Excessive Daytime Sleepiness and Epilepsy: A Systematic Review

Background. Sleep complaints are common in patients with epilepsy (PWE). Excessive daytime sleepiness (EDS) is one of the most reported complaints and its impact is still a matter of debate. Objective. Evaluate the relationship between EDS and epilepsy, with emphasis on prevalence, assessment, and causes. Methods. A systematic review on PubMed database in the last 10 years (2002 to 2012). The search returned 53 articles and 34 were considered relevant. After citation analysis, 3 more articles were included. Results. Most studies were cross-sectional and questionnaire based. 14 papers addressed EDS as the primary endpoint. 14 adult and 3 children studies used subjective and objective analysis as methodology. The number of studies increased throughout the decade, with 21 in the last 5 years. Adult studies represent almost three times the number of children studies. EDS prevalence in PWE varies from 10 to 47.5%. Prevalence was higher in developing countries. Conclusion. EDS seems to be related more frequently to undiagnosed sleep disorders than to epilepsy-related factors, and although it affects the quality of life of PWE, it can be improved by treating comorbid primary sleep disorders.


Introduction
Sleep disorders are now recognized as a major impairment in quality of life and work productivity. These complaints are especially common in patients with epilepsy (PWE) normally with more severe consequences than in the general population [1]. Fragmented or inadequate sleep can exacerbate daytime drowsiness and memory dysfunction, which are already present in epilepsy either because of the pathological substrate or the use of antiepileptic drugs (AEDs) and also contribute to intractable seizures [1]. There is a cycle created by sleep disruption leading to worsening seizures, which in turn leads to an even greater impairment of sleep [1]. One of the most reported sleep-related complaints in PWE is excessive daytime sleepiness (EDS) and it has been mainly interpreted as a side effect of AEDs treatment and frequent seizures [2,3]. This does not seem to be the case. Several studies have shown that EDS in PWE is comparable to that of controls and that symptoms suggestive of obstructive sleep apnea (OSA) and restless legs syndrome (RLS) are stronger predictors of subjective daytime sleepiness than frequency of seizures or AED [2,3].
One of the problems faced by studies addressing EDS specifically is that a uniform operational definition is still lacking, probably because it is not a disease or a disorder but a symptom presented in primary sleep disorders, such as narcolepsy, OSA, and RLS. The International Classification of Sleep Disorders 2nd edition (ICSD-2) defines EDS as "the inability to stay awake and alert during major waking episodes of the day, resulting into unintended lapses into drowsiness or sleep" [4]. Since EDS has a subjective interpretation, it could be mistaken as tiredness or fatigue, thus presenting a diagnostic challenge [5].
EDS can be studied by subjective or objective evaluation. From a subjective point of view, structured questionnaires and scales that are validated in the general population are commonly used. The Epworth Sleepiness Scale (ESS) is the most common instrument used in sleep research [6]. It is a quick self-applied scale with 8 typical dozing situations with a 0 to 3 point score for each question. With a maximum of 24 2 Epilepsy Research and Treatment points, scores >10 are considered abnormal for most authors [6,7] and it has been validated to Brazilian Portuguese [8]. For the objective evaluation, the multiple sleep latencies test (MSLT) and maintenance of wakefulness test (MWT) are commonly used. Although these tests were developed to address daytime sleepiness in hypersomnia syndromes, they also have been widely used to study sleepiness in other situations. The tests should be performed on the following day of a full night polysomnography (PSG). The patient is given five 20-minute opportunities to sleep (MSLT) with a 2-hour interval between the recordings. The mean sleep latency is calculated and values <10 minutes are considered abnormal. For the MWT, four 40-minute opportunities to stay awake are given and a mean sleep latency <8 minutes is considered abnormal [9]. Other types of objective evaluation for EDS are pupillometry, cognitive, and psychomotor function, evoked potentials, and alpha attenuation test (AAT). Of those, pupillometry and long latency cortical and cognitive evoked potentials are less used either because of lack of studies demonstrating its clinical utility or by large degree of intersubject variability [10]. Cognitive and psychomotor tasks are used to test the effects of sleep disruption on cognitive and psychomotor performance. They can also be used to assess response to treatment, but they only evaluate the effects of sleepiness on the efficiency of brain function and do not measure sleepiness directly. AAT is based on changes of the alpha frequency that occur on the transition from wakefulness to sleep. Patients are instructed to open and close their eyes 8 times, with each movement lasting 1 minute, while being in an illuminated room. AAT correlated significantly with MSLT, even higher than subjective instruments [10].
The objective of this review is to analyze the current evidence on EDS and epilepsy, with emphasis on prevalence, assessment, and causes.

Methods
We performed a systematic review to identify all articles addressing EDS in PWE. A PubMed database search was performed for all publications existing and "in press, " in the last 10 years (October 2002 to October 2012). The search terms used were "sleepiness" OR "sleep disruption" OR "disturbed sleep" AND "epilepsy" AND "sleep." The keywords were restricted to title and abstract. No language restriction was made and citations of relevant studies were checked. Exclusion criteria included review papers and articles that do not use a validated questionnaire for subjective analysis as part of the research methodology. The search returned 53 articles that were reviewed and 34 articles were selected. After citation analysis, three more articles were added because of their relevance [2,3,11].

Results
Of the 37 articles selected, 27 were adult-based and 10 were children-based studies. Regarding the study design, the articles were 23 cross-sectional studies, 11 prospective cohort studies, 2 case reports, and one case series.

Evidence from Adult-Based Studies.
Most of the questionnaire-based studies addressed the prevalence of either EDS or primary sleep disorders. Table 1 summarizes these studies. Evidence shows that sleep disorders (especially OSA and insomnia) are twice to 3 times more common in PWE than in controls [12][13][14]. On the other hand, the data on EDS is not so straightforward, with studies showing that the prevalence is similar to controls [2,3,12,13] and others reporting a high prevalence in PWE [14][15][16][17][18]. An ESS score >10 was reported in 18 to 47% of PWE and 12 to 17% of controls [2,15,16,18,19], with a trend for higher ESS scores in patients with intractable seizures [17]. The most common sleep-related complaints in PWE were maintenance insomnia and EDS [13,14,16]. Sleepiness in PWE may have a multifactorial origin and studies show that symptoms of OSA and RLS are independent predictors of an ESS score >10 [2,3,20]. This has been also demonstrated in patients of older age and by the fact that the treatment of OSA with continuous positive airway pressure (CPAP) can improve EDS and seizure control [21]. Another study prospectively analyzed the effect of CPAP treatment in 29 epilepsy patients with OSA diagnosed on PSG and with EDS evaluated with the ESS. The median followup was 26 months and 52% had an ESS score >10. In 21 patients, OSA symptoms coincided with a clear increase in seizure frequency or the occurrence of status epilepticus. Patients with good CPAP compliance had a significant reduction in EES score and in seizure frequency [22]. In contrast, a study with 125 PWE determined the cut-off of the sleep apnea scale of the sleep disorders questionnaire (SA-SDQ) in this population and suggested that in PWE the values should be lower than in the general population and that EDS may not be a good predictor in epilepsy as in the general population [23]. Very few studies addressed EDS in PWE with subjective and objective analysis. A study conducted in Brazil evaluated EDS with ESS and PSG/MSLT in 39 patients with temporal lobe epilepsy (TLE). The most frequent complaints were daytime sleepiness (85%), frequent awakenings (79%), and nocturnal seizures (69%), 13% had OSA, and ESS scores correlated with MSLT mean latencies. EDS was found in 36% (ESS score >10) [24].
Another common finding is the relationship between EDS and depression or anxiety in PWE. A study with 247 patients showed that anxiety, depression, and sleep disturbance had greater effect than seizure control in quality of life (QOL) scores and also found underdiagnosis of psychiatric comorbidity in PWE [25]. EDS complaints were reported in 47.5% of 99 PWE and correlated with anxiety and neck circumference [19].
Although previous reports suggest that seizure type, AEDs, and nocturnal seizures could contribute to disrupted sleep and EDS, the presence of primary sleep disorders seems to play a significant role in this complex interaction. Studies have demonstrated that nocturnal seizures could disrupt sleep structure and even brief seizures could result in prolonged alterations in sleep architecture that lasts longer than the postictal period and their treatment results in improvement of sleep parameters [1]. This effect could help explain why patients with only nocturnal seizures have daytime complaints of sleepiness. On the other hand, some studies suggest that nocturnal seizures were not significant   predictors of EDS [2,13,16,26]. In one study of patients with nocturnal frontal lobe epilepsy (NFLE), EDS was similar between patients and controls, although daytime symptoms of sleepiness could be more frequent in a subgroup of patients with higher ESS scores, irrespective of seizure frequency [26]. In a study evaluating sleep disturbances in patients with refractory and controlled epilepsy with subjective and PSG evaluation, EDS and OSA were more frequent in the refractory group, and it also found shorter average sleep time in contrast to self-reported longer sleep time in this group of patients, suggesting that OSA and sleep fragmentation could have a negative impact in seizure control [17]. The role of AEDs in EDS has also been studied and the evidence shows that patients with a stable medication 6 Epilepsy Research and Treatment regimen do not report more EDS [18,19]. A study with levetiracetam showed that it can initially reduce sleep time and disturb sleep architecture, but as the treatment proceeded this effect was less important [27], and there was a report of a patient with severe but potentially reversible hypersomnia without subjective EDS, with levetiracetam as an add-on treatment [28]. The effects of topiramate and pregabalin on daytime sleepiness have also been studied. 14 patients were prospectively studied with a short monotherapy course with topiramate 200 mg/day and it did not impair daytime vigilance assessed by MSLT and visual reaction time [29]. Pregabalin, in a prospective study with 12 patients, improved seizure control, increased REM sleep, and reduced stage N2. There was also an increase in ESS score but within the normal limits, suggesting mild daytime sleepiness [30]. Another prospective double-blind, randomized study evaluated the effect of pregabalin 300 mg/day versus placebo on polysomnographic variables in 17 patients with well controlled partial seizures and reported sleep disturbance. The pregabalin group showed improvement in sleep continuity with reduction of awakenings and improvement in wake time after sleep onset, and also improvement in subjective scales [31].
The role of epilepsy surgery has also been addressed by some studies. On a retrospective review of 21 patients treated with resective surgery for drug-resistant focal epilepsy (NFLE), 9 had complaints of EDS was preoperatively and after surgical treatment all patients were seizure-free and EDS resolved in those who previously reported it, despite the maintenance of AED treatment, suggesting that the arousal instability that caused sleep fragmentation and EDS was related to recurrent epileptic discharges not detectable on scalp EEG [32]. The same group reported a case where the patient presented EDS and periodic leg movements that were related to epileptic discharges revealed only by stereo-EEG and that improved after surgical treatment [33]. In a prospective cohort of TLE patients EDS was evaluated before, 3 months, and 1 year after surgical treatment. Results showed decrease in ESS and PSQI scores postoperatively and analysis for nocturnal seizures was not statistically significant [34]. The effects of vagus nerve stimulation (VNS) were also studied and daytime vigilance after 6 months of VNS treatment (evaluated with MSLT and visual reaction time) with low intensity stimulus (≤1.5 mA) improved, with positive impact on QOL [35].

Evidence from Children-Based Studies.
In children with epilepsy (CWE), like adults, questionnaire-based studies reveal that they are more likely to have sleep complaints than the general pediatric population [11,[36][37][38][39]. In Table 2, a summary of the studies addressing the relationship between epilepsy and EDS is displayed. A combination of nocturnal seizures disrupting sleep architecture, sedative medication, and the presence of primary sleep disorders may represent a part of the multifactorial nature of the problem [40,41]. EDS and sleep initiating/maintaining complaints are the most reported issues [37,39]. A study with 26 patients showed EDS, parasomnia, and sleep-disordered breathing (SDB) to be more frequent in the epilepsy group and also defined parasomnia and SDB as independent predictors of EDS, while epilepsy syndrome, AEDs, and seizure freedom were not statistically significant [36]. The role of EDS and behavioral impairment has also been studied in CWE and their parents. In a group of 97 CWE, the autistic spectrum disorder (ASD) was screened together with sleep disorders. The worst behavior and daytime sleepiness were seen in children with greater risk of having ASD, suggesting that behavioral difficulties and EDS could affect their ability to learn [42]. A recent study evaluated disrupted sleep in CWE and their parents (105 households and 79 controls) and reported high rates of parent-child room sharing and cosleeping in the epilepsy group. CWE had greater sleep disturbance (especially parasomnia, EDS, and sleep walking) and their parents had more fatigue and sleep dysfunction. Severity of epilepsy correlated with child and parent sleep dysfunction and also with parent fatigue [43]. Maternal depression has also been evaluated. In a study of 52 mothers of CWE, 45% had high scores on depression scales with 25% in the moderate/severe range. Maternal depression correlated with behavioral problems (attention deficit) but not with epilepsy [44].
The effect of medication (AED and other sleep-active drugs) on daytime sleepiness of CWE is also a matter of discussion. One of the most studied sleep-active drugs in children is melatonin. Melatonin can reduce excitability of glutamate secreting neurons and increase activity of GABAergic inhibitory neurons, and it also has metabolites that act as an anticonvulsant and a free radical scavenger. It has been suggested that melatonin may have an anticonvulsant effect [38,45]. A study with 37 patients (23 with intractable seizures) addressed the role of melatonin in improving sleep-related complaints. Melatonin was associated with improvements in EDS and sleep-related complaints and also with significant reduction in seizure severity [38]. The role of AED in EDS of CWE is less straightforward. A study with 46 children taking valproic acid (VPA) assessed sleep duration and sleep behavior before and after tapering it in patients treated for more than 6 months. Subjective and objective (actigraphy) evaluation was performed at baseline, 8, and 12 week after treatment. Actigraphy data showed reduction in the average sleep without VPA in 33 children and increase in 13 children. Mean actual sleep time per day was reduced after VPA termination, but the reduction was only significant in children older than 6 years [46].

Conclusion
The relationship between EDS and epilepsy is still a matter of discussion. Although EDS in PWE is comparable to the general population, it carries more severe consequences in QOL and the presence of primary sleep disorders seems to contribute more to EDS symptoms than epilepsy-related factors. In CWE, EDS seems to contribute to parent-child cosleeping and to behavioral and learning disabilities. Only a few studies done in the last decade test the patients with subjective and objective evaluation, probably because of the logistic difficulties of performing PSG and MSLT in a bigger  (population) set. Difficulties apart, EDS in epilepsy is a prosperous area of research and more knowledge is needed to understand (1) a better definition of EDS and fatigue in research methodology; (2) more studies combining subjective and objective evaluation; (3) the role of medication; (4) the role of psychiatric comorbidity; and (5) impact of treatment of coinciding primary sleep disorders.