Restless legs syndrome and its relationship with insomnia symptoms and daytime distress: Epidemiological survey in Sweden


*Jan-Erik Broman, PhD, Department of Neuroscience, Psychiatry, University Hospital, SE-75185 Uppsala, Sweden. Email:


To investigate prevalence estimates of restless legs syndrome (RLS) in relation to insomnia complaints and daytime distress a questionnaire was sent to a randomly selected sample of 1962 inhabitants of Uppsala, Sweden. The questionnaire included questions about sleep and daytime distress and the standardized four-question set for epidemiological settings recommended by the International RLS Study Group. A positive diagnosis of RLS was established in 18.8% of all responders. When the optional question about frequency was applied 5.8% reported frequent symptoms. Insomnia symptoms and daytime distress were significantly associated with the frequency of RLS symptoms.

RESTLESS LEGS SYNDROME (RLS) is a sensorimotor disorder characterized by an almost irresistible urge to move the legs that may cause considerable sleep disruption and insomnia complaints.1 Surveys of the general population have shown that RLS is a common disorder, and has a prevalence ranging between 2.5 and 10%.2

Standardized criteria were proposed by the International Restless Legs Study Group (IRLSSG) in 19953 and these were slightly modified in 20034 when a four-question set recommended for epidemiological surveys also was presented. The first three questions were built upon those validated by Rothdach et al.5 and a positive diagnosis requires affirmative answers to each of these three questions. The fourth, optional question was designed as a convenient measure of severity and had seven response categories.

The proposed four-question set does not yet seem to have been used in epidemiological surveys in the normal population. Therefore, the aims of the present study were (i) to estimate the prevalence of RLS according to the proposed four-question set; and (ii) to further investigate the frequency of RLS symptoms, and the relationship between frequency of RLS symptoms and sleep disruption and daytime distress.


A questionnaire was sent out by post to a randomly selected sample of 1962 subjects aged 20–59 years living in Uppsala Municipality. Approximately every 50th person was drawn from the national registration records. Responses were obtained from 586 men and 749 women, with mean (±SD) ages of 39 ± 12 years and 38 ± 11 years, respectively.

The questionnaire consisted of items from the Uppsala Sleep Inventory.6 The four-question set recommended for use in epidemiological settings by the IRLSSG3 was also included. These questions had been translated into Swedish by a professional translator, discussed in an expert group, and adequately back-translated into English by another professional translator. Last, the Hospital Anxiety and Depression Scale7 was included. The study protocol was approved by the Ethics Committee of the Faculty of Medicine at Uppsala University.


Prevalence rates of RLS symptoms and their relation to frequency

Results showed that 18.8% of all responders gave affirmative answers to the first three questions, indicating an RLS diagnosis. The prevalence was higher in women than in men (21.6% vs 15.2%, χ2 = 18.8, P < 0.01) and higher in subjects aged 40–59 than in subjects aged 20–39 (23.5% vs 14.8%, χ2 = 16.1, P < 0.0001).

The distribution of subjects according to frequency of symptoms is shown in Fig. 1. These subjects were further divided into those with infrequent symptoms (occurring less than twice a week) and those with frequent symptoms (occurring at least twice a week). The prevalence of frequent symptoms according to sex and age groups is shown in Fig. 2. It was reported by 5.8% of all responders and was more common in subjects aged 40–59 than in subjects aged 20–39 (9.8% vs 2.5%, χ2 = 32.2, P < 0.0001), while there were no significant difference between women and men (6.3% vs 5.3%, χ2 < 1; n.s.).

Figure 1.

Restless legs symptoms according to frequency of occurrence.

Figure 2.

Prevalence rate of frequently occurring restless legs symptoms (i.e. at least twice a week), stratified for sex and age. (○), Women; (●), men.

RLS symptoms in relation to sleep disturbance and daytime distress

Subjects with infrequent and frequent RLS symptoms, respectively, were compared with subjects without RLS with regard to several sleep and distress variables. Self-reported habitual sleep latency had overall differences between groups (one-factor ANOVA; F = 10.3; P < 0.0001) and post-hoc Scheffé test indicated significantly longer sleep latency in subjects with frequent RLS compared with subjects without RLS (31.5 ± 35.3 min vs 19.2 ± 21.4 min; F = 10.2; P < 0.0001), but no significant difference between subjects with infrequent RLS and subjects without RLS (20.7 ± 20.6 min vs 19.2 ± 21.4 min; F = 0.3; n.s.) With regard to sleep duration, overall differences were found between groups (F = 13.2; P < 0.0001) and also significantly shorter sleep duration in subjects with frequent RLS compared with subjects without RLS (6.11 ± 1.48 h vs 6.80 ± 1.16 h; F = 12.2; P < 0.0001), but no significant differences between subjects with infrequent RLS and subjects without RLS (6.62 ± 1.04 h vs 6.80 ± 1.16 h; F = 1.7; n.s.).

As can be seen in Table 1, logistic regression of screening positive for frequent RLS indicated a number of associations with insomnia symptoms and daytime distress, while none of these variables was associated with screening positive for infrequent RLS.

Table 1.  Logistic regression of screening positive for RLS
 Infrequent RLS (n = 173)Frequent RLS (n = 78)
  • Affirmative answer to the question ‘Do you have sleeping difficulties?’

  • Score of at least 11 on the subscale of the Hospital Anxiety and Depression Scale.

  • CI, confidence interval; frequent RLS, at least twice a week; infrequent RLS, less than twice a week; OR, odds ratio (age adjusted); RLS, restless legs syndrome.

Insomnia claim1.120.77–1.63n.s.3.502.17–5.64<0.0001
Major problems with:
– difficulties falling asleep1.400.85–2.28n.s.1.991.59–2.51<0.0001
– night awakenings1.170.73–1.89n.s.2.841.68–4.81<0.0001
– not being rested by sleep1.140.78–1.68n.s.2.131.28–3.55<0.01


RLS was found to have an overall estimated prevalence of 18.8%, while the prevalence of frequently occurring symptoms was 5.8%. Frequent RLS was also significantly related to various insomnia symptoms and daytime distress. The prevalence of frequent RLS of 5.8% is in accordance with other published estimates when IRLSSG criteria have been applied. Frequency of symptoms, however, is more rarely reported.

Ulfberg et al. reported a similar prevalence (5.8%) in 2608 Swedish men aged 18–64.8 In a Scandinavian study on 2005 individuals, aged 18–99 years, the overall prevalence was 11.5% but only approximately half of these subjects had moderate to severe RLS according to a severity scale.9 In a recent international study of 15 391 adults 7.2% of all responders met diagnostic criteria for symptoms of any frequency, while 4.1% met criteria for symptoms occurring at least twice every week.10 Further, Ekbom, who described and named the syndrome, reported a prevalence of 5.2% already in 1945.11

It is a limitation that the question set has been validated only for the German version, and that specificity and sensitivity of the Swedish version have not yet been established. Another limitation is the restricted age range of the study population.

In accordance with previous reports the present study found a high prevalence of RLS, and a clear association between RLS, sleep complaints and daytime distress. This association, however, was restricted to those with frequent symptoms and it is possible that these individuals are more likely to warrant medical attention. We suggest that the fourth optional question to the proposed three-question set should be regarded as mandatory to delineate a clinically significant condition in future studies.


This research was supported by the Swedish Research Council and the Nicke and Märta Nasvell Foundation.