Sleep Medicine 13 (2012) 409–413

Contents lists available at SciVerse ScienceDirect

Sleep Medicine

journal homepage: www.elsevier .com/locate /s leep

Original Article

Descriptors of restless legs syndrome sensations

Samantha Kerr a,b,⇑, Warrick McKinon b, Alison Bentley a

a Wits Dial.a.Bed Sleep Laboratory, Brain Function Research Group, School of Physiology, South Africab School of Physiology, Faculty of Health Sciences, University of the Witwatersrand Medical School, 7 York Road, Parktown, 2193, South Africa

a r t i c l e i n f o a b s t r a c t

Article history:Available online 17 March 2012

Keywords:Restless legs syndromeTerminologySemanticsSleep disorderDiagnosisSensations

1389-9457/$ - see front matter � 2012 Elsevier B.V. Ahttp://dx.doi.org/10.1016/j.sleep.2011.11.020

⇑ Corresponding author at: School of Physiology,University of the Witwatersrand Medical School, 7South Africa. Tel.: +27 (011) 717 2464; fax: +27 (011

E-mail address: Samantha.Kerr@wits.ac.za (S. Kerr

Background: Restless legs syndrome (RLS) is characterised by an urge to move in response to unusualsensations in the legs. Patients experience difficulty describing their RLS sensations, resulting in a diverserange of descriptors which have not been fully categorised. The purpose of this study was to describe RLSsensations and to evaluate the accuracy of current diagnostic descriptors.Methods: Forty-one RLS participants completed an interview which involved: providing spontaneousdescriptions of RLS sensations, completing the McGill Pain Questionnaire (MPQ), and selecting descrip-tors from a list of previously published RLS terms (prompted descriptors).Results: The most frequent spontaneous descriptors were: ‘‘irritating’’ (17%), ‘‘painful’’ (17%), and ‘‘urge tomove’’ (24%); prompted descriptors were: ‘‘restless’’ (88%), ‘‘uncomfortable’’ (78%), and ‘‘need to stretch’’(76%); and MPQ words were: ‘‘tingling’’ (56%) and ‘‘jumping’’ (54%).Discussion: The most frequently cited descriptors in this study differ from the terminology used in theRLS diagnostic criteria. Inclusion of these frequently used descriptors may improve the diagnostic accu-racy of RLS. Our data emphasise the need for an international, large scale, multicultural study to deter-mine the most accurate diagnostic descriptors to define RLS more clearly.

� 2012 Elsevier B.V. All rights reserved.

1. Introduction

Restless legs syndrome (RLS) is a condition characterised by anurge to move in response to unusual sensations normally experi-enced in the legs. In the 19th century RLS was deemed a psychiatricdisorder known as ‘‘anxietas tibiarum’’ by Wittmaack owing to thebizarre descriptions of symptoms given by those experiencing RLS[1]. Karl Ekbom described RLS clinically in 1945, giving rise to thename of ‘‘Ekbom syndrome,’’ or the more commonly used ‘‘restlesslegs syndrome’’ [2–4]. RLS was formally classified as a sleep disor-der in the International Classification of Sleep Disorders in 1990[5]. More recently, diagnostic criteria based on patients’ subjectivesymptoms have been developed and validated by the InternationalRestless Legs Syndrome Study Group (IRLSSG) in 1995 [6] and,later, revised in 2003 [7]. The critical components of these criteriaare ‘‘an urge to move’’ and ‘‘uncomfortable and unpleasant sensa-tions’’ in the legs, although some patients report the urge to moveindependent of any associated sensations [7]. It is unclear wherethe terminology used to describe the sensory and motorcomponents in the current diagnostic criteria originate and, to

ll rights reserved.

Faculty of Health Sciences,York Road, Parktown, 2193,

) 643 2765.).

our knowledge, there has never been a comprehensive study doneassessing the scope of the sensations experienced by RLS patients.

In previous studies RLS sensations were classified as parasethe-sia or dysesthesia or described by patients anecdotally as ‘‘tingling,burning, jittery, and prickling,’’ and even as ‘‘ants or coca-cola inthe bones and veins’’ [8]. Parasethesia and dysesthesias are usuallysymptoms of neuropathic pain and, so, the relationship betweenRLS and pain has been investigated. To this end, RLS symptomshave been described as painful in 50% to 80% of RLS patients [9].It has been suggested that the sensations may represent a subclin-ical form of pain as the McGill Pain Questionnaire (MPQ) can beused to assess the severity and quality of these RLS symptoms[10]. The alleviation of RLS symptoms with analgesic medications,amongst other treatments, further supports the concept that painpathways are involved in the sensory symptoms of RLS [11]. RLSdoes not however fit neatly into existing pain models and the rela-tionship between RLS and pain sensations needs furtherinvestigation.

The aetiology of RLS is still largely unknown. While it is obviousto researchers in this field that the sensations of RLS are unique,very little work has been done to look for any common sensoryfeatures to define the sensations more precisely. The purpose ofthe study was to characterise the range of descriptors used by anEnglish speaking South African population to describe their RLSsensations.

Table 1Characteristics of RLS participants (Mean [SD]).

RLS participants Range

Number 41 –Male: female (n) 10:31 –Age (years) 47.0 (13.2) 24–74Duration of RLS (years) 17.6 (12.4) 1–48Age of RLS onset (years) 29.8 (13.6) 10–72RLS severity scale score 20.0 (6.4) 4–30MPQ score 18.8 (9.9) 4–51

410 S. Kerr et al. / Sleep Medicine 13 (2012) 409–413

2. Methods

The methods are divided into two parts. The first part describesthe compilation of a list of RLS descriptors, which is subsequentlyincluded in a structured interview, described in part 2.

2.1. Part 1

We conducted a comprehensive literature search includingpeer-reviewed articles and relevant blogs and websites andextracted all words, phrases, and expressions used in the literatureto describe RLS. The majority of the descriptors from the peerreviewed articles were obtained from a small number of papersthat had extensive lists of terms [7,12–15]. Descriptors thatwere synonyms were condensed into one single descriptor (e.g.,‘‘got to move’’ and ‘‘have to move’’). These descriptors were splitinto two separate lists, one for RLS words (n = 113) and theother for RLS phrases (n = 31). The order of the words and phraseswere randomized on several copies of the lists so that eachparticipant had a different arrangement to avoid biasing any ofthe descriptors.

2.2. Part 2

2.2.1. ParticipantsParticipants were recruited on a voluntary basis by local adver-

tisement. They completed a screening questionnaire which in-cluded the four essential RLS diagnostic criteria questions usingthe exact wording as defined by the International Restless LegsSyndrome Study Group (IRLSSG) [7] and basic demographic dataand confirmed, on history, the exclusion of any known secondarycauses or mimics of RLS. Participants who answered all four ofthe diagnostic questions in the affirmative and had no history ofsecondary RLS or RLS mimics were included in the study. Partici-pants were questioned at the interview stage to again confirmthe absence of secondary RLS and RLS mimics. All participants werefluent in English. Ethical clearance (clearance number M070452)was obtained from the University of the Witwatersrand HumanResearch Ethics Committee and participants signed a writteninformed consent form. All data from participants were coded inorder to preserve participant anonymity.

2.2.2. Study designAll RLS participants were given an interview booklet and were

interviewed by the first author, who has experience in diagnosingRLS and who asked them to complete the following tasks:

1. Describe their RLS symptoms in their own words (spontaneousdescriptors). The interviewer recorded these responses usingboth a digital recorder and by making written records.

2. Complete the International Restless Legs Syndrome SeverityScale.

3. Complete the McGill Pain Questionnaire (MPQ). Participantswere asked to select the most relevant word per group thatrelated to their RLS sensations and to leave out any groups thatdid not feature relevant words.

4. Identify all words and phrases from the lists compiled in part 1of the study which could describe their RLS sensations(prompted descriptors). Participants were told to select asmany words and phrases as they felt were relevant to thedescription of their RLS sensations.

2.2.3. Data organisation and analysis2.2.3.1. Descriptors. The spontaneous descriptors that the partici-pants used to describe their own RLS sensations were split into

two lists, one for words and the other for phrases. Descriptors thatwere synonyms were condensed into one single descriptor (e.g.,got to move and have to move) and the frequency of choice of eachdescriptor was determined. All the duplicate terms were subse-quently removed to provide the sum of spontaneous descriptorsoffered. The number of new descriptors contributed by this partic-ipant group was calculated. The frequencies for each of the chosenprompted descriptors (words and phrases) and words on the MPQwere also calculated. The mean (±SD) number of words andphrases selected was calculated for the spontaneous, prompted,and MPQ descriptors.

Phrases were divided into primarily sensory or primarily motorin description and the number of each determined. Manual step-wise regression was used to obtain the minimum number of wordsthat could be used to describe 100% of participants’ sensationsfrom the list of prompted words (starting with the most frequentlyselected word).

3. Results

3.1. Participant information

Forty-one participants, fulfilling all four of the essential RLSdiagnostic criteria, took part in the study. (Six participants whodid not meet all four of the diagnostic criteria were excluded).The characteristics of the participants, their RLS history, and theirseverity scores are presented in Table 1. There was a broad rangefor age of onset, level of pain, and severity of RLS. A positive familyhistory of RLS was reported by 41% of the participants. The majorityof the participants were treatment naive, 10% of the participantshad previously tried dopaminergic treatment for their RLS(discontinued due to adverse side effects or not obtaining repeatprescriptions), and 27% had tried other treatments (mainly overthe counter medications). None of the participants were currentlyreceiving RLS treatment. All the participants had a minimum levelof a completed secondary education.

3.2. Descriptors

The participants spontaneously provided 62 words and 39phrases in total. Seventeen of these descriptors were not presentin the list of previously published terms (part 1). The most frequentspontaneously offered and selected prompted words and phrasesas well as the words selected from the MPQ are shown in Tables2 and 3. The mean number, per participant, of spontaneous wordswas 2.6 ± 2.0, prompted words was 21.8 ± 12.5, and MPQ wordswas 10.1 ± 4.3. The mean number of spontaneous phrases per par-ticipant was 2.4 ± 1.9 and prompted phrases was 8.5 ± 5.1.

For the spontaneous phrases there was an equal number givenfor the motor and sensory descriptors. Of the top 10 promptedphrases, 80% were motor descriptors, including the first six choices,despite there being an almost equal split of sensory and motordescriptors on the phrases list. The number of descriptors, both

Table 2The eight most frequent words given by RLS participants (n = 41) when asked todescribe their RLS sensations spontaneously, prompted with words from a literaturederived list of RLS descriptors, and asked to choose words from the McGill PainQuestionnaire (MPQ).

TOP ‘‘8’’ words (% of participants that offered/chose the word)

Spontaneous words (%) Prompted words (%) MPQ words (%)

Irritating 17 Restless 88 Tingling 56Painful 17 Uncomfortable 78 Jumping 54Crawling 15 Twitchy 63 Nagging 51Uncomfortable 15 Unpleasant 59 Tiring 44Discomfort 12 Irritating 56 Annoying 41Restless 10 Nagging 56 Tugging 39Tingling 10 Fidgety 46 Dull 34Twitching 10 Jerky 46 Gnawing 29 Res

tless

Uncom

forta

ble

Twitchy

Creep

y-cr

awly

Legs nee

dto

stre

tch

Urge to

move

Legs w

ant t

o move

on their

own

Need to

kick

out l

egs

0

70

80

90

100

WordsA B Phrases

Cu

mu

lati

ve c

ho

ice

of

des

crip

tors

(%

)

Fig. 1. The minimum words required from a literature derived list of RLS

S. Kerr et al. / Sleep Medicine 13 (2012) 409–413 411

words and phrases, required to include all participants are shownin Fig. 1. Four words and four phrases were sufficient to cover allthe participants as far as descriptive words and phrases areconcerned.

descriptors to cover. 100% of the population. The most frequently selected word(A) and phrase (B) is on the left and the additional words and phrases added to theright.

4. Discussion

In this cohort of English speaking South African RLS partici-pants, with a broad range of ages, symptom duration, age-of-onset,and severity of RLS, there appears to be some common themeswhen participants describe RLS associated sensations. Participantsonly provided a small number of descriptors when asked todescribe their RLS, but selected a much larger sample of descriptorswhen provided with a selection of established terms. Differentwords were chosen spontaneously, when prompted, and fromthe MPQ. Spontaneous and prompted phrases, however, were sim-ilar. The participants spontaneously described an equal number ofsensory and motor phrases, however, upon prompting, the motorphrases were the dominant choice.

These findings represent the results from a small sample ofEnglish speaking South African RLS participants. A larger, more cul-turally diverse, sample size may have provided a better represen-tation of the RLS population as a whole, although there is noevidence that RLS descriptors change between populations. We feltthat it was necessary to restrict the initial study to English speak-

Table 3The 10 most frequently selected phrases chosen by RLS participants (n = 41) whenasked to describe their RLS sensations spontaneously and when selecting words froma literature derived list of RLS descriptors (prompted).

TOP ‘‘10’’ phrases (% of participants that offered/chose the phrase)

Spontaneous phrases Prompted phrases

Have/urge to move 24 Legs need to stretch 76Need to stretch 17 Just an urge to move 73Pins and needles 12 Legs want to move on their own/

won’t be still66

Legs refuse to sit still 10 Need to kick out your legs 59Have to walk 7 Legs need to walk/jog 59Worms inside leg 7 The got to moves 51Columns of ants marching up

and down5 Anxiety in your legs 49

Feels as if you’ve hit yourfunny bone

5 Feels like your legs want to jumpoff your body

34

Feels like want to run/exercise/use muscle

5 Legs have too much energy/arefull of energy

34

Muscles aren’t relaxed 5 Nervous legs 32Spontaneous tapping 5There’s something deep inside

the leg5

Want to get up and go 5

ing participants to avoid the confounding variables associated withmultiple language translations and back translations. The findingscould have been biased by the fact that the word ‘‘restless’’ is themost frequently chosen descriptor which emanates from theparticipants familiarity with the phrase ‘‘restless legs syndrome’’.Other words chosen could also have been influenced by priorknowledge of RLS descriptors. Although the study included a rangeof RLS severities, the majority of our participants suffered frommild to moderate RLS and this may have influenced the choice ofdescriptors, although the relationship between descriptors andseverity of RLS is unknown. Another limitation may be the use ofhistory to exclude RLS mimics and secondary RLS, although sixpotential mimics were excluded. The data would indicate that thisgroup of RLS participants are quite consistent in their choice ofwords and phrases, thus reducing the chance of mimics occurringin the sample. The words chosen on the MPQ must also beviewed with caution as, unlike the other selections, participantsare restricted in the number of words they can choose – 20 beingthe maximum. Thus, other words may have been valid but in thesame group as the word chosen and therefore not available forselection.

The word ‘‘restless’’ was only offered spontaneously by 10% ofthe participants, but when these same participants were providedwith a choice of pre-selected descriptors, restless was the most fre-quently chosen word (88%) to describe their RLS sensations. Giventhe name restless legs syndrome, it is assumed that patients wouldexperience restlessness; however, the word restless is not includedin the diagnostic criteria. The official IRRLSSG diagnostic criteria,asks whether a patient experiences an ‘‘uncomfortable’’ or‘‘unpleasant’’ sensation [7]. Uncomfortable was chosen by 78% ofparticipants in this cohort and unpleasant by only 59%. Thesetwo words were selected together by 56.1% of participants. Thewords ‘‘uncomfortable’’ and ‘‘unpleasant’’ are not frequently of-fered when participants spontaneously described their symptoms.One RLS diagnostic criterion requires a patient to feel the ‘‘urge tomove’’ [7]. The relevance of this criterion is reflected by its selec-tion as the most frequent spontaneously offered phrase and thesecond most commonly chosen prompted phrase. The phrase‘‘need to stretch’’ is not included in the diagnostic criteria; how-ever, it was as frequently selected, both spontaneously and whenprompted, as the phrase ‘‘urge to move.’’

412 S. Kerr et al. / Sleep Medicine 13 (2012) 409–413

There appear to be some differences between the terms used inthe diagnostic criteria and those that patients use. It is unclearwhere the wording in the IRRLSSG diagnostic criteria originates.We suspect that the terms ‘‘uncomfortable’’ and ‘‘unpleasant’’ arederived from the original definition of RLS as ‘‘parasethesia’’ or‘‘dysesthesia.’’ Dysesthesia are defined by the European Federationof Neurological Societies as ‘‘abnormal and unpleasant’’ sensations,whereas parasethesia are ‘‘abnormal but not unpleasant’’ sensa-tions [16]. Ekbom’s original description of RLS using his original34 cases of patients with RLS reported their sensations as ‘‘crawl-ing,’’ ‘‘unpleasant,’’ ‘‘irritating,’’ and ‘‘disagreeable,’’ but rarely aspainful, and that patient’s found it ‘‘impossible to keep their legsstill’’ [3]. Previously, far greater emphasis has been placed on thesensory components than the urge to move. Current literature indi-cates that the need to move is far more useful diagnostically thanthe RLS sensations and, as such, diagnostic criteria state that theurge to move may or may not be accompanied by uncomfortablesensations [7]. This may also be due to the lack of extensive worktrying to classify the sensory discomfort more accurately.

A relationship between the sensations of RLS and those of painwas not confirmed by these data. The only common wordsbetween those selected from the MPQ and the spontaneous andprompted words were ‘‘tingling’’ and ‘‘nagging,’’ which were notin the top five words chosen in either the spontaneous or theprompted word group. This may indicate that, although theMPQ may be valid to measure the severity of RLS, the descriptorsin that questionnaire are not sufficient as diagnostic criteria[10]. These data also question the prevalence data of patients com-monly reporting ‘‘pain’’ as a descriptor, although a composite ofwords may indeed feel like pain. In the early literature reportingon RLS, Ekbom repeatedly indicated that sufferers of RLS havedifficulty describing the sensations they experience [3] and thisproblem persists with patients today. A great deal more work isrequired to tease out the relationship between RLS sensationsand pain.

The phrasing of the diagnostic criteria, although a purelysemantic exercise, may influence a clinician’s diagnosis of a pa-tient. Popat et al. pointed out that ‘‘using a consistent set of ques-tions (with respect to wording, response choices, and algorithm forclassifying RLS status)’’ is essential in being able to compare RLSepidemiological studies [17]. Using the published wording of theIRRLSSG should ensure that researchers are using a standard setof descriptors to diagnose RLS, but these descriptors may not besufficiently accurate. Physicians may be primed to relate to thecurrent wording and may not recognise other, more unusual, yetvalid, descriptors.

An extensive study done in numerous countries in the Northernhemisphere (Europe and USA) using the descriptors ‘‘uncomfort-able’’ and ‘‘unpleasant’’ (included in the diagnostic criteria)reported a marked under-diagnosis of RLS [18]. These researchersconcluded that true RLS could be misinterpreted as a variety ofother conditions such as varicose veins, neuropathy, leg cramps,chronic venous insufficiency, or damage to the lumbar spine [18].Alternatively, it has been suggested that other conditions are diag-nosed as RLS as the four diagnostic criteria, as they stand, are notable to exclude RLS mimics [19]. Creating a better word selectionso that patients can better describe their symptoms could assistclinicians in distinguishing RLS from RLS mimics and a differentialdiagnosis could possibly be made based on the word choice of pa-tients. The same process done here for RLS mimics should be donefor the conditions that may mimic RLS to bring clarity to the situ-ation. Having a list of validated descriptors for each condition,including paraesthesia and dysethesia, would assist the primary

care doctor in differentiating the conditions and would thus im-prove the diagnostic accuracy of RLS.

The descriptors selected in this study reflect the choices of anEnglish speaking South African population. The descriptors maydiffer between different cultural and language groups, as well asbetween primary and secondary RLS. There is a need to performa number of studies in English speakers of different culturalbackgrounds and geographical areas to determine the consistencyof the English descriptors shown in this study in order to improvethe diagnosis of RLS. Patient descriptors also need to be assessed inother languages as well as changes that may occur after treatmentand with augmentation.

In conclusion, our participants had a wide selection of words todescribe their RLS, with some unique to individual participants, buta few common descriptors which covered the whole sample couldbe found. Expanding the word choice in the diagnostic criteria mayimprove the accuracy of RLS diagnosis and exclude mimics moreeasily.

Conflict of Interest

The ICMJE Uniform Disclosure Form for Potential Conflicts ofInterest associated with this article can be viewed by clicking onthe following link: doi:10.1016/j.sleep.2011.11.020.

Acknowledgement

Funding provided by the Wits Dial.a.Bed sleep laboratory.

References

[1] Ekbom KA. Asthenia crurum parasthetica (irritable legs). Acta Med Scand1944;118:197–209.

[2] Allen RP, Earley CJ. Restless legs syndrome: a review of clinical andpathophysiologic features. J Clin Neurophysiol 2001;18(2).

[3] Ekbom KA. Restless legs: a clinical study. Acta Med Scand1945;158(Suppl.):1–123.

[4] Wetter TC, Pollmacher T. Restless legs and periodic movements in sleepsyndrome. J Neurol 1997;244(Supp. 1):S37–45.

[5] Thorpy MJC. The International classification of sleep disorders, revised:diagnostic and coding manual. In: Taaaa MJ, editor. RochesterMinnesota: The American Sleep Disorders Association; 1990.

[6] Walters AS. Towards a better definition of restless legs syndrome. TheInternational Restless Legs Syndrome Study Group. Mov Disord1995;10:634–42.

[7] Allen RP, Picchietti D, Hening WA, et al. Restless Legs Syndrome: diagnosticcriteria, special considerations, and epidemiology. A report from the RestlessLegs Syndrome diagnosis and epidemiology workshop at the NationalInstitutes of Health. Sleep Medicine 2003;4:101–19.

[8] Earley CJ, Allen RP, Beard JL, et al. Insight into the pathophysiology of restlesslegs syndrome. J Neurosci Res 2000;62:623–8.

[9] Winkelmann JW, Wetter TC, Collado-Seidel V, et al. Clinical characteristics andfrequency of the hereditary restless legs syndrome in a population of 300patients. Sleep 2000;23:597–602.

[10] Bentley AJ, Rosman KD, Mitchell D. Can the sensory symptoms of restless legssyndrome be assessed using a qualitative pain questionnaire? Clin J Pain2007;23:62–6.

[11] Hening WA, Allen RP, Earley CJ, et al. The treatment of restless legs syndromeand periodic limb movement disorder. An American Academy of SleepMedicine Review. Sleep 1999;22:970–99.

[12] Arbuckle R, Abetz L, Durmer J, et al. Development of the paediatric RestlessLegs Syndrome Severity Scale (P-RLS-SS): a patient-reported outcome measureof paediatric RLS symptoms and impact. Sleep Med 2010;11:897–906.

[13] Cotter PE, O’Keeffe ST. Restless leg syndrome: is it a real problem? Ther ClinRisk Manage 2006;2(4):465–75.

[14] Patrick L. Restless legs syndrome: pathophysiology and the role of iron andfolate. Altern Med Rev 2007;12:101–12.

[15] Picchietti D, Allen RP, Walters AS, et al. Restless Legs Syndrome: prevalenceand impact in children and adolescents – the Peds REST study. Pediatrics2007;120:253–66.

[16] Cruccu G, Anand P, Attal N, et al. EFNS guidelines on neuropathic painassessment. Eur J Neurol 2004;11:153–62.

S. Kerr et al. / Sleep Medicine 13 (2012) 409–413 413

[17] Popat R, Van Den Eeden S, Tanner C, et al. Reliability and validity of two self-administered questionnaires for screening restless legs syndrome inpopulation-based studies. Sleep Med 2010;11:154–60.

[18] Hening WA, Walters AS, Allen RP, et al. Impact, diagnosis and treatment ofrestless legs syndrome (RLS) in a primary care population: the REST (RLS

epidemiology, symptoms, and treatment) primary care study. Sleep Med2004;5(3):237–46.

[19] Hening WA, Allen RP, Washburn M, et al. The four diagnostic criteria forRestless Legs Syndrome are unable to exclude confounding conditions‘‘mimics’’. Sleep Med 2009;10:976–81.