Prevalence of ‘poor sleep’ among patients with multiple sclerosis: An independent predictor of mental and physical status☆

1. Introduction 

Multiple sclerosis (MS) is a chronic demyelinating disease of the central nervous system with onset generally between the ages of 20 and 50years. Patients with MS report sleep disturbances more frequently than the general population, and may be affected by the entire spectrum of sleep disorders (i.e., insomnia, excessive daytime sleepiness, periodic leg movements, restless legs syndrome, abnormal sleep–wake regulation, sleep-disordered breathing, narcolepsy and rapid eye movement sleep behaviour disorder, etc.) [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12]. Two polysomnographic studies performed on patients with MS showed that sleep macro- and microstructure were markedly altered in this patient group, with a higher number of nocturnal awakenings, a higher level of wake time after sleep onset and a higher total arousal index, and a lower sleep efficiency index compared with controls [7], [8]. Even though disrupted nocturnal rest might be directly related to specific sleep disturbances, many other demographic and clinical conditions (e.g. presence of comorbidities, MS duration and progression, medication effects, experience of pain and sexual and/or bladder dysfunction, anxiety and depression) should be considered as possible causes of sleep disorders in patients with MS. To our knowledge, the role of demographic, socioeconomic and clinical variables on sleep quality in patients with MS has never been carefully evaluated.

Several studies, performed in Italian patients affected by MS, have shown that this neurological disease has a highly negative impact on quality of life (QoL) [13], [14]. Many factors seem to be implicated in the impairment of QoL in patients with MS: a progressive disease course [13], [15], [16], [17], extended disease duration [18], physical disabilities [16], [19], [20], emotional and cognitive impairment [13], [17], [20], sexual and/or bladder dysfunction [21] and interferon (IFN) beta treatment [22] represent the main predictors of poor QoL in this population. Unfortunately, only one of the studies conducted to date has included sleep quality among the variables potentially related to QoL [23], but in this study the only independent predictors of the global quality of life index score were found to be depressive mood and physical disability; quality of sleep was an independent predictor only of physical well-being.

The aims of our study were: (i) to examine the prevalence of “poor sleepers” and (ii) to evaluate the role of demographic, socioeconomic and clinical variables on sleep quality among patients affected by MS, and (iii) to analyse the association between quality of sleep and QoL while controlling for known and unknown possible predictors of QoL in patients with MS.

2. Methods 

3. Results 

3.1. General characteristics 

Demographic, socioeconomic and clinical characteristics of the 120 patients with MS included in our study are reported in Table 1. Most of our patients were affected by relapsing-remitting MS, whereas no patients presented primary progressive MS and 23 patients (19.2%) had secondary progressive MS.

Table 1. Demographic, socioeconomic and clinical characteristics of the study population, “good sleepers” and “ poor sleepers”

		All subjects (n=120)	Good sleepers (n=63)	Poor sleepers (n=57)	p
	Age (years)	44.2±11.1	42.5±11.6	46.1±10.2	0.08
	Male (%)	27.5	28.6	26.3	0.7
	Employeda (%)	76.7	77.8	75.4	0.7
	Living alone (%)	32.5	39.7	24.6	0.08
	High or very high educationb (%)	57.5	63.5	50.9	0.1
	Relapsing-remitting MS (%)	80.8	81	80.4	0.9
	Duration of MS (months)	109.4±94.8	98.3±85.7	121.6±103.2	0.3
	EDSS score	2.7±1.4	2.3±1.4	3.1±1.4	0.009
	Interferon-beta or glatiramer acetate treatment (%)	60	58.7	61.4	0.7
	Frequency per week of interferon-beta or glatiramer acetate treatment	3.25±2.0	3.1±1.8	3.4±2.2	0.4
	Immunosuppressive treatment (%)	12.5	15.9	8.8	0.2
	Pain due to MS (%)	25.0	17.7	33.3	0.05
	Sexual and/or bladder dysfunction due to MS (%)	5.0	1.6	8.8	0.1
	Presence of lesions in the brain only (%)	27.5	28.6	26.3	0.7
	Presence of combined lesions (%)	60.8	55.6	66.7	0.2

n, Number of patients; MS, multiple sclerosis; EDSS, Expanded Disability Status Scale.

a Including home workers. b ⩾13years of school attended.

With regard to treatment, a large part of MS patients were treated using IFN beta and/or glatiramer acetate (72 patients) and immunosuppressive drugs (15 patients), whereas no drug specific for MS was taken by the remaining 33 subjects (27.5%). The mean number of drugs other than IFN beta, glatiramer acetate and immunosuppressives was 1.4±0.9. In our sample, 33 patients had demyelinating plaques only in the brain and 14 only in the brainstem and/or spinal cord, whereas 73 showed combined lesions. We observed in MS patients a mean CCI score of 0.11±0.3 (range 0–2). Based on the neuropsychological evaluation, 12 (10%) and 57 (47.5%) patients were defined as affected by anxiety and depression, respectively. MS patients diagnosed as “anxious” showed at the HARS a mean score of 21.1±2.8, while the mean score at the HDRS in the 57 depressed MS patients was of 16.2±7.4.

The 120 MS patients had a mean global PSQI score of 6.83±4.53, ranging from 1 to 18. The mean score for each PSQI component was of: 1.38±0.92 for sleep quality, 1.48±1.20 for sleep onset latency, 1.03±0.90 for sleep duration, 0.82±1.00 for sleep duration, 1.15±0.54 for sleep disturbance, 0.40±0.87 for use of hypnotic drugs and 0.58±0.95 for daytime dysfunction. The mean scores for each SF-36 domain and for the MCS and PCS among the 120 patients are shown in Fig. 1.

View Large Image Download to PowerPoint

Fig. 1. Mean scores for each Short Form-36 domain and for the mental component summary (MCS) and physical component summary (PCS) among all subjects, “good sleepers” and “poor sleepers”. PF, physical functioning; RP, role limitations due to physical health problems; BP, bodily pain; GH, general health; VT, vitality; SF, social functioning; RE, role limitations due to emotional health problems; MH, mental health.

3.2. Quality of sleep 

Fifty-seven MS patients (47.5%), reporting a global PSQI score>5, were defined as poor sleepers.

Demographic and socioeconomic characteristics did not differ between good sleepers and poor sleepers. In poor sleepers, EDSS score and the prevalence of pain due to MS were significantly higher than in good sleepers, whereas there were no significant differences between the two groups with regard to the other variables related to MS, including therapeutic information (Table 1, Table 2). CCI score was significantly higher in poor sleepers than in MS patients with a global PSQI⩽5 (0.19±0.4 vs. 0.03±0.2, p=0.009). The mean number of drugs other than IFN beta, glatiramer acetate and immunosuppressives taken by patients did not differ between the two groups (good sleepers, 1.44±0.8 vs. poor sleepers, 1.41±0.9, p=0.8). A diagnosis of anxiety or depression was more common among poor sleepers than among good sleepers (17.5% vs. 3.2%, p=0.009 and 77.2% vs. 20.6%, p<0.001, respectively). In the MS patients diagnosed as anxious we did not observe a significant correlation between the global PSQI and HARS scores (r=−0.35, p=0.2); in contrast, the global PSQI score showed a significant direct correlation with the HDRS score (r=0.64, p<0.001).

Fig. 1 shows the mean score for each SF-36 domain and for the MCS and PCS among good and poor sleepers. Comparisons between patients with MS with good and poor sleep quality were significant for all variables (p<0.001).

Among continuous variables, the global PSQI score showed a significant direct correlation with age (r=0.30, p=0.001), CCI (r=0.20, p=0.03) and EDSS global score (r=0.38, p<0.001), whereas there was a significant inverse correlation with each SF-36 domain (PF, r=−0.52, p<0.001; RP, r=−0.40, p<0.001; BP, r=−0.35, p<0.001; GH, r=−0.63, p<0.001; VT, r=−0.56, p<0.001; SF, r=−0.54, p<0.001; RE, r=−0.54, p<0.001; MH, r=−0.52, p<0.001) and with the MCS and PCS (see Fig. 2). No significant correlations were observed with the duration of MS, number of drugs other than IFN beta, glatiramer acetate and immunosuppressives, and frequency per week of IFN beta or glatiramer acetate treatment.

View Large Image Download to PowerPoint

Fig. 2. Correlations between the global Pittsburgh Sleep Quality Index (PSQI) score and the Short Form-36 mental component summary (MCS) (A) and the Short Form-36 physical component summary (PCS) (B).

3.3. QoL (MCS and PCS) 

Among demographic and socioeconomic variables, there was a significant inverse correlation between age and mean MCS and PCS scores (MCS, r=−0.21, p=0.02; PCS, r=−0.40, p<0.001). Moreover, the two SF-36 summary scores were significantly higher in patients attending school for ⩾13years (MCS, 44.10±9.65 vs. 39.92±9.87, p=0.02; PCS, 44.70±10.92 vs. 38.58±11.13, p=0.003). Secondary progressive MS impaired physical (PCS, 33.23±9.19 vs. 44.40±10.73, p<0.001) but not mental status. No significant correlations were observed between MS duration and MCS and PCS scores, whereas MCS and PCS scores were inversely correlated with the mean EDSS score (MCS, r=−0.32, p<0.001; PCS, r=−0.67, p<0.001). The type and frequency of treatment for MS had no effect on either MCS or PCS scores, and pain due to MS impaired physical status only (37.60±11.88 vs. 43.79±10.77, p=0.01). Patients reporting sexual and/or bladder dysfunction due to MS had significantly lower scores both for MCS and for PCS than those without urogenital dysfunction (MCS, 30.97±10.57 vs. 42.92±9.57, p=0.03; PCS, 31.07±10.00 vs. 41.68±11.19, p=0.03). Mean MCS score was significantly lower in patients with demyelinating plaques in both the brain and brainstem and/or spinal cord (combined lesions) than in MS patients with isolated lesions in the brain or in the brainstem and/or spinal cord (40.80±8.93 vs. 44.70±10.98, p=0.04). There was no correlation between measure of comorbidity (CCI) and MCS or PCS scores. There was an inverse correlation between PCS and the number of drugs other than those specific for MS (r=−0.31, p=0.01). Patients affected by anxiety and depression showed significantly lower scores both for the MCS and for the PCS than those without a clinical diagnosis of psychological disturbances (for anxiety: MCS, 32.06±9.57 vs. 43.47±9.32, p<0.001; PCS, 29.52±8.03 vs. 43.50±10.85, p<0.001; for depression: MCS, 37.50±9.46 vs. 46.70±8.22, p<0.001; PCS, 35.95±9.53 vs. 47.66±10.02, p<0.001).

The presence of sexual and/or bladder dysfunction due to MS and global PSQI score were the only independent predictors for MCS, whereas age, EDSS score and global PSQI score were the only independent predictors for PCS (Table 2).

Table 2. Quality of life predictors obtained by means of multiple linear regression models with the Short Form-36 mental component summary (MCS) and physical component summary (PCS) as outcome variables

	Variable	Standardized β value	p
	MCS predictors (R2=0.35; F=31; p<0.001)
	Age	−0.02	0.7
	Level of educationa	0.08	0.2
	EDSS score	−0.10	0.2
	Sexual and/or bladder dysfunction due to MS	−0.17	0.03
	Presence of combined lesions	−0.12	0.09
	Global PSQI score	−0.54	<0.001
	Anxiety	−0.12	0.2
	Depression	−0.12	0.3
	PCS predictors (R2=0.46; F=15; p<0.001)
	Age	−0.21	0.04
	Level of educationa	−0.05	0.6
	Secondary progressive MS	−0.12	0.3
	EDSS score	−0.48	<0.001
	Pain due to MS	0.04	0.7
	Sexual and/or bladder dysfunction due to MS	0.15	0.2
	Number of drugs other than interferon-beta, glatiramer acetate and immunosuppressives	−0.16	0.1
	Global PSQI score	−0.30	0.005
	Anxiety	−0.08	0.4
	Depression	−0.13	0.3

MS, multiple sclerosis; EDSS, Expanded Disability Status Scale; PSQI, Pittsburgh Sleep Quality Index.

a ⩾13years of school attended.

4. Discussion 

Sleep disturbances are common in MS; in fact up to 54% of patients report sleep-related problems [2]. These data were confirmed by the results of our study, in which almost half of MS patients were affected by non-restorative sleep. In addition, our study showed for the first time that, in patients with MS, poor sleep is not only associated with specific sleep disorders, but also with several clinical conditions, both related and unrelated to MS. Furthermore, we observed that poor sleep has a direct impact on both mental and physical status in this patient group.

Considering each component of the PSQI, the most important complaint among MS patients appeared to be a difficulty in falling asleep. Recently, Stanton et al. reported that initial insomnia, on at least two nights per week, occurred in 42% of 60 outpatients with MS [1]. Although several conditions could explain this disturbance, an important role might be played by restless legs syndrome. In fact, the unpleasant sensations associated with the urge to move reported by individuals with this sleep-related movement disorder represent a main cause of initial insomnia [33], and the prevalence of this condition has been found to be more frequent in MS patients than in the general population [5], [6]. However, the effect of other variables in determining why patients with MS have difficulty falling asleep, such as pain and anxiety (both found to be associated with poor sleep in our study) should be carefully considered by clinicians [1]. The use of hypnotic drugs and daytime dysfunction were two components of the PSQI that appeared less prominent contributors to this phenomenon in our patient population.

As suggested by Fleming and Pollak, many factors could influence the quality of sleep in this population [34]. Although several studies have focused on the role of specific sleep disturbances in MS patients, to our knowledge the possible effects of general (demographic and socioeconomic) and clinical characteristics on quality of sleep in individuals affected by this chronic demyelinating disease have never been specifically addressed. Our study shows that, unlike demographic and socioeconomic variables, clinical conditions, both related and unrelated to MS, are associated with an altered quality of sleep in individuals affected by MS. A significantly higher mean EDSS score was found among poor sleepers than among good sleepers and was correlated directly with the global PSQI score. These findings support those previously reported by Lobentanz et al. [23], however, other authors have found the presence of sleep disturbances in patients with MS to be independent of the EDSS score [1], [2], [4]. Since EDSS is a measure of impairment, our results suggest that in MS patients disability may have a primary role in affecting sleep. Our suggestion appears to be likely if we consider that the harmful effect of disability on sleep has already been reported in patients with different chronic disabling disorders. Recently, a large study performed on primary care patients showed that a “not good” physical health condition lasting more than six days in the previous month was associated with all the sleep disorder symptoms considered [35]. Obviously, the impact of physical impairment on life and, consequently, also on sleep is particularly high on subjects with clinical conditions, causing more frequent disability, such as rheumatological and neurological diseases. In patients with juvenile rheumatoid arthritis disability was associated with subjective and objective measurements of sleep disruption [36], [37]. Parkinson’s disease, poliomyelitis and cerebral palsy are neurological disorders able to produce major physical problems. Several studies performed on subjects affected by these complaints showed that the quality of sleep depended directly on disability [38], [39], [40]. Despite the clinical condition determining disability, physical impairment seems to disrupt sleep by means of a compromised bed mobility able to cause difficulty in turning in bed or in getting out of bed unaided. We hypothesize that this mechanism might also be present in severely disabled MS patients.

Pain is a distressing symptom associated with MS. In patients with MS, pain is related to demyelization (e.g. Lhermitte’s sign, trigeminal neuralgia, dysesthetic extremity pain), spasticity (e.g. back pain and painful spasm), weakness, incoordination and immobility [41]. Its prevalence among MS patients seems to be approximately 50%. Pain and sleep represent two clinical conditions strictly associated. Acute (abdominal surgery, burns and coronary heart disease) and chronic (rheumatic diseases) pain may produce sleep fragmentation and a lower percentage of deep and paradoxical sleep. These sleep consequences seem to be due to repetitive microarousals and sleep shifts caused by painful stimuli. It is important to underline that not all painful stimuli provoke cortical activation during sleep; the occurrence or not of microarousals depends on the type and duration of sensory stimulation (long chemical or mechanical stimulations have higher effects on arousal) and on the stage/phase of sleep (light sleep is more easily affected by arousal). In addition, disrupted sleep may interfere with pain processes enhancing pain sensitivity. In MS patients the effects of chronic pain seem to be relevant to sleep. In fact, Stanton et al. showed an important negative effect of pain on each indicator of insomnia, with a main effect on the difficulty of falling asleep [1]. Our study confirms these results, showing a significant impact of pain on nocturnal rest. Therefore, suitable pharmacological or non-pharmacological treatment for pain should be administered to patients with MS who are experiencing pain, not only to control this symptom, but also to facilitate sleep.

This is the first study to consider comorbidity as a risk factor for disrupted sleep in patients with MS. Studies performed in dialysis patients have shown that comorbidities, measured by means of the CCI, impair nocturnal sleep [42], [43]. Obviously, mean CCI scores in these dialysis patients were much higher than those in our patient population, but the role of comorbidities as a factor capable of influencing quality of sleep remains unaltered and requires more attention in clinical practice.

Associations between sleep disturbances and anxiety [1] and depression [3] have been previously reported in patients with MS. Our results showed that the number of MS patients with a clinical diagnosis of anxiety and depression was significantly higher among poor sleepers than among good sleepers. However, the impact of these psychological conditions on sleep seems to be different in our sample. Indeed, our correlation analyses demonstrated that the degree of quality of sleep was not dependent on the severity of anxiety, whereas it was strictly correlated with the level of depression. Sleep disruption and depression are notoriously associated. The relationship between sleep and depression seems to be bidirectional; in fact, poor sleep can predict depression and, on the other side, a depressed mood may be a predictor of a disrupted sleep. Patients affected by poor sleep and associated depression should be treated with antidepressants that have significant effects on nocturnal rest by treating the underlying mood disorder and by directly modifying nocturnal sleep. In our opinion this therapeutic approach is correct for all kind of patients, including those affect by MS. Since a large part of MS patients are depressed, we suggest a careful evaluation and treatment of this psychological condition in order to improve their sleep as well.

In patients affected by chronic diseases (e.g. cancer, renal and heart failure, chronic pain, rheumatic diseases, neurodegenerative disorders, etc.) the quality of sleep seems to have a primary role into impairing QoL. Specifically, advanced cancer patients presenting a poor QoL may experience a poor quality of sleep, which has been demonstrated to be related to functional status, use of opioids, presence of metastasis and pain [44]. Hemodialysed patients with a poor quality of sleep, in contrast with good sleepers, have a higher prevalence of depression, lower levels of hemoglobin and lower scores in all SF-36 domains [45]. In patients with stable heart failure, quality of sleep is related to the severity of heart failure, measured by means of the New York Heart Association classification, and to both physical functions and mental health evaluated using SF-36 [46]. Theoretically these results come to the conclusion that any chronic and/or progressive diseases could be able to affect QoL also by means of a poor quality of sleep. To the best of our knowledge, similar findings have not been reported in MS patients. Indeed, the study by Lobentanz et al. is the only one to have included quality of nocturnal sleep among the possible predictors of QoL in patients with MS [23]. According to the statistical analysis conducted by these researchers, the impact of sleep on QoL was limited; in fact the authors did not detect a significant difference in QoL between good and poor sleepers, and the global PSQI score represented an independent predictive factor only for the QoL domain of physical well-being. On the basis of our clinical experience, we hypothesized that disrupted sleep in patients with MS could have a crucial role in determining poor QoL; consequently we decided to verify the findings of Lobentanz et al. Statistical analysis confirmed our clinical impression, showing that poor sleepers, compared with good sleepers, reported poorer scores for QoL on each SF-36 domain. In addition, the global PSQI score was an independent predictor of both mental (MCS) and physical (PCS) status in patients with MS, together with sexual and/or bladder dysfunction (for MCS), and age and EDSS (for PCS).

The observation that type and frequency of treatment for MS did not appear to affect MCS or PCS scores is in contrast with previous observations that IFN beta treatment was the main predictor of poor QoL in MS patients [22]. On the contrary, if our findings should be confirmed, this would even indicate an indirect positive effect of immunomodulating treatments on the QoL of MS patients, given their effectiveness in reducing the progression of EDSS score, which was identified in this study as a strong and independent predictor of PCS score.

One limitation of our study should be discussed. Despite the fact that PSQI is used throughout the world (including Italy [47]) in sleep medicine studies, an Italian version of this questionnaire is still lacking. However, we consider that the impact of this limitation in our study is minimal. In fact, the Italian version that we used should be reliable, having been translated from English into Italian and then retranslated into English and compared with the original by two independent professional translators blinded to the previous version.

In conclusion, as suggested by Crayton et al., we are of the opinion that, in patients with MS, individual symptoms, if left untreated, may worsen or precipitate other symptoms, producing a vicious circle [48]. Although symptomatic treatment of MS should be comprehensive and integrated, in our opinion disrupted sleep should be treated immediately since (unlike other complaints, such as disability, sexual and/or bladder dysfunction and spasticity) the quality of nocturnal sleep could quickly improve following suitable treatment, thus breaking this vicious circle and improving patients’ mental and physical status. Consequently, we recommend fast, careful diagnosis and treatment of disrupted sleep in patients with MS, bearing in mind that in addition to the effects of specific sleep disturbances, other clinical conditions (both related and unrelated to MS) might affect nocturnal rest in this population.