Introduction

Restless legs syndrome (RLS) is a vexing disorder for patients and clinicians. It is traditionally defined as a discomfort, predominantly in the evening and localized to the lower extremities, which compels the afflicted to move. Movement relieves the symptoms although only momentarily. For patients, RLS can be notoriously difficult to describe, as language may fail to characterize the dysesthesia. Symptoms are reported in a seemingly infinite number of ways including: restless, tingling, cramping, painful, numbing, burning, aching, creepy-crawly, itching, and so on. Often patients are fully unable to localize or describe the sensations at all, but instead state that there is something wrong causing an urge to move.

Clinical investigators have long reported that RLS patients have other, non-motor compulsions. Most notably the majority of patients with RLS will describe an urge to eat that coincides with evening motor symptoms [1••, 2, 3•]. This urge prevents the individual from falling asleep until food is ingested, at which point the feeling abates and sleep may be initiated [2, 4–6•]. Karl Ekbom, in his seminal 1960 publication describing RLS wrote the following. “They often have to get up and walk, “like a caged bear,” to quote one of my patients, or they go into the kitchen and get something to eat…” [1••]. Further, this nocturnal eating is not merely “killing time” because patients with other forms of insomnia are more likely to have awakenings than patients with RLS but less likely to eat [3•]. The International Restless Legs Syndrome Study Group has suggested that the name itself be changed from RLS to Willis-Ekbom’s syndrome in part to recognize its non-motor manifestations [7]. In this report, for the sake of simplicity the disorder will be referred to as RLS.

The cryptic nature of this common disorder (approximately 10 % of the population) often results in misdiagnosis and mistreatment leading to complicated amnestic behaviors [5]. Frequently, the only complaint a patient with RLS will give a clinician is that they “can’t fall asleep”. The sleep initiation difficulties are then easily misattributed to a cognitive hyper-vigilant insomnia such as psychophysiological insomnia. Medications for pathophysiological insomnia, such as the ubiquitously prescribed benzodiazepine receptor agonists, suppress memory and executive function. Thus, it may be expected that when RLS patients are prescribed hypnotic agents, complicated amnestic behaviors such as walking and eating emerge.

The objectives of this clinical review are: to demonstrate that restless eating is a common manifestation of RLS, that RLS is commonly mistaken for psychophysiological insomnia, and that mistreatment of RLS with sedative hypnotic medication would be expected to result in the amnestic binge eating of sleep related eating disorder.

Sleep Related Eating Disorder (SRED)

In 1991, SRED was first described and essentially characterized as sleepwalking with eating. It is defined as a parasomnia (abnormal behavior at night) with recurrent episodes of eating after an arousal from sleep occurring in an unconscious, out of control manner. Similar to sleepwalking, patients could not be easily awakened or redirected [8••]. Adverse consequences have included: weight gain, inedible food consumption, dangerous food preparation, dental caries, and hyperglycemia in diabetic patients [5, 8••, 9]. Also similar to sleepwalking, numerous cases in both the original and subsequent reports noted a strong association with sedative medications. Most notably, a rise in SRED cases has paralleled the utilization of benzodiazepine receptor agonists (BRA) [5, 8••–11].

SRED is a common condition. A self-administered questionnaire determined a prevalence rate of 5 % among a group of college students [12]. This finding was similar to a survey of 1235 general psychiatry patient’s which noted a 4 % lifetime prevalence of SRED [13]. In patients with daytime eating disorders, the prevalence is even higher at 17 % among in-patients and 9 % among out-patients [12].

SRED patients described a long history of nocturnal eating (mean duration >10 years) and nearly all report eating on a nightly basis [9]. A substantial proportion (23 %) describes eating greater than five times a night [14]. The majorities (60-83 %) of reported cases are female and nocturnal foods are higher in carbohydrates and fats then daytime ingestions [8••, 9, 14]. This condition is associated with weight gain and obesity; however, a causality has not been established. In the original 1991 case series nearly half of all patients fulfilled established criteria for being overweight [8••] and in a follow up report 44 % of patients claimed that greater than 20 % of their excess weight was related to nocturnal eating [4].

Amnestic food preparation can be dangerous. SRED patients have reported injuries such as drinking excessively hot liquids, choking, and lacerations. Furthermore, inedible and noxious compounds have been consumed such as: egg shells, coffee grounds, sunflower shells, cigarettes, glue, and cleaning solutions. Finally, patients with food allergies have ingested substances that during the daytime they take extreme precautions to avoid [4, 9, 15].

SRED and Sedating Medications

Reports have noted that patients with SRED are often primed by sedating medications and that wakeful nocturnal eating typically occurs in the absence of these hypnotic agents. The first case of amnestic nocturnal eating in 1981 was associated with a combination of chlorpromazine, amitriptyline, and methyprylon [16]. Further, the majority of patients in the original SRED series were taking hypnotic medication [8]. Subsequently, SRED has been induced with triazolam, lithium, olanzapine, and risperidone [4, 13, 17–19]. Conversely, among 26 subjects with fully consciousness nocturnal eating all were noted to be sedative free [14].

Recently, the benzodiazepine receptor agonists have been implicated in the vast majority of SRED cases as well as the most dramatic and prolonged episodes [10•, 11, 20–23]. Benzodiazepine receptor agonist’s enhance GABA activity at central GABA A receptors and robustly suppress frontal and hippocampal cortical regions. Zolpidem is the most commonly reported agent to induce SRED and has the highest binding affinity among the benzodiazepine receptor agonists. Of note, zolpidem and similar agents do not by themselves activate SRED but instead disinhibit the behavior in a patient population at risk for nocturnal walking and eating [3•].

SRED and RLS

A report in 2002 described five middle-aged patients with intermittent episodes of conscious nocturnal eating prior to starting zolpidem. Soon after initiating zolpidem each patient described amnestic nocturnal eating that stopped with discontinuation. Interestingly all had a history of RLS [10•]. This first series of zolpidem induced SRED along with a judicious review of the literature would suggest a causal relationship between RLS and SRED. This conclusion is based upon similarities clinical course, epidemiology, polysomnography, and treatment response. The evidence is especially compelling in cases of SRED induced by sedative medications where the mistreatment of RLS as psychophysiological insomnia is plausible and, as explained below, amnestic eating the expected result.

As RLS affects approximately one individual out of ten it is a common cause of sleep initiation and maintenance difficulty [5, 24, 25]. Similar to SRED, RLS has a higher prevalence in women [5, 24].

Patients with RLS have a high prevalence of both amnestic SRED and wakeful restless nocturnal eating. A survey of 100 RLS patients reveals a high prevalence of SRED in RLS (33 %) compared to normal controls (1 %) [6•]. The authors pondered whether the compulsive nocturnal eating was related to underlying RLS brain pathology or whether nocturnal eating was merely “killing time”. This question was addressed in another study of 130 patients with either RLS or psychophysiological insomnia who presented to a sleep disorders center. This report noted that 61 % of RLS patients described either restless nocturnal eating (25 %) or SRED (36 %). Conversely only 12 % of patients with psychophysiological insomnia described nocturnal eating, and no patients met criteria for SRED. This study suggests that nocturnal eating in RLS is not merely “killing time” as psychophysiological insomnia patients were more likely to have prolonged (> 5 minute) nightly awakenings (93 %) compared to patients with RLS (64 %) [3•].

While restless nocturnal eating and SRED are common in RLS, RLS is the most frequently reported comorbidity among patients with SRED. RLS was noted in original and subsequent follow up SRED reports [4, 8••]. Further, the largest polysomnographic study of SRED 77 % had confirmation of wakeful RLS and/or periodic limb movements during sleep [14]. The relationship between RLS and SRED has been particularly profound among patients with medication induced SRED (see below) [3•, 10•, 11, 23, 26–28].

As noted above, RLS is challenging to diagnose as symptomatic criteria fails to identify many cases and atypical variants could go unrecognized [29–31]. Thus it may be expected that many cases of subclinical RLS exist in the SRED population. Intriguingly, among SRED reports where RLS was not addressed, there is a peculiar incidence of conditions frequently comorbid with RLS such as Parkinson’s disease and narcolepsy [32–34]. Further medications such as quetiapine that have been demonstrated to induce RLS [35] have also been noted to induce SRED [36].

Restless nocturnal eating closely resembles RLS motor activity. To review, RLS discomfort prevents sleep and compels movement. Movement then subsequently relieves the discomfort [37]. In restless eating, patients state that either prior to falling asleep or after an awakening they have a compulsion to eat (without hunger) that prevents them from falling back asleep. After food is ingested the feeling abates and sleep may be initiated [2, 4–6•].

Compulsive nocturnal eating is not unexpected as RLS patients often describe other non-motor nocturnal urges. Among tobacco smokers the presence of RLS predicts the likelihood of nocturnal restless smoking. This compulsion to smoke is distinct from daytime smoking urges and not present among smokers without RLS [38]. Affected individuals claim that they wake up from sleep and are unable to return without smoking. Intriguingly, among RLS patients with nocturnal smoking, SRED is common (83 %) and both phenomena often begin simultaneously [2].

Similar to RLS [39], SRED appears to be caused by a dopaminergic mechanism [4, 14, 39–42]. First, motor restlessness, smoking, and binge eating are all dopamine mediated behaviors [39, 40]. Further, as noted above, a polysomnography study of 35 SRED patients demonstrated that 77 % had confirmation of RLS and/or periodic limb movement during sleep [14]. Also, rhythmic masticatory muscle activity and bruxism are both dopaminergic phenomena [14, 41] associated with RLS [42] and commonly seen in SRED [4, 14].

It has been debated whether the restless eating and SRED seen in patients with RLS are caused by dopamine agonists as these agents are known to trigger daytime impulsive behaviors such as gambling [43–47]. However the preponderance of evidence suggests that dopamine agents are not the cause of nocturnal feeding behaviors and instead that restless eating is a manifestation of RLS. First, dopamine agents suppress feeding behavior in animal models [48]. Second, several reported cases of SRED have resolved with dopamine agonists [4, 27, 49, 50]. Third, in a survey of patients with both SRED and RLS, ten patients reported that nocturnal eating emerged prior to or concomitant with motor restlessness and none reported that nocturnal eating arose after the start of dopaminergic therapy. Also, subjects whose nocturnal eating symptoms were under control were more likely to be on dopaminergics than subjects who continued to have nocturnal eating [6•]. Fourth, a double-blind placebo controlled trial of pramipexole, a dopamine agonist, in SRED demonstrated improved sleep and reduced nighttime activity and no increased feeding activity [51]. Fifth, another series prospectively monitored therapy outcome in 44 RLS patients previously unexposed to dopaminergics. In this population the frequency of both restless eating and SRED diminished by half with dopaminergics and consistent with other reports, demonstrated a clinical response in parallel to motor RLS symptoms [3•]. Finally, treatment with dopaminergic agents appears to improve other non-motor RLS compulsions that frequently coexist with SRED. In particular, all patients who reported a resolution of nocturnal smoking had been treated with dopaminergic agonists [2].

As the majority of RLS patients are predisposed for eating, greater than 60 % in one survey [3•], then amnestic SRED would be the expected result when RLS patients are treated with agents that suppress memory as well as executive function, such as benzodiazepine receptor agonist’s. Further, while RLS is distinct from, it is commonly mistaken for psychophysiological insomnia. Thus, it may be expected that many RLS patients will be inappropriately treated with therapies such as benzodiazepine receptor agonists that potently suppress hippocampal and frontal cortices. Thus it is not a surprise that in one report, 80 % of RLS patients exposed to sedative-hypnotics had subsequent amnestic SRED or sleepwalking behavior [3•]. In addition to the first case series (describe above) of five zolpidem induced SRED patients all of whom were noted to have RLS, others have commented that RLS appears to be ubiquitous in the setting of zolpidem induced SRED [10•, 28]. In fact, there have been no reports of zolpidem induced SRED where RLS was explicitly considered, and subsequently not discovered [3•, 10•, 11, 23, 26–28, 50, 52].

Tellingly, SRED among psychophysiological insomnia patients treated with sedative hypnotics is rare. Among 25 psychophysiological insomnia patients treated with either a benzodiazepine or benzodiazepine receptor agonist, only two reported amnestic behavior, and in neither case did the events persist [3•]. These findings are consistent with previous reports where SRED and sleepwalking are rare (1 % or less) in zolpidem treated psychophysiological insomnia patients when RLS has been carefully excluded [53].

Two recent case reports from two different groups of investigators help illustrate the intimate relationship between the motor restlessness, restless eating, and SRED. The first is a 48 year-old female with a 10-year history of difficulty falling asleep due to compelling motor restlessness. In addition to RLS symptoms she developed nocturnal eating in the absence of hunger and when treated with hypnosedatives consumed inedible substances such as soap. When she was taken off of triazolam (a short acting benzodiazepine) she became more aware of her behavior but the restless eating persisted. Later, despite the addition of clonazepam (a long acting benzodiazepine) her restlessness (both motor and eating) continued until she was started on low doses of pramipexole. This dopaminergic therapy immediately resolved the nocturnal eating as well as improved her motor symptoms [50]. The second case is of a 74 year old female who presented 20 years prior with history of severe RLS and SRED. Both the SRED and RLS were well controlled with anti-RLS opioid and dopaminergic therapy. Subsequently, the patient had a relapse of RLS (bilateral symptoms) as well as SRED with the eruption of a right lower extremity zoster. Compellingly, the motor restlessness, restless eating, amnestic eating, and skin lesions all waxed and waned in parallel. Medications were not adjusted prior to, during or after the zoster event [27].

SRED, RLS and NES

These findings may help unify the two conditions of dysfunctional nighttime eating, SRED and the night eating syndrome (NES). NES presents with evening hyperphagia, wakeful nocturnal eating, and morning anorexia [54]. Both SRED and NES share many similarities: a chronic course, familial inheritance, comorbid neuropsychiatric disease, and weight gain [55]. In clinical practice differentiating the two is often difficult, as individual patients may have overlapping features.

The purported pathophysiological mechanisms of both NES and SRED are similar and complementary. NES is attributed to an abnormality in the circadian timing of caloric intake relative to sleep while SRED is a breakdown in nocturnal fasting mechanisms [55, 56]. These explanations are not mutually exclusive and the restless nocturnal eating of RLS may also explain the circadian nature of NES. RLS symptoms, both motor and non-motor, have circadian oscillations that reach a late evening crescendo in parallel to the evening hyperphagia of NES [29, 57]. One study investigated NES patients with polysomnography and reported reductions in sleep duration and efficiency suggesting an underlying sleep disrupting process. No comment was made regarding the presence or absence of RLS or periodic limb movements [58].

The divergent clinical fields in which these disorders were identified helps explain their current classification. Clinical investigators in sleep medicine characterized SRED while NES was discovered and researched by experts in eating disorders. Historically, SRED has been distinguished from NES by amnestic eating, or “sleepwalking with eating”. However as demonstrated above, SRED predominantly occurs among RLS patients on sedating agents. Stated another way, most SRED cases are merely adverse drug events [59].

The natural variation in the timing of peak nighttime restlessness, either prior to (evening), or following (nocturnal) the onset of sleep could explain how a spectrum of nighttime eating became split into two disorders. In particular, patients with early nighttime (evening) restless eating would most likely arrive at an eating disorder center as their distressful behavior occurs during wakefulness. Conversely, patients who have late nighttime (nocturnal) restless eating would more likely present to a sleep disorders center as the difficulty is a failure to maintain somnolence. I speculate that NES and SRED may ultimately be unified under a spectrum of restless nighttime eating. NES as a manifestation of predominantly evening restlessness while SRED is the result of sedative medications prescribed in the setting of nocturnal restlessness.

Regardless, SRED and NES are reversible conditions and thus represent unique opportunities in the struggle against obesity [55]. Integration between sleep and eating disorder medicine is of paramount importance to better unravel pathophysiological mechanisms and reverse the contribution of nighttime eating to weight gain. Engagement between investigators would enhance clinical research and help achieve the ultimate goal of effective therapy for all patients.

Conclusion

RLS is an enigmatic disorder with various presentations and is easily misdiagnosed. These challenges are nicely summarized in the title of a recent RLS report, “rarely diagnosed and barely treated” [32]. Unfortunately, when RLS is mistakenly treated as psychophysiological insomnia SRED is often the unintended result. Table 1 provides a summary of the evidence suggesting that nocturnal eating is a manifestation of RLS and SRED the result of mistreating RLS with a sedative hypnotic agent.

Table 1 Evidence suggesting that nocturnal eating is a non-motor manifestation of RLS and SRED the result of RLS treated as psychophysiological insomnia
Full size table