Restless legs syndrome and periodic limb movements in 86 patients with multiple sclerosis

Abstract

Study Objectives

To assess the frequency of restless legs syndrome (RLS), periodic limb movements during sleep (PLMS), and their overlap in a large sample of patients with multiple sclerosis (MS). To compare clinical and paraclinical findings among four subgroups of patients: RLS−/PLMS− (patients without RLS and PLMS), RLS+/PLMS− (patients with RLS and without PLMS), RLS−/PLMS (patients without RLS and with PLMS), and RLS+/PLMS+ (patients with both RLS and PLMS).

Methods

In this cross-sectional, observational, instrumental study, 86 patients (M/F: 27/59; mean age 48.0 ± 10.8 years) with a diagnosis of MS underwent a telephone interview assessing the five standard diagnostic criteria for RLS. Seventy-six participants underwent polysomnography (PSG) and maintenance of wakefulness test (MWT). Instrumental and clinical findings were subsequently statistically compared to investigate their association with RLS and PLMS index (PLMSI).

Results

RLS and PLMS (PLMSI ≥15/h) frequency in patients with MS were of 31.4% and 31.6%, respectively. Among patients with RLS, 37.5% had a PLMSI of at least 15/h. RLS−/PLMS+ group showed higher wake after sleep onset (p = 0.01), stage shifts per hour (p = 0.03), increased stage N1 (p = 0.03), and reduction in stage N3 (p = 0.01) compared to RLS−/PLMS−. RLS had no influence on clinical and PSG parameters (p = 0.45).

Conclusions

RLS is highly frequent in patients with MS. The frequency of PLMS is comparable to the general population. The low percentage of patients with RLS having a high PLMSI, together with the absence of correlation between RLS and female gender and older age, supports the existence of a distinct symptomatic form of RLS in MS.

restless legs syndrome, periodic limb movements, multiple sclerosis, polysomnography

Statement of Significance

This study investigated restless legs syndrome (RLS) and periodic limb movements during sleep (PLMS) in multiple sclerosis, through the largest instrumental, nocturnal, and diurnal (MWT) polysomnographic study. RLS is very common and severe in patients with MS, while the frequency of PLMS was comparable to the general population. The results support the existence of a causal relationship between MS and the sensory but not the motor component of RLS.

Introduction

Restless legs syndrome (RLS) is a common sleep-related movement disorder, with a prevalence in the general population ranging between 3% and 10% and a female preponderance [1, 2]. RLS is characterized by an uncomfortable urge to move the legs emerging or worsening at rest and improving or disappearing with movement [3]. Symptoms worsen or appear at night or in the evening. Chronic RLS affects sleep [4], mood [5], cognition [6], and quality of life (QoL) [7]. Approximately 90% of idiopathic RLS is associated with periodic limb movements during sleep (PLMS), which are repetitive dorsiflexion of feet and big toes, running in series of at least four and occurring periodically approximately every 20 s [3]. PLMS are sensitive but not specific for RLS, being present also in other medical conditions and in elderly, with male predominance. PLMS might generate sleep and autonomic instability, favoring insomnia and hypertension [8].

Several medical conditions, including some neurological diseases, represent a risk factor for the so-called “symptomatic” forms of RLS. Solid and numerous epidemiological evidence documented a high frequency of RLS in multiple sclerosis (MS) ranging between 13.3% and 41% [9]. Among patients with MS, RLS is associated with older age, longer disease duration, more severe disability measured by Expanded Disability Status Scale (EDSS) [10], lower sleep quality, depression, fatigue, and worse QoL levels [11–15]. Most of the available studies assessed RLS and sleep complaints in MS by using questionnaires or interviews, focusing on the subjective report of patients and leaving out the objective component. Despite demanding and not costless, there is a compelling need for polysomnographic objective data on large MS cohorts in order to implement the current knowledge on the relationship between MS and sleep disorders. In particular, similarly to other conditions like sleep apnea or rapid eye movement (REM) sleep behavior disorders, the overlap between RLS and PLM might differ in MS in comparison with idiopathic RLS. Consequently, the diagnostic and prognostic value of PLM in the context of RLS or other sleep disorders in MS needs to be better evaluated. This also in consideration of the fact that RLS might be difficult to differentiate from other sensory mimics in MS.

The purpose of this study was to assess the frequency of RLS and PLMS by using polysomnography (PSG) in a large sample of patients with MS and to estimate the overlap between the two conditions. Secondary aims were to evaluate the impact of RLS/PLM on daytime vigilance and to identify possible associated factors between RLS/PLMS and sleep quality, mood, QoL, and objective daytime sleepiness in patients with MS.

Methods

Participants

A cross-sectional, observational, instrumental, single-center study in a sample of patients older than 18 and affected with MS according to McDonald criteria [16] or clinically isolated syndromes [17] was carried out at the Neurocenter of Southern Switzerland. Additional inclusion criteria were an EDSS score of less than 7.0 [18] (range 0–10) and a brain magnetic resonance imaging performed within the 12 months preceding the evaluation. Exclusion criteria were Mini-Mental Status Examination (MMSE) score lower than 24; recent (within the last 3 months) clinical MS relapse; radiologically isolated syndrome; history of drug and/or alcohol abuse; any serious general medical condition such as decompensated cardiopulmonary disease, cancer, or decompensated renal failure, as well as any major neurological condition other than MS that could interfere with the correct execution of the study design. Based on its clinical course, MS was classified as primary progressive, secondary progressive, or relapsing remitting.

Standard protocol approvals, registrations, and patient consents

According to the regulatory requirements of Switzerland, the protocol of the study was approved by the local Independent Ethics Committee. Eligible patients signed a written informed consent to participate in the study.

Study design

At the screening visit, patients were interviewed concerning their medical history; received a complete clinical and neurological examination, including the EDSS evaluation, and an MMSE assessment. All patients filled in the following self-administered questionnaires: Epworth Sleepiness Scale (ESS; range: 0–24; cutoff for normality: ≤10), Beck Disease Inventory (BDI-II; range: 0–21; cutoff for normality: <7), and Multiple Sclerosis Quality of Life (MSQoL; range: 0–84; cutoff for normality: <38). Within 1 week from the screening visit, the patients underwent a full-night PSG, followed by a maintenance of wakefulness test (MWT) in the following day. Additionally, all patients underwent a structured telephone interview evaluating the presence of RLS, conducted by one PSG-blinded neurologist expert in sleep medicine. A patient was considered to be affected by RLS if she/he met the five standard diagnostic criteria [19]. On this basis, patients were classified into four groups: (1) patients without RLS symptoms and without PLMS index (PLMSI) greater than 15/h (RLS−/PLMS−), (2) patients with RLS without PLMSI greater than 15/h (RLS+/PLMS−), (3) patients without RLS and with PLMSI at least 15/h (RLS−/PLMS+), and (4) patients with RLS and PLMSI at least 15/h (RLS+/PLMS+). All patients affected by RLS underwent the validated self-administered International RLS Rating Scale [20]. The severity of RLS symptoms was scored as mild (total score of 1–10), moderate (11–20), severe (21–30), and very severe (31–40).

PSG and MWT

All patients underwent a full-night home PSG by using a portable device (Embletta ST + Proxy). Participants were not allowed caffeinated beverages in the afternoon preceding the recordings and were allowed to sleep in until their spontaneous awakening in the morning and light-out time was based on individual habitual bedtime. The following parameters were included in the PSG montage: electroencephalogram (EEG) (at least six channels, including F3 or F4, C3 or C4, and O1 or O2, referred to the contralateral mastoid), electrooculogram (electrodes placed 1 cm above the right outer cantus and 1 cm below the left outer cantus and referred to the left mastoid), electromyogram (EMG) of the submentalis muscle, EMG of the right and left tibialis anterior muscles (bipolar derivations with two electrodes placed 3 cm apart on the belly of the TA muscle of each leg, impedance was kept less than 10 KΩ), electrocardiogram (one derivation), oral and nasal airflow by means of nasal pressure cannula, thoracic and abdominal respiratory effort strain gauge, and oxygen saturation (pulse-oximetry). Sleep signals were sampled at 200 Hz and stored on a hard disk in European data format for further analysis.

For MWT that objectively explores the ability to remain awake in sleep-promoting environmental conditions [21], recordings included only EEG, EMG, and EOG from the previous PSG and were carried out in a dim room, with low-level illumination. Patients are instructed to stay awake as long as possible. Four sessions were recorded every 2 h, each lasting until the patient fell asleep or up to 40 min of continuous awake state. Sleep latency was assessed in each of the four recordings [22].

Sleep stages were visually scored following standard criteria on 30-s epochs using the sleep analysis software Hypnolab 1.2 (SWS Soft, Italy). Limb movements during sleep were first detected automatically by the software and then manually checked and corrected following the World Association of Sleep Medicine 2016 scoring criteria [23]. In particular, the PLMSI was calculated as the number of leg movements (LMs) included in a series of four or more, separated by 10–90 s (without any LM preceded by an interval <10 s interrupting the PLMS series), per hour of sleep. Two expert sleep scorers, blind to the patient identity, scored sleep (K.T. and M.M.) and visually edited the limbs movements (K.T. and A.C.) detected automatically, before computing the final results.

Statistical analysis

The frequency of RLS was calculated as the percentage of the total sample (n = 86), while the frequency of PLMS was calculated on patients who performed PSG (n = 76). Direct logistic regression was performed to assess the impact of antidepressant medications (selective serotonin reuptake inhibitor (SSRI) and tricyclic antidepressants) on PLMS and RLS. Only Participants taking SSRI and/or tricyclic antidepressants but not receiving antiepileptic drugs or opioid medications (known to have a favorable effect on RLS) were considered in this analysis. Two one-way, between-group multivariate analysis of variance (MANOVA) were conducted separately for PLMS and RLS categorical variables, comparing the two conditions on the combination of instrumental and clinical dependent variables. Preliminary assumption testing was conducted to check for normality, linearity, univariate and multivariate outliers, homogeneity of variance–covariance matrices, and multicollinearity, with no serious violations noted. Standard multiple regression was performed evaluating the impact of eight independent factors—EDSS, BDI total score, MWT latency, ESS, MSQoL, sleep efficiency (SE), total sleep time (TST), and respiratory distress index (RDI)—on the PLMSI and the score at the International RLS Study Group rating scale (IRLSSG) for RLS severity. Finally, a one-way, between-group analysis of variance, with post-hoc comparisons using the Tukey HSD test, was conducted to explore the impact of RLS/PLMS on PSG sleep parameters. The level for statistical significance was set at p < 0.05.

Data availability

The data that support the findings of this study are available from the corresponding author, upon reasonable request.

Results

Eighty-six patients affected by MS (59 women, 27 men, mean age 48.0 ± 10.8 years) were included in the study (Table 1). All patients were screened for RLS symptoms, and 76 had complete instrumental assessments including clinical questionnaires, PSG, and MWT. Ten patients dropped out without further measurements for personal reasons.

Table 1.

Demographic parameters of the participants enrolled in this study

Attribute	Quantity
Participants, n	86
Age, years, mean (SD)	48.0 (±10.8)
Female, n (%)	59 (68.6%)
Male, n (%)	27 (31.4%)
Disease duration, mean (SD)*	10.5 (±7.9)
EDSS, mean (SD)*	2.6 (±1.4)
CIS, n (%)*	5 (5.8)
RRMS, n (%)*	65 (79.3)
PPMS, n (%)*	6 (7.9)
DMTs* none, n (%)	10 (13.1)
Injectables, n (%)	31 (40.8)
Orals, n (%)	18 (23.7)
Antibodies, n (%)	17 (22.4)
Psychotropics intake*, n (%)	39 (51.3)

EDSS = Expanded Disability Status Scale; CIS = clinical isolated syndrome; RRMS = relapsing-remitting multiple sclerosis; PPMS = primary progressive multiple sclerosis; DMT = disease modifying therapy.

*Data referred to 76 of 86 patients.

RLS frequency

Twenty-seven patients out of 86 (31.4%) met the five standard diagnostic criteria for RLS and were classified as “MS/RLS+.” Patients with MS who never experienced RLS symptoms (57 patients, 66.3%) and patients who reported one or fewer but not all five diagnostic criteria for RLS (2 patients, 2.3%) were classified as “MS/RLS−” subgroup (59 patients, 68.6%). In the MS/RLS+ group, the mean IRLS score was 18.7, in particular, 17.5% of participants presented a mild (1–10 points) severity, 43.5% had a moderate (11–20 points) severity, a severe form was found in 35.7% (21–30 points), and a very severe form of RLS in 4.3% (31–40 points). In females, the frequency of RLS was 30.5% (18 of 59), while in males it was 33.3% (9 of 27). No correlation between age and presence of RLS (odds ratio [OR] = 1.017; confidence interval [CI]: 0.96–1.07, p = 0.53) was found with a logistic regression analysis.

Frequency of PLMS

PLMS were screened in 76 patients for whom the PSG study was available. Twenty-four patients with MS (31.6%) had PLMSI of at least 15/h (Figure 1) and formed the “PLMS+” subgroup, while those with PLMSI less than 15 were classified as “PLMS−” subgroup. In females, PLMS frequency was 20.8% (11 of 53), while in male patients, it was 56.5% (13 of 23). Fifty-two patients (68.4%) formed the MS/PLMS− subgroup. Frequency of PLMS depended on the PLMSI cutoff used, and increased to 38.2% with the threshold set at at least 10/h, up to 48.7% for PLMSI of at least 5/h and decreased to 7.9% with a threshold set at PLMSI of at least 30/h. In the overall group, the mean PLMSI was 18.8/h. The mean PLMSI was 52.0/h in patients with PLMSI of at least 15/h, and it was 37.2/h, 45.1/h, and 82.1/h in patients with PLMSI of at least 5, 10, and 30, respectively. There was no evidence of a linear correlation between patients’ age and PLMSI at multiple regression analysis (p = 0.69; β = 0.52; 95% CI).

Figure 1.

Frequency of RLS and PLMS in the 76 patients with MS for whom PSG was available.

Overlap between RLS and PLMS

As illustrated in Figure 2, the frequency of PLMS in patients with RLS decreased with the increase of PLMS threshold chosen. With a cutoff of PLMSI of 15, the frequency of PLMS in the whole group of RLS was 37.5%, while it was 28.8% in non-RLS patients. Among patients with severe/very severe RLS, 12.5% had PLMSI of at least 15/h. This percentage rises to 46.1% in patients with mild/moderate RLS. Although this latter result seems paradoxical, it may be explained by the lack of correlation between PLMSI and IRLSSG score (p = 0.81).

Figure 2.

Comparison of the frequency of PLMS in patients with MS with or without RLS.

As shown in Figure 3, among patients with PLMSI of ≥5/h, ≥10/h, ≥15/h, and ≥30/h, the percentage of participants with RLS was 45.9%, 41%, 37.5%, and 25%, respectively.

Figure 3.

Comparison of the frequency of RLS in patients with MS with or without PLMS.

Participants presenting both RLS and PLMS of ≥5/h, ≥10/h, ≥15/h, and ≥30/h represented the 22.4 %, 15.8%, 11.8%, and 3.9% of the overall sample, respectively.

Mean PLMSI in RLS patients was 13.2/h and was 23.0/h in those without RLS (p = 0.22).

RLS+ versus RLS−

MANOVA analysis showed no difference between RLS+ and RLS− groups on the combination of seven dependent variables: EDSS, BDItot, MWT, ESS, MSQoL, SE, and TST (F = 0.99, p = 0.45; Wilks’ Lambda = 0.89; partial eta squared = 0.11). When the results for the dependent variables were considered separately, no difference reached statistical significance (Table 2). Standard multiple regression was subsequently performed evaluating the impact of the eight independent factors (EDSS, BDItot, MWT, ESS, MSQoL, SE, TST, and RDI) on the severity of RLS as measured by the IRLSSG. We observed an association toward greater EDSS in patients with MS with a higher score at the IRLSSG after applying the linear regression model (p = 0.002), even when patients with an EDSS at least 7 were excluded. No associations were found between RLS severity and other independent variables.

Table 2.

Comparison between MS patients with (RLS+) and without (RLS−) RLS

	RLS+		RLS−		One-way MANOVA
	Mean	SD	Mean	SD	p<
Expanded disability status scale	2.70	1.71	2.60	1.30	0.66
Epworth sleepiness scale	7.59	4.02	9.73	4.67	0.94
Maintenance of wakefulness test	34.5	7.2	34.0	9.0	0.1
Beck disease inventory	12.25	11.42	14.83	10.87	0.47
MS quality of life physical	59.29	22.55	59.28	20.92	0.08
Sleep efficiency	78.6	11.78	80.2	10.76	0.80
Total sleep time	337.5	64.84	373.28	64.90	0.98

MANOVA test was performed comparing severe RLS (IRLSSG score >20) versus RLS− on the same dependent variables and there was no difference between groups.

PLMS+ versus PLMS−

MANOVA analysis showed no difference between the PLMS+ and PLMS− groups (cutoff value PLMSI ≥15) on the combination of seven dependent variables: EDSS, BDItot, MWT, ESS, MSQoL, SE, and TST (F = 0.32, p = 0.94; Wilks’ Lambda = 0.96; partial eta squared = 0.04). As summarized in Table 3, when the results for the dependent variables were considered separately, no difference reached statistical significance. Standard multiple regression was then performed evaluating the impact of eight independent factors (EDSS, BDItot, MWT, ESS, MSQoL, SE, TST, and RDI) on the PLMSI. We observed an association toward greater RDI in patients with MS with higher PLMSI after applying the linear regression model (p = 0.0001). No association was found between PLMSI and the other independent variables.

Table 3.

Comparison between MS patients with (PLMS+) and without (PLMS−) PLMS

	PLMS+		PLMS−		One-way MANOVA
	Mean	SD	Mean	SD	p<
Expanded disability status scale	2.66	1.50	2.65	1.33	0.71
Epworth sleepiness scale	9.74	5.59	9.00	3.87	0.78
Maintenance of wakefulness test	33.7	8.2	33.7	8.3	0.9
Beck disease inventory	9.67	8.25	17.56	1.30	0.85
MS quality of life physical	65.92	20.74	53.07	19.35	0.61
Sleep efficiency	76.61	11.16	81.42	10.28	0.24
Total sleep time	357.26	77.28	366.78	62.92	0.35

MANOVA test was performed comparing patients with PLMSI of at least 30/h versus those with PLMSI of less than 5/h on the same dependent variables and there was no statistically significant difference between groups.

Antidepressant intake was not correlated to PLMS (OR = 0.60; p = 0.41) nor RLS (OR = 0.34; p = 0.13).

RLS−/PLMS− versus RLS+/PLMS− versus RLS−/PLMS versus RLS+/PLMS+

As given in Table 4, there was no difference between the four groups of participants on the combination of seven dependent variables: EDSS, BDItot, MWT, ESS, MSQoL, SE, and TST.

Table 4.

Comparison of clinical and paraclinical parameters based on the presence of RLS/PLMS

	RLS+/PLMS+ (n = 9)		RLS+/PLMS− (n = 16)		RLS−/PLMS+ (n = 15)		RLS−/PLMS− (n = 36)		One-way ANOVA
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	p (between groups)
Expanded disability status scale	2.63	0.74	2.93	1.44	3.07	1.53	2.29	1.36	0.23
Epworth sleepiness scale	8.25	3.62	9.71	5.34	10.87	5.89	8.77	4.01	0.46
Maintenance of wakefulness test	30.69	12.06	31.59	11.84	36.98	4.08	34.27	8.30	0.32
Beck disease inventory	16.57	13.10	14.43	9.96	11.87	10.71	14.54	12.06	0.81
MS quality of life physical	51.18	25.41	53.60	19.68	61.37	22.38	64.15	20.05	0.28
Sleep efficiency	78.41	9.45	79.56	10.68	74.01	11.53	82.80	10.33	0.08
Total sleep time	342.93	63.15	362.10	55.54	351.80	69.51	374.7	66.98	0.51

Polysomnographic sleep parameters

As given in Table 5, there was a statistically significant difference in wake after sleep onset (W), stage shift per hour (SSh), and percentage of N1 and N3 between the four groups. Post-hoc comparisons using the Tukey HSD test indicated that the mean score of the above-mentioned parameters for the RLS−/PLMS+ group was significantly different from that of the RLS−/PLMS− group, with PLMS+ patients having higher W, SSh, increased N1, and reduction in N3. The same analysis was performed comparing patients with and without PLMS of at least 5 and patients having or not RLS symptoms. RLS was found to have no influence on PSG parameters, while there was higher W (p = 0.03), SSh (p = 0.01), and increased N1 (p = 0.07) in patients with PLMS compared to patients without PLMS. The analysis of nocturnal respiratory events revealed a mean RDI of 5.82/h ± 8.08, with 35.4% of the patients having an RDI of at least 5.

Table 5.

Comparison of PSG sleep architecture based on the presence of RLS/PLMS

	RLS+/PLMS+ (n = 9)		RLS+/PLMS− (n = 16)		RLS−/PLMS+ (n = 15)		RLS−/PLMS− (n = 36)		One-way ANOVA
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	p (between groups)
Sleep latency, min	26.15	17.26	27.21	27.63	37.53	32.42	29.91	35.22	0.78
Stage W, min	12.47	7.66	12.46	8.8	18.22*	11.79	9.56*	6.77	0.01†
Awakenings/hour	6.53	2.21	5.93	2.59	6.82	3.43	5.95	2.96	0.75
Total sleep time, min	349.50	81.38	326.06	55.88	367.69	77.74	375.71	60.88	0.24
Sleep period time, min	404.95	58.22	418.57	51.89	431.07	63.23	410.49	70.18	0.71
REM latency, min	129.30	91.02	106.21	75.34	155.13	111.91	95.23	66.13	0.11
Stage shifts/hour	23.18	4.90	18.15	5.64	23.51	10.10	19.03	5.03	0.03†
Stage N1, %	12.01	3.16	10.51	4.74	14.78*	11.04	9.21*	3.81	0.03†
Stage N2, %	37.71	11.05	41.46	11.03	41.78	11.54	42.78	7.75	0.54
Stage N3, %	21.05*	5.12	18.46	11.19	12.09*	7.60	20.13*	6.30	0.01†
Stage R, %	16.75	8.36	17.10	5.95	13.13	7.25	18.30	6.27	0.11
Sleep efficiency, %	76.33	9.66	79.20	13.74	77.13	11.90	82.02	9.48	0.40
PLMS index, n/hour	27.60	12.69	5.53	4.13	65.76	58.90	2.41	3.31
Respiratory disturbance index	3.78	4.16	3.99	3.85	8.85	13.12	4.97	5.59	0.43

*Difference in mean scores statistically significant between two groups after post-hoc analysis (Tukey HSD test).

†Difference in mean scores statistically significant (p < 0.05) for the four groups.

Maintenance of wakefulness test

Seventy-six out of 86 participants performed MWT. Mean sleep latency in the overall sample was 34 min; 11.3% and 25.4% showed pathological results on MWT considering a threshold of 20 and 30 min, respectively.

There was no difference (p = 0.18) in mean sleep latency between RLS+ (31.3 min, SD: 7.2 min) and RLS− groups (35.1 min, SD: 9.03 min). Among patients with and without RLS, a pathological mean sleep latency was found in the 22.7% and 6.0% (threshold of 20 min; p = 0.04) and 31.8% and 20% (threshold of 30 min; p = 0.22), respectively.

No difference in mean sleep latency was found between patients with (33.73 min, SD: 8.16 min) or without PLMSI of at least 15 (33.73 min, SD: 8.26). Among patients with and without PLMSI of at least 15, pathological mean sleep latency was found in the 7.4% and 12.2% (threshold of 20 min) and 18.5% and 26.5% (threshold of 30 min), respectively.

Discussion

This is the largest polysomnographic study conducted in MS so far. We focused on RLS and PLMS, finding a frequency of 31.4% and 31.6%, respectively. Among patients with RLS, 40% presented a severe form of the disease. The high frequency of RLS and its severity, compared to the general population (3%–10%), strongly suggest the need to always investigate this pathological condition in patients with MS. Although a prevalence of RLS in MS varying from 13% to 41% [9, 10, 24, 25] has been reported in the literature, only two case–control studies found higher PLMSI in patients with MS than in controls [26, 27]. Other four case–control studies found no difference in PLMSI between patients and controls [28–31]. None of these studies involved more than 65 patients.

Accordingly, with the latest epidemiologic data, RLS prevalence in the American and European populations ranges from 3% to 10% [1, 2], with female predominance, while about 28.6% of the general population presents a PLMSI of more than 15/h with frequency increasing with age and in association with male gender [32]. In our cross-sectional study, RLS and PLMS (defined by PLMSI ≥15) were approximately 4- and 1.14-fold more frequent in patients with MS than in the general population, respectively.

As reported in the updated IRLSSG consensus criteria [19], approximately 80%–84% of RLS patients present periodic leg movements (PLMS ≥5/h), while other studies even show a frequency of 90% [3]. The presence of PLMS is a supportive criterion for the diagnosis of RLS [19] and might contribute to sleep and autonomic instability documented in RLS patients. This study highlighted a lower frequency of PLMS in patients with MS-related RLS: only 37.5% presented a PLMSI of at least 15/h and 70.8% presented at least 5/h.

The latest findings in the literature show a prevalence of idiopathic RLS in patients with PLMSI of at least 15/h of about 25% [33]. Differently, in patients with PLMSI of at least 15/h, we found RLS frequency of 37.5%, indicating a slightly higher predictive value of PLMS for RLS in MS than in the general population.

RLS, but not PLM, was associated with daytime sleepiness when investigated objectively by means of MWT. The interpretation of this data is hampered by the fact that no previous study investigated idiopathic RLS using MWT. This finding might seem at odds with the observation that self-reported daytime sleepiness, as measured by the objective ESS questionnaire, was similar between RLS and RLS-free patients. This latter data are widely acknowledged, as also mentioned by the supportive diagnostic criteria for RLS, and it is in line with the previous literature as well as with the observation that RLS patients do not usually nap during the day, even in face of a significant sleep loss [19]. The apparent discrepancy between MWT and ESS findings in MS might be related to the fact that MWT measures the ability of the patient to stay awake, while ESS measures the tendency to fall asleep in representative daily activities. Alertness and sleep propensity may behave somewhat independently, as MWT correlates poorly or moderately with ESS, also in other disorders such as obstructive sleep apnea syndrome and narcolepsy [34].

RLS can be classified as primary/idiopathic and secondary/symptomatic forms, such as those associated with iron deficiency, renal failure, rheumatoid arthritis, drug-related, and pregnancy [35–37]. Among all neurological disorders, peripheral neuropathies [38], spinocerebellar ataxia [39], Parkinson’s disease [40], myelopathies [41], and MS have been reported to have more frequently RLS than the general population. The high-frequency rates of RLS in our sample of MS support the concept that MS is a risk factor for RLS. Furthermore, our data suggest that patients with MS present a specific RLS phenotype characterized by (1) a higher frequency of severe and very severe RLS, (2) a lower overlap between RLS and PLM, (3) a significant correlation between RLS frequency and global EDSS score, (4) the absence of correlation between RLS and female gender as well as older age, and between age and PLMSI in MS patients, and (5) an atypical association between RLS and daytime sleepiness (MWT). It might be speculated that the neurological involvement of MS lesions at the level of the cervical cord might account for the increased prevalence of RLS in MS and that the preferential involvement of sensitive fibers explains the specific phenotype we observed. Previously, the REMS study (2008) provided data both in favor and against the hypothesis of a secondary MS-related RLS. In line with our findings, the authors showed an association between RLS, longer MS duration, and more severe disability, besides the higher severity of RLS in MS. On the other hand, other results from the previous study supported the hypothesis that the nature of RLS may be idiopathic. In fact, a positive correlation between RLS and age and female sex was found, and approximately 14% of MS/RLS+ patients reported the occurrence of RLS symptoms in at least one of their first-degree relatives [10]. Future studies should further address the specific RLS phenotype in MS and the casual relationship between MS and MS-related RLS.

We found no correlation between RLS, PLMS, and measures of mood, QoL, and neurological disability. These findings are at odds with the results reported by other three studies, where self-reported sleep quality, depression, clinical disability (EDSS), and QoL were found to be worse in patients with RLS compared to their RLS-free counterparts [11, 42, 43]. Divergent results might be due to differences in study populations, for example, all three studies enrolled patients younger and with a more severe MS compared to our patients.

Polysomnographic data in this study revealed that PLMS themselves resulted in higher W, with an increased percentage of light sleep (N1) and slow-wave sleep loss. The analysis of nocturnal respiratory events showed high mean RDI, with 35% of patients having more than five respiratory events per hour, mostly obstructive. In the literature on MS, apnea–hypopnea index values were not exceedingly high and varied considerably between studies, ranging from normal [26, 29, 44] to moderately increased [15]. Future studies should investigate further the prevalence and the features of sleep breathing disorders in patients with MS.

This study has some limitations. First of all, the absence of an age-matched control group, although the percentages we found are far above well-known values for the general population. Spine and brain magnetic resonance images were not evaluated. It will be important to include this information in future investigations as one study found that the presence of spinal cord lesions increased the risk of RLS in patients with MS [45]. The exclusion of participants with an EDSS of at least 7 and the relatively low median EDSS (2.4) in this MS population limits the generalizability of our results. The lack of a structured assessment of MS-related sensory–motor RLS “mimics,” such as neuropathic pain and spasticity, is a crucial point that will require further more systematic and possibly interventional study designs. Indeed, patients with MS are known to have a high prevalence of sensory symptomatology that can mimic RLS (RLS “mimics”). The low frequency of PLMS in our patients with RLS may suggest that some of them have been misdiagnosed. However, an expert in sleep medicine carefully investigated RLS through a structural interview in our sample. All RLS criteria must be fulfilled to be in the RLS group, while doubtful cases were listed among the non-RLS participants. This raises the concern that the current five RLS diagnostic criteria might not be appropriate for MS and that other criteria, such as a positive trial response to dopamine agonists, should be considered in these patients. Finally, the lack of data on ferritin levels limits the possibility to speculate on the possible causal links between RLS and MS. In the REMS study (2008), iron-storage indicators, as well as plasma levels of creatinine and folate, did not differ between MS patients with and without RLS [10]. Although being a complex topic, iron involvement in MS-related RLS pathogenesis is indeed a source of interest and should be investigated by further studies.

In conclusion, this study confirms that RLS is highly prevalent and severe in patients with MS, supporting the importance of accurately investigating this pathological condition in MS patients. RLS in MS was also associated with an impairment of vigilance only when objectively investigated by MWT, suggesting that subjective ESS test might be not enough sensitive to identify this clinical feature. The frequency of PLMS in patients with MS is comparable to the general population. These findings and the low proportion of patients with RLS and PLMS support the hypothesis that RLS in MS represents a specific entity, i.e. an RLS variant secondary to MS mediated central nervous system damage. However, RLS in MS should be carefully differentiated from sensory mimics, maybe by implementing the current RLS diagnostic criteria with standardized pharmacological tests. According to our results, the value of PSG in RLS diagnosis remains instead rather limited to MS cases with refractory RLS, or if there is suspicion of sleep breathing or other sleep disorders.

Acknowledgments

The authors want to thank all the patients and participants involved in the study. Special thanks are due to Ente Ospedaliero Cantonale (EOC), for the financial support (Grant ABREOC).

Funding

Grant ABREOC from Ente Ospedaliero Cantonale (EOC) and Swiss MS Society (SMSS).

Disclosure Statement

Financial disclosure: The authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest in the subject matter or materials discussed in this manuscript.

Nonfinancial disclosure: The authors report no conflict of nonfinancial interest.

Conflict of interest statement. None declared.

Author Contributions

M.M. and C.G. contributed to study concept and design and revised the manuscript for intellectual content. D.S. drafted the text for intellectual content and contributed to data acquisition and analysis. R.F., A.C., S.M., K.T., N.T., C.C., G.R., G.D., and C.Z. contributed to data acquisition and analysis. M.M., D.S., and R.F. contributed to figure and table preparation and interpretation of data.

Author notes

These authors contributed equally to this work.