Legs Syndrome in End-Stage Renal Disease

The pathogenesis of RLS in ESRD is largely unknown, although treatment studies suggest that iron and anemia status are significant factors. A number of studies have assessed large groups of such patients and compared a variety of variables in ESRD patients with and without RLS, looking for factors that might distinguish the two groups. The majority of such analyses have been negative. Although Collado-Seidel and colleagues¹³ found higher parathyroid hormone (PTH) levels in ESRD patients with RLS than in those without RLS, Stepaniski and colleagues¹⁴ found the opposite result, with lower PTH levels in ESRD-related RLS patients, suggesting that this finding and that by Collado-Seidel and colleagues were probably type 1 errors resulting from inadequate statistical correction for the number of factors evaluated. Walker and colleagues¹⁵ found that BUN and creatinine levels were higher in those with RLS and ESRD, but again these results did not use statistics with adequate control for the multiple comparisons and the elevation in creatinine levels did not meet the usual criteria for statistical significance (p >.05). These results, not surprisingly, have not been confirmed in other studies. Similarly, some authors have found that duration of dialysis dependence or number of dialysis sessions predicts RLS in ESRD, whereas others have not.^16,17 This disagreement cannot be explained by the problem of not correcting for multiple comparisons, but there are significant confounding factors affecting dialysis duration that likely differ between dialysis programs. Of note is that most studies have not found a consistent relationship between ESRD-related RLS and other common predictors of RLS in the general population such as age, gender, and iron indices.^10,15,18,19 Similarly, although some authors have found a higher prevalence of (only) severe RLS in diabetics with ESRD,^5,16 others have not.^18,20 One study found an increased history of use of calcium channel blockers.²¹

Given the inconsistency and negative findings of the majority of investigations into correlates of RLS in ESRD, our understanding of the pathogenesis of RLS in this condition is limited to treatment studies. Extrapolation of theories regarding primary RLS pathophysiology to ESRD-related RLS seems reasonable, although little work has been performed testing such theories. For instance, no neuroimaging or neuroendocrine assessments of dopamine function have been performed in ESRD-related RLS. Although rates of familial RLS in those with ESRD appear to be similar to those in the general population,²² genetic studies, which have been of value in defining chromosomal markers in familial primary RLS, may still help define an underlying genetic predisposition to develop ESRD-related RLS.

Only cursory examination of peripheral nerve functioning has been performed in ESRD patients with RLS. The almost universal, complete, and immediate relief from RLS produced by renal transplantation does not occur for peripheral neuropathy associated with ESRD. Thus, peripheral nerve status does not seem likely to be a significant factor explaining the high rate of RLS in ESRD.

Finally, although normalization of the hematocrit with supplemental erythropoietin has been shown to reduce the PLM index during sleep,²² serum iron measures do not predict those patients with ESRD who develop RLS.^15,18 However, some of the usual serum measures of iron status (e.g., ferritin) are distorted by the disease process in ESRD. Alternative evaluations available for iron as significant factors in the pathogenesis of RLS are the clinical responses to aggressive iron treatment (combined with erythropoietin) and the post-transplantation status. Fortunately, here we have better data. First, there is a blinded, placebo-controlledation study of intravenous iron administration that showed effective an dramatic reduction in RLS symptoms,²³ even though iron markers were normal before intravenous iron treatment. Second, there is and anecdotal impression that rates of RLS in ESRD have decreased since the widespread adoption of supplemental intravenous iron administration combined with erythropoietin. Third, RLS rarely occurs among post-transplantation patients, but the small number with RLS have significantly lower hemoglobin levels and more ID than those without RLS.²⁴ Further work in this area, in particular with assessment of central nervous system indices of iron storage and function, will be important. At this point, iron insufficiency is the only putative cause of RLS in ESRD that has some experimental support.

Restless Legs Syndrome Morbidity in End-Stage Renal Disease

RLS has been associated with a number of adverse outcomes in ESRD: interference with sleep, excess daytime sleepiness, premature dialysis sign-offs, reduced quality of life, increased rates of depression and anxiety disorders, and increased mortality rates. Unfortunately, most of the studies are open to the criticism of not controlling for confounding factors, and thus it is possible that some of these adverse outcomes may not be caused by RLS.

Although it is not a criterion for RLS, sleep disturbance is one of the most salient symptoms of RLS, and the one that often drives patients to seek treatment. All studies examining self-reported sleep in patients with RLS in ESRD have demonstrated severe impairments in sleep, including difficulty falling asleep, frequent nocturnal awakenings, early morning awakening, and shortened total sleep time, with some data suggesting worsened sleep with increasing severity of RLS.³ Studies using polysomnography to assess sleep in ESRD patients with RLS have generally shown prolonged sleep latency, increases in the percentage of light sleep (stages I and II), and reductions in total sleep time and sleep efficiency,^22,25 which is similar to the sleep of RLS patients in general.²⁵ ESRD patients with RLS also consistently take more hypnotic medications than those without RLS.^5,14,16,17

The sleep disturbance in these patients is certainly a result of the sensory dysesthesia and motor restlessness present during wake but may also be related to periodic leg movements of sleep (PLMS), which are present in the majority of RLS patients but also in some patients with ESRD without RLS. In those with ESRD and RLS, the index of PLM per hour of sleep has been noted to be higher than in idiopathic RLS and to be inversely correlated with reports of subjective sleep quality,²⁴ as well as to quality of life.²⁶

Excessive daytime sleepiness (EDS) is present in 30% to 50% of patients with ESRD, by either physiologic (Multiple Sleep Latency Test [MSLT]) or subjective (Epworth Sleepiness Scale [ESS]) measurement.^27,28 Although other features of ESRD, such as blood urea nitrogen,²² and obstructive sleep apnea, certainly relate to this finding, it appears that RLS and/or PLMS is closely associated to patients’ EDS. Both Hui and coworkers²⁹ and Gigli and coworkers²⁰ found an association between ESRD patients with RLS and excessive daytime sleepiness, the latter reporting that scores on the ESS greater than 9 were over five times (16% versus 3%) as likely for those patients reporting RLS as for those without these symptoms. Nevertheless, the average ESS values reported (of ≅4) are certainly within the normal range. PLMS may also relate to daytime sleepiness. Hanly and coworkers²⁸ investigated 24 patients with ESRD and found that the PLM index was significantly and substantially higher in those with EDS than those without EDS (57 versus 6 per hour of sleep), whereas multiple indices of sleep apnea severity were not significantly associated with measures of daytime sleepiness. Stepanski and coworkers,¹⁴ in a smaller group of ESRD patients without obstructive sleep apnea, found similar results.

Compliance with hemodialysis (hemodialysis) treatments, which requires relative immobility for 4 hours, 3 times per week, is suboptimal for the general group of patients with ESRD. Roughly 8% of patients in the United States miss at least one hemodialysis session per month, and nearly 20% shorten a session by at least 10 minutes.³⁰ However, for those with RLS, adherence to hemodialysis is one of the most frustrating and uncomfortable aspects of their life. Winkelman and coworkers¹⁸ found that reported premature discontinuation of dialysis was substantially more likely in patients with ESRD-related RLS compared with those without RLS (p <.0001). Nonadherence with the dialysis prescription has been associated with a variety of poor intermediate outcomes (e.g., hospitalization), as well as with premature mortality, the latter being 14% to 30% more likely with just one skipped session per month.^{24,26,30–32}

The studies with an adequate sample size (more than 10 RLS patients) have shown that multiple features of quality of life are worse in patients with ESRD-related RLS than in those without RLS.^5,26 Unruh’s⁵ epidemiologic study of 145 patients with RLS found that mental health composite scores and vitality were substantially worse in those with RLS, even after controlling for sleep quality. Mucsi and associates³³ also found that a kidney disease–specific quality of life was impaired in those with RLS, compared with those without RLS, even after controlling for sleep quality.

The two reports of increased mortality in ESRD-related RLS and a similar finding in ESRD patients with PLMS are possibly the most important and surprising findings related to consequences of RLS. Winkelman and associates¹⁸ first reported a significantly higher mortality rate (odds ratio = 1.85) at 2-year follow-up in ESRD patients with at least moderate RLS compared with those without RLS (total N = 204), which remained significant even after controlling for age, gender, and duration of dialysis (p <.02) (Fig. 25-1). Unruh and associates⁵ replicated this finding in a study of 894 patients with ESRD, even after controlling for multiple potential confounds including coincident disease burden, although the increased mortality risk was present only for those with severe RLS (hazard ratio = 1.39). Benz and associates³⁴ also found elevated mortality rates in those ESRD patients with elevated rates of PLMS by overnight polysomography.

FIGURE 25-1. Increased mortality in those end-stage renal disease patients with restless legs syndrome compared with those without restless legs syndrome.

Reprinted with permission from Winkelman JW, Chertow GM, Lazarus JM. Restless legs syndrome in end-stage renal disease. Am J Kidney Dis 1996;28:372-378.

The causes of this increased mortality in ESRD-related RLS are unclear. The deaths in the Unruh study were not accounted for by an increased mortality from cardiovascular disease or withdrawal of dialysis. Possibilities include a direct effect of RLS, PLMS or poor sleep on longevity, increased rates of sign-offs in RLS leading to increased mortality, or that RLS is a marker for more advanced ESRD. Although the connection with dialysis nonadherence is tantalizing, Unruh’s patients with severe RLS had the same adequacy of dialysis as those without RLS, suggesting that underdialysis was not related to their increased mortality. These issues are obviously of substantial importance in this population, and future studies should address whether effective treatment of RLS improves both hemodialysis adherence and long-term survival.

Work-up of the End-Stage Renal Disease Patient With Restless Legs Syndrome

Once the diagnosis of RLS is made in a patient with ESRD, an evaluation for potential underlying causes and/or important comorbidities is recommended. The primary reversible causes of RLS in this population are ID and medications. Iron deficiency is common in patients on hemodialysis and is believed to be related to chronic phlebotomy, dialysis-related blood loss, and hemolysis.³⁵ It may have become more pronounced since the introduction of erythropoietin administration. Its importance is underscored by its association as an independent predictor of dialysis sign-offs.¹⁸ Iron deficiency is assessed in different ways in patients with ESRD than in those with primary RLS, given the unreliability of a ferritin determination as a reflection of underlying iron status, due to its role as an acute phase reactant. For this reason, morning fasting serum iron and transferrin saturation are commonly used.³⁶

A variety of medications can exacerbate RLS, including serotonergic antidepressants,³⁷ dopaminergic antagonists,³⁸ and withdrawal from opiates.³⁹ Avoidance of the former medications when possible, and avoidance of daytime use of short-acting opioids may reduce the severity or prevent the appearance of RLS altogether.

Obstructive sleep apnea is common in ESRD, and the associated nocturnal awakenings, sleep deprivation, and daytime napping seen with obstructive sleep apnea may exacerbate RLS. Avoidance of the supine sleeping position, weight loss, or use of CPAP may assist with RLS treatment. There are no data linking treatment of underlying diabetes or co-existing peripheral neuropathy to outcomes in RLS.

Treatment of Restless Legs Syndrome in End-Stage Renal Disease

The goals of treatment in the patient with ESRD are to relieve waking discomfort, allow patients to get their prescribed dialysis treatment, and optimize sleep quality and quantity. Both pharmacological and nonpharmacological approaches are of value.

The place of dialysis adequacy on the severity of RLS is unknown. Although K_t/V has not predicted the presence of RLS in patients with ESRD, it is not known whether dialysis time or adequacy has any effect on RLS severity. As described earlier, the importance of treatment of medical comorbidities is unclear except for the importance of iron status. Other nonpharmacological guidelines in those with ESRD are similar to individuals with primary RLS (e.g., avoidance of alcohol and sedating antihistamines, avoidance of sleep deprivation, attempts to avoid immobility except when necessary so as to focus treatments on dialysis and sleep times), and therefore are not further elucidated here.

Given the distinctive pharmacokinetics and metabolism of medications, and therefore the potential for toxicity in the patient with dialysis dependence, pharmacological trials in ESRD have a number of considerations that are distinct from their use in primary RLS.

Four trials using L-dopa have been published for ESRD-related RLS.^40–43 Three were crossover trials—one with an active comparator, comparing L-dopa SR with ropinirole, and the other two against placebo. The other trial was an open-label retrospective comparison of L-dopa against gabapentin. One crossover trial was open label, and the placebo comparisons were double blinded. All studies were small, with only 41 patients receiving L-dopa in all four studies combined. Differing outcome measures were used in the studies to assess RLS severity (none of them used the IRLSSG severity scale). Three of the four studies found an overall reduction in RLS severity of 18% to 34%, using L-dopa doses of 146 to 190 mg, given once per night. The fourth study, comparing L-dopa with placebo, which did not find a treatment difference, used only 100 mg of L-dopa SR. Substantial improvements in sleep were observed in three of four studies with L-dopa. In the study using L-dopa SR, total sleep time was improved 41% (from 4.1 ± 1.7 to 5.8 ± 1.6 hours) but was still significantly less than that obtained with ropinirole (7.9 ± 1.0). Again, in the study using L-dopa SR 100 mg, no differences in sleep duration or latency were observed. In the short-acting L-dopa study employing polysomnography, the PLM index was significantly decreased in the first 4 hours of the night (from 119 ± 56 to 69 ± 61), but had much less effect in the final half of the night, whereas the study using the low-dose SR, the PLM index was decreased throughout the night. L-Dopa was generally well tolerated in these studies, with only one patient withdrawing due to side effects (vomiting).

The dopaminergic agonist ropinirole demonstrated remarkable efficacy in an open-label crossover trial versus L-dopa (noted previously) in 11 patients with ESRD-related RLS.⁴² Dosages ranged from 0.25 to 2.0 mg, with a mean of 1.45 mg, given 2 hours before bedtime. Self-reported severity with the RLS-6 decreased nearly 75%, sleep duration increased 88% (to nearly 8 hours), patient-rated CGI mean improvement was 2.7 (between much improved and very much improved), and four patients reported elimination of all RLS symptoms. No adverse events were reported. Less than 10% of ropinirole is excreted as unchanged drug in the urine, the drug is almost entirely inactivated by metabolism in the liver, and none of the major circulating metabolites have pharmacological activity, thus decreasing the potential toxicity of this agent in ESRD.

Gabapentin has also shown promising efficacy in ESRD-related RLS in one placebo-controlled double-blinded crossover trial⁴⁴ and one open-label comparison with L-dopa.⁴¹ In the crossover study, 12 of 16 patients responded to 200 to 300 mg of gabapentin, whereas only 2 of 16 responded to placebo. Two subjects dropped out due to gabapentin-related lethargy. Gabapentin is not metabolized and is eliminated exclusively by renal excretion. Gabapentin serum concentrations are decreased roughly by one half by 4 hours of dialysis.⁴⁵ Thus, when gabapentin is administered to anuric patients, it is recommended that it be given only after dialysis treatments, which will then replace the drug eliminated during the hemodialysis treatment and maintain steady-state concentrations as if it were being administered TID or QID.

In one short crossover trial,⁴⁶ the α₁-antagonist clonidine (at an average dose of 0.075 mg BID) was superior to placebo in reducing the symptoms of RLS in patients with ESRD. Eight of 10 patients had complete relief of RLS symptoms in the clonidine group versus only 1 of 10 for placebo. However, clonidine is excreted by the kidneys, and thus the half-life is substantially lengthened and serum levels increased in those with ESRD.

To our knowledge, there are no studies of opioids in ESRD-related RLS, although these medications are considered a good second-line treatment in primary RLS. The elimination half-life of opioids is prolonged in renal failure, and thus the potential for sedation and respiratory suppression (including central sleep apnea) is increased. Although meperidine, in particular, has been associated with toxicity, fentanyl, oxycodone, and codeine are believed to be the safest of these compounds. None of the opioids is known to be removed through dialysis.

In a patient with low serum iron, either the oral or intravenous route is used as supplemental sources. For oral iron, Fe²⁺ 150 to 200 mg of elemental iron per day is recommended to keep hemoglobin at 11 to 12 g/dl, ferritin greater than 100, transferrin saturation greater than 20%, and/or reticulocyte hemoglobin content (CHr) greater than 29 pg. Intravenous iron is often given at the time of dialysis sessions, and 1 g is generally administered over 10 dialysis sessions when poor absorption or side effects of oral iron occur.⁴⁷

A careful double-blind placebo-controlled treatment trial of intravenous iron in ESRD-related RLS found a significantly larger therapeutic effect of 1000 mg than with placebo.²³ The treatment benefit was present 1 week after infusion, was maximal at 2 weeks, and had lost significance by 4 weeks. No excess adverse effects were observed in the intravenous iron group.

Improvement in symptoms of RLS in patients with ESRD after kidney transplantation was first noted in 1986 in a single case.⁴⁸ More recently, this has been confirmed in larger case series.^49,50 Winkelman and colleagues¹⁸ described 11 patients with ESRD and RLS whose kidney transplantation produce rapid (within 1 to 21 days) relief of RLS. In some patients, failure of the transplant led to reappearance of RLS, and in one patient, retransplantation again led to relief of symptoms. Molnar and colleagues⁵⁰ replicated these findings in a cross-sectional assessment of 816 transplanted patients. They found that estimated glomerular filtration rate predicted reappearance of RLS, with declining kidney function predicting symptoms, such that 2% who had a glomerular filtration rate greater than 60 ml/min/1.73 m² had RLS, whereas 24% of those with a glomerular filtration rate of less than 60 ml/min/1.73 m² had RLS.

Conditions With Iron Deficiency and Restless Legs Syndrome

There is increasing evidence that there is a relationship between ID and RLS. In fact, ID may be a major cause of all secondary RLS not associated with neurologic diseases, including ESRD and pregnancy. The patient with ID may not have anemia. Once anemia is present, the marrow iron stores are zero and the ferritin level is often less than 10 µg/L,⁵¹ but marrow iron stores may be zero even with ferritin levels as high as 100 µg/L. O’Keeffe^52,53 noted the values for normal serum ferritin are usually set by population samples ignoring the biological issues of relation between serum ferritin and bone marrow iron status. When Guyatt and associates⁵⁴ evaluated the standard receiver-operator characteristic (ROC) curve to assess prediction of bone marrow ID from serum ferritin, they found 45 µg/L to be the optimal criterion (Fig. 25-2). These data were obtained in a population with suspected ID, which of course alters the usual ROC curve, but RLS patients can be assumed to represent a population with increased risk of ID. Thus, it becomes appropriate to apply these criteria to all clinical management of RLS. Moreover, both ferritin and percent transferrin saturation should be used to evaluate iron status in RLS. A low value in either one indicates probable ID for an RLS patient: that is, ferritin less than 50 μ/L or transferrin saturation less than 16% indicates probable ID.

FIGURE 25-2. Receiver-operator characteristic curve for serum ferritin predicting iron deficiency determined by bone marrow aspiration from a consecutive series of 259 patients with suspected anemia.

Reprinted with permission from Guyatt GH, Patterson C, Ali M, et al. Am J Med 1990;88:205-209.

Several lines of research demonstrate the importance of iron for RLS. Research data indicate that a major cause of RLS may be a decreased level of availability and acquisition of iron in the basal ganglia (substantia nigra and putamen), adversely affecting dopamine synthesis.^55,56 Some other studies have found that there is a relationship between the concentration of serum ferritin and the severity of the symptoms.^57,58 It is also hypothesized, in one study, that there is a defect in transport of iron into the brain leading to reduced brain iron. This study suggests that RLS is a disorder of blood-brain barrier iron transport.⁵⁹

A causal relationship between iron metabolism and RLS would also explain the circadian pattern seen in the symptomatology of RLS, as there is a drop in iron availability at night, with a 30% to 50% drop in serum iron concentration.⁶⁰

Studies have also investigated if iron supplement is an effective treatment of RLS. In two studies where oral iron was given to those with reduce serum ferritin (below 45 µg/L in one⁶¹ and below 75 µg/L in the other⁶²) both showed reduced RLS symptoms with the treatment. In a poorly designed study that did not control for low serum ferritin there was no benefit for oral iron treatment,⁶³ but in this study the serum ferritin values were high enough that there would be little, if any, absorption of the oral iron and thus any observed treatment benefit would have been the same as the placebo treatment. Intravenous iron has been reported to be effective in treatment in two studies.^64,65 However, there is a need for more data on trials with oral or intravenous iron in RLS, using a randomized placebo-controlled protocol.

The first report in the literature on a claimed relationship between ID and restlessness in the legs was written by an anonymous author as a letter to the editor in the British Medical Journal in 1946.⁶⁶ This short document is also informative about how this patient experienced her symptoms, particularly in relation to her pregnancies, and is quoted in full below:

SIR,-From my own experience of the many minor discomforts to which the flesh is heir, jittery legs is one of the most annoying. I suffered considerably in this way during my first pregnancy; the symptoms cleared up, only to recur with increasing severity during the next pregnancy, and then persisted for the following six or seven years. Medical opinions, inside and outside the family, were vague, and not helpful as regards treatment. However, I did gain complete relief from a course of iron and vitamins, which was prescribed for a definite secondary anemia (confirmed by blood count). Once the count reached normal the “jitters” stopped.

I am writing this in case such a line of treatment would help other cases. The diagnosis of the anemia was arrived at accidentally. I was in good general health. Having “jittery legs“ sounds a stupid form of complaint, but it can be a real burden physically, mentally, and socially, and like chilblains is one of the things for which the medical profession offers sympathy but little relief.-I am, etc.,

“A DOCTOR’s WIFE.”

The lack of awareness of RLS was certainly bigger at that time, only 1 year after Ekbom published his thesis “Restless Legs—A Clinical Study of a Hitherto Overlooked Disease.”¹ Accordingly, there was no therapy known for physicians to offer these patients until many years later. In his 1945 thesis, Ekbom reports three patients with RLS who also had anemia, but he did not recognize the relationship between ID and RLS. Therefore, 10 years later, he alluded to the “Doctor’s Wife” as the first person in the literature who recognized the relationship between iron and restless legs.⁶⁷

In 1953, Nordlander⁶⁴ published a sample of 22 patients with restless legs, who received one or two intravenous injections of iron, which proved very effective, regardless of whether the patient had anemia or ID. In the same paper, he reported that he observed three cases of anemia from different causes (bleeding gastric ulcer, chronic nephritis, and leukemia) where restless legs developed in parallel with the anemia and disappeared when the anemia was cured by blood transfusions. In the patient with chronic nephritis, this parallelism was repeated several times.⁶⁴

In 1955, Ekbom⁶⁷ reported on two men with malignancies (cancer in the urinary bladder and cancer in the gut) who had restless legs. In both of these cases, the main reason for them to see a physician was restlessness in their legs, which had a severe negative impact on their sleep. At examination, anemia and ID were observed. The symptoms of restless legs were improved in these two cases with intravenous injection of iron. In the same paper, Ekbom reported 34 cases (25 female, 9 male) with RLS, where 13 (38%) had low serum iron values, all below the lower limit of normal.

Due to these findings, together with the observations of the cancer cases, Ekbom claimed that there often is reason for the physician to ask for symptoms of restless legs in patients presenting with anemia and/or ID and to perform analyses of iron status in RLS cases. If ID is diagnosed, its cause should always be pursued, at least in males.⁶⁷

In a population study of individuals 25 years old or older, all persons with a suspicion of ID were questioned in 1964 by Aspenstrom⁶⁸ about the compulsory phagomania called pica and about symptoms of restless legs. Pica was almost never mentioned spontaneously, so this looks like the situation with RLS subjects! During 1 month, 80 cases of verified ID were found. Two thirds (42 of 64) of the women and 25% of the men admitted the presence of typical pica. Restless legs were reported in 27% of the women and in 25% of the men. Thus, in this study, pica was found to be the most common specific complaint in female ID, but restless legs was a complaint in nearly half of the female subjects who were examined. Restless legs, particularly in women, blood donors, and gastrectomized persons, should probably be initially treated with iron according to the author.⁶⁸

In 1966, Karl Ekbom Jr.⁶⁹ approached 317 patients during a follow-up examination 7 to 9 years after partial gastrectomy by the Billroth I or II method for gastric or duodenal ulcer. Restless legs syndrome had developed after the operation in 40 cases (12.6%).

In 18 patients, 15 of them women, symptoms of pica and restless legs were present concurrently. The hemoglobin and serum iron values were determined in 13 of these cases; a low hemoglobin level and ID were found in all of them. In the remaining 5 cases, the history gave reason to suspect previous ID anemia. After operation, many patients had repeated, periodic, anemia with concurrent symptoms of restless legs. The latter disappeared when the anemia was treated.

Regular blood donation may constitute a significant cause of ID, especially in female donors.⁷⁰ The recommended intake of iron to substitute for iron loss after one blood donation is 2000 mg. This recommendation is based on the fact that there is a loss of iron of 200 to 250 mg at each blood donation and that the uptake from the gastrointestinal tract of the iron taken orally is 10%. Thus, if the mean intake of iron is much lower than what is recommended after each donation, net losses of iron occur over several years and may lead to a substantial negative impact on iron metabolism in the human body.

Ekbom⁷¹ was the first to report two cases of blood donors, both without anemia, but with low serum iron values, with concomitant RLS. Silber and Richardson⁷² and Kryger and colleagues⁷³ have also reported on cases series with repeated blood donations, and it was hypothesized that their RLS was a result of their ID, due to an insufficient iron substitution after blood donations.

In a Swedish survey, 946 consecutive blood donors ages 18 to 64 years were assessed with use of a questionnaire.⁷⁴ Red blood cell distribution width (RDW) was also recorded, a value of blood cell production that increases in relation to ID. RLS affected 14.7% of the male blood donors and 24.7%, of the female blood donors. The mean intake of iron among the blood donors after each blood donation was only 781 mg, thus much lower than recommended. Among the women, 7.4% presented with an RDW greater than 14.5%, which strongly indicates ID. In this group of women, 37.5% were affected by RLS. The female RLS sufferers were more affected than the female non-RLS subjects with both problems initiating and maintaining sleep, and were also more unrefreshed on awakening. Although the intake of iron was higher among women than among men, the women showed a higher frequency of RLS than the men. Moreover, a relationship between ID and RLS was only shown among the female blood donors. It appears that female blood donors have a gender-based susceptibility for ID.

For those who have RLS, blood donation may necessitate a greater iron intake to attain even higher ferritin levels. This recommendation was made by Ekbom to RLS subjects donating blood as early as 1956!⁷¹

Low-density lipoprotein (LDL)–apheresis (LA) is an extracorporeal procedure used to treat severe hyperlipidemia in patients with coronary heart disease who fail to respond to lipid-lowering diet and drug therapy. LA treatment may induce a lack of iron.

In the course of routine patient care, Tings and collaborators⁷⁵ in Goettingen, Germany, noted that numerous LA patients also had RLS. This observation prompted them to explore the occurrence of RLS in a group of LA patients.

Of 25 LA patients, 12 (48%) were found to have RLS. One patient already had RLS before the LA procedure was initiated. The other 11 patients developed RLS during regular LA treatments. Of the affected patients, 5 (42%) had a positive family history for RLS. Laboratory investigation showed that 11 of 12 patients with RLS had ferritin levels below, or at the lower limit of, the normal range, indicating a depletion of body iron stores. Consistent with this finding, transferrin levels were elevated or in the upper range of normal in these patients. However, there were no significant differences in the iron-binding parameters between the LA patients both with and without concomitant RLS. These data corroborate the data on RLS in Swedish blood donors,⁷⁴ that depletion of iron may be sufficient to induce RLS only in susceptible individuals. These data give further support to the hypothesis that ID is primarily a trigger for RLS in patients who have a genetic predisposition for the syndrome.

Rheumatoid Arthritis

Rheumatoid arthritis is another medical condition associated with anemia and iron status. The relation to RLS has not, however, been well established. Salih and colleagues⁷⁸ in a survey of arthritis patients, reported an RLS prevalence of 25% for rheumatoid arthritis compared with 4% for osteoarthritis. These two conditions share some symptom overlap, but included in the differences between them is the relation to iron. Rheumatoid arthritis commonly occurs with anemia and iron compromise not generally seen with osteoarthritis. Györfi and colleagues⁷⁹ examined the potential relations between iron status and RLS. They found in a small group of 19 rheumatoid arthritis patients that 9 (47%) had RLS, and these had a statistically significant elevation in serum transferrin receptor indicating an ID. These small studies hardly suffice to demonstrate a connection between rheumatoid arthritis, iron status, and RLS, but they are consistent with the body of literature suggesting that when ID occurs with any medical condition it increases the risk of RLS. Future studies of associations between RLS and both arthritis and other pain conditions need to be done with evaluation of the iron status of the patients.

An increased frequency of RLS has been suggested in additional rheumatic autoimmune conditions such as fibromyalgia,^80,81 scleroderma,⁸² and Sjögren’s syndrome.⁸³

It should be noted that even if RLS is associated with these conditions, they are sufficiently uncommon and do not manifest with RLS, so that routine screening of RLS patients for these conditions is not useful.⁸⁴

Diabetes

Some epidemiologic studies^85,86 and clinical series⁸⁷ have found a higher frequency of RLS in diabetics. More compelling evidence has emerged from studies of diabetic clinical populations,⁸⁸ including a case-control study⁸⁹ that showed a three-fold increase in diabetics compared with other endocrinology controls (17.7% versus 5.5%). One factor that has been highlighted in recent studies was the presence of diabetic neuropathy as a risk factor for RLS in diabetic populations.⁸⁹ The likelihood of this risk factor is also supported by a study of patients with diabetic neuropathy who had an RLS frequency of 33%.⁸⁹