If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Diagnosing abnormal nocturnal events or sleep complaints can be challenging.
We reviewed the utility of combining PSG with 18-channel EEG in selected patients.
This one diagnostic test can be helpful in differentiating these complaints.
Standard polysomnography (PSG) typically utilizes 4–6 channels of electroencephalography (EEG), which is inadequate to evaluate focal epileptiform activity. Though technical capability has long existed for more extensive EEG recording, few sleep laboratories have utilized this technique. The objective of this study was to determine the utility of combining PSG with 18-channel EEG in select patients with paroxysmal nocturnal events or other symptoms concerning for sleep disorders, nocturnal seizures or both.
Consecutive combined PSG–EEG studies (n = 237) were performed between 10/1/2005 and 8/1/2009. Demographics, referral source, indications, and results were reviewed and analyzed.
Of the 237 consecutive combined PSG–EEG studies performed, 93% revealed the presence of a primary sleep disorder, 38% were shown to have abnormal EEGs, and 37% had both. Among the 221 subjects (93%) shown to have sleep disorders, the majority of these cases were obstructive sleep apnea (OSA) 89%, followed by periodic limp movements of sleep (PLMS) 22% and rapid eye movement behavior disorder (RBD) 6%. Significantly more patients with known epilepsy were diagnosed with OSA then were patients without a seizure history.
Combined PSG–EEG, utilizing 18-channel EEG, is an under-utilized technique which can assist in diagnosing paroxysmal nocturnal events, and differentiate between the presence of a primary sleep disorder, seizure activity, or both. Our study further illustrates the importance of considering sleep disorders in epilepsy patients.
Case #1: JM is a 50yo man with a history of partial seizures involving hypermotor movements and dyscognitive features, fairly well controlled on anticonvulsant medications with only rare breakthrough events. He reports to his neurologist that he often feels sleepy during the day. His wife mentions that the patient often has leg twitching and abrupt body movements in his sleep.
Case #2: TS is a 68yo obese woman who was in a severe motor vehicle accident several months ago which resulted in brief loss of consciousness. Since the accident she has experienced poor sleep. She reports not feeling rested upon waking, and often has headaches in the morning. Sometimes she wakes up in the middle of the night with an intense feeling of panic and nausea, she wonders if these are nightmares. Her roommate says she snores loudly.
Reports of paroxysmal nocturnal events are not uncommon for neurologists, sleep specialists and primary care physicians to hear, yet can be quite challenging to diagnose. There is an increased prevalence of sleep disturbances in the epilepsy population, along with increased sleep complaints.
Reports of excessive daytime sleepiness, insomnia, or atypical nocturnal behaviors are also commonly reported among patients with epilepsy. Additionally, many types of movements or events that occur in sleep can mimic seizures and be mistakenly diagnosed as such. Common examples include non-rapid eye movement (NREM) parasomnias, sleep-related movement disorders, and even REM behavior disorder.
Conversely, epilepsy syndromes can be mistaken for a sleep disorder when the events occur exclusively in sleep or at the sleep–wake transition. An example includes paroxysmal nocturnal dystonia, which was initially described as a movement disorder, however, the patients used for the initial publication describing the disorder were ultimately found to have frontal lobe seizures.
Nocturnal seizures and parasomnias share some similar characteristics; both present at night, may be associated with amnesia for the event, can impair sleep, and be provoked by stress or sleep fragmenting factors. Occasionally, sleep disorders and seizures can also occur together. The importance of treating sleep disorders, especially OSA, in seizure patients has been demonstrated by Malow and others’ work, which brought to light this important association.
It is often very challenging by history alone to determine whether a patient is suffering from a primary sleep disorder, epilepsy (especially of frontal lobe origin), or possibly both. Distinguishing between the two categories of disorders can also be extremely difficult by merely witnessing the movements, either by a bed partner or by video. Routine electroencephalography (EEG) has traditionally been the primary diagnostic test for patients with suspected epilepsy, though its use is limited both in its sensitivity and by the low likelihood of capturing a typical event during the short recording. Some epileptiform abnormalities appear on EEG exclusively or more frequently during drowsiness or sleep, though these states are often not captured on routine EEG studies, which typically record for only 20–30 min. Longer, ambulatory EEG studies do not provide sufficient information regarding whether a potential sleep disorder is present due to lack of respiratory and electromyography (EMG) channels. In the United States (US), standard polysomnography (PSG), used to diagnose sleep disorders and/or evaluate sleep architecture, typically utilizes only 4–6 EEG channels (ROC, LOC, and referentially connected F3, C3, and O1 connected to the contralateral mastoid), with backup electrodes F4, C4, O2, or alternatively Fz, Cz, and Oz. While adequate for the purposes of determining sleep state, these are often insufficient to detect focal ictal activity,
aimed to determine the validity of abbreviated EEG montages for seizure detection during PSG. They retrospectively reviewed ictal recordings from 56 patients with epilepsy and 60 non-epileptic events that were acquired using the 10–20 electrode system on videoEEG. All 116 files were then displayed in each of two montages, one with only 8 channels and the other with the full 18 channels, and reviewers were asked to determine whether the file contained a seizure. They concluded that abbreviated EEG montages fail to adequately differentiate seizures from non-epileptic nocturnal events during PSG. Our study, the first of its kind to our knowledge, is a review of 237 prospectively obtained PSGs with 18 channel EEG examining whether this one combined test can be used alone for diagnostic purposes in a very select group of patients. We included not only ictal patterns in our study, but also epileptiform discharges and other EEG abnormalities.
In Case #1 above, the patient with epilepsy is experiencing daytime sleepiness and there are reports of leg and abrupt body movements in his sleep. What is causing the daytime sleepiness? Could it be from OSA? Could the leg movements be periodic limb movements of sleep (PLMS)? What if these are nocturnal seizures? A comprehensive evaluation in this clinical setting must include PSG, however, that alone with the standard 4–8 channel EEG is not sufficient to exclude nocturnal seizures. In such a situation, full head EEG recordings are also crucial, but currently, at least in the US, must be obtained separately in the form of either routine, ambulatory EEG or inpatient videoEEG monitoring. Similarly, the patient in Case #2 above requires both a PSG to evaluate for possible OSA or other sleep disturbances as well as full head EEG, as temporal lobe nocturnal seizures are also on the differential diagnosis. Obtaining the correct diagnosis in such cases is very important, as treatment and counseling for each is quite different, and missing the diagnosis could potentially be dangerous for the patient.
Case #3: AA is a 21yo woman with a history of anxiety, depression, and recent increased stress. She has a history of unusual behaviors from sleep, which were attributed to dream enactment in the setting of initiation of a selective serotonin reuptake inhibitor. She reported many sleep problems, in part related to her anxiety and stress. She had fragmented sleep, was resistant to treatment with trazodone and other anti-depressants, clonazepam lead to intolerable sedation. There were recent reports of possible staring episodes of unclear etiology.
The patient in Case #3 had an 18-channel EEG combined with PSG in our sleep laboratory. No evidence of a primary sleep disorder was present, but a seizure was captured during the recording (see Fig. 1 below), solidifying her diagnosis. Though the technical capability has long existed for more extensive EEG recording during sleep studies and is utilized in some European countries, few sleep laboratories in the US have utilized this technique. Clearly if only OSA or other sleep disorders (i.e. PLMS) are suspected in a patient, extended EEG data is not necessary or helpful. In this study, however we sought to determine whether a diagnosis could be made in this select group of patients, where both sleep disorders and nocturnal seizures are on the clinician's differential diagnosis by using this one diagnostic test: combined PSG with 18-channel EEG.
The study was approved by the human subjects committee at our institution. We reviewed all combined PSG–EEG studies, performed between 10/1/2005 and 8/1/2009 at a sleep laboratory affiliated with a major academic hospital (Brigham and Women's Hospital, Boston, MA). We examined demographics, referral source, study indications, and results. Data were recorded on a Nihon-Kohden system. Recording montage included standard EOG (electrooculography/eye movements), EMG, airflow (oro-nasal thermistor and nasal pressure), snoring, respiratory effort (thoracic and abdominal motion), arterial oxygen saturation, body position, EKG, and leg movements (anterior tibialis EMG). Video-EEG monitoring was performed throughout the sleep study recording using anterior temporal, as well as standard 10–20 EEG electrodes. Spike and seizure detection was performed visually and with the aid of a computer program to evaluate for any epileptiform features. The studies were interpreted by a neurologist dually trained in epilepsy and sleep medicine (M.P.). Though the recordings were obtained prospectively for clinical purposes, this was a retrospective review of the combined PSG–EEG reports. We do not have clinical information regarding details of the patients diagnosed with PLMS in the study. The diagnosis of RBD was made by a sleep specialist (M.P.) using the clinical data provided (typically unusual behaviors in sleep) combined with REM sleep without atonia on the recording, or by actually capturing the events in question. Data were not available to determine exactly how the results of the recordings impacted the patients’ clinical care. We presumed, however, that when a sleep disorder was diagnosed, it was treated accordingly. We further presumed that if no seizures or EEG abnormalities were found on the study, that the physician was able to exclude, or presume much less likely, nocturnal seizures as cause of the abnormal behaviors (i.e. limb movements) or symptoms (i.e. morning headache, daytime sleepiness) reported. Our analyses were based on the diagnoses made as a result of the combined PSG–EEG study. Fisher's exact test was used for comparison of frequencies.
Patients were referred mainly by a neurologist or primary care physician for evaluation of unusual nocturnal behaviors and/or daytime sleepiness, where nocturnal seizures, a primary sleep disorder, or both were on the differential diagnosis. Information regarding the referring physicians was known for 172 of the 237 patients. Neurologists comprised approximately 75% of the referring physicians (130 patients), primary care physicians 20% (34 patients), and psychiatrists 5% (8 patients). The mean age of the subjects was 45 (range 16–85). There were 107 men, and 130 women. None of the patients reported any adverse events related to the additional EEG leads. The frequency of sleep disorder diagnoses and EEG abnormalities is summarized in Table 1. Of the 237 consecutive combined PSG–EEG studies performed over four years, 93% revealed the presence of a primary sleep disorder, 38% were shown to have abnormal EEGs, and 37% had both. EEG findings were categorized as either showing definite interictal epileptiform discharges or as abnormal but without definite discharges, the latter referring to focal slowing, rhythmic patterns or any asymmetry. Two patients had subclinical electrographic seizures captured during the study. Among the 221 subjects shown to have sleep disorders (93%), the most frequent diagnosis was OSA (89%), followed by PLMS (22%), and RBD (6%). Detailed clinical information regarding a history of seizures was only available on 53 patients; 27 of which had a known seizure disorder, 26 did not have a history of seizures. None of the patients in this cohort had normal PSGs. Of the known epilepsy patients, 23 (85%) were diagnosed with OSA in the study, while only 13 (50%) of the patients without a seizure history showed OSA. Both of the patients who had seizures recorded during the study were diagnosed with OSA. Unfortunately the details regarding which antiepileptic drugs (AEDs) the epilepsy patients diagnosed with OSA were on was not available to us for this analysis, though would have been interesting to determine whether patients on certain AEDs, such as those associated with weight gain, were more likely to have OSA. Table 2 delineates the degree of the OSA in this subset of patients by respiratory disturbance index (RDI), an index used to assess the severity of sleep apnea based on the total number of apneas (complete cessation of breathing) and hypopneas (partial breathing obstruction) noted per hour during sleep. About a third of the patients with known epilepsy had moderate severity of disease (RDI > 20); more than double that of the group without seizures (12%). Fifteen (55%) had periodic limb movements of sleep (PLMS). Six (23%) had definite interictal epileptiform discharges. Of the 4 patients with no known seizure disorder who had discharges or an otherwise abnormal EEG, all of them were diagnosed with OSA, and 2 were also diagnosed with PLMS. Of the 7 patients with known seizure disorders and discharges or otherwise abnormal EEGs, 4 were diagnosed with OSA, one of those 4 was also diagnosed with PLMS (Table 3). Diagnosing these sleep disorders in these patients indicate to the clinician that the nocturnal movements or complaints of daytime sleepiness were less likely to be seizure-related, but rather due to the sleep disorder.
Table 1Demographic distribution and frequency of diagnoses made among all participants.
n = 237 (%)
Definite epileptiform discharges
Abnormal EEG, but no definite discharges
Any sleep disorder
Abbreviations: EEG: electroencephalogram; PLMS: periodic limb movements of sleep; OSA: obstructive sleep apnea; RBD: rapid eye movement behavior disorder.
In this cohort of 237 consecutive patients, we found that combined PSG with full EEG can be very helpful diagnostically in a select population including (1) patients with known epilepsy with sleep complaints or daytime sleepiness, (2) patients with undiagnosed nocturnal events, and (3) patients with unusual daytime alterations of awareness and no definitive diagnosis. The test is well tolerated by patients and informative for the clinician. Our findings extend those of Foldvary-Schaefer et al.
, who demonstrated that 18-channel EEG is superior to 8-channel EEG when attempting to identify seizures on PSGs. Most of the discharges and other EEG abnormalities detected in our study were focal, thus would likely have been missed, misinterpreted, or required more evaluation if a limited EEG montage was used.
We found a substantially higher prevalence of OSA among patients with known epilepsy as compared to those without; this was statistically significant (p = 0.006), and consistent with the findings of Malow's group.
Another important finding is the degree of severity seen in the OSA, as determined by RDI. Many of these patients were found to have moderate disease, indicating the necessity of CPAP or other therapies in this population.
Our analysis has several limitations. We analyzed a clinical cohort rather than an experimental population, thus we could not adequately calculate sensitivity or specificity. For the same reasons, we could not randomly assign patients to different types of tests. In future studies, a longitudinal design may allow more detailed measurement of outcomes, such as cost effectiveness relative to serial EEGs or other tests, and health outcomes due to optimized treatments as a result of this combined study. We can report, however, that after obtaining the combined 18-channel EEG–PSG study, only 2 of these patients required referral to our epilepsy monitoring unit (ours being the primary affiliated epilepsy center for this cohort of patients) to further assist with diagnosis, suggesting that the combined study was sufficient for diagnosis in the vast majority of cases. Further study is needed to determine whether these assumptions are valid, as the positive diagnostic yield is relatively low given the time and effort involved. The yield appears to be comparable to the sensitivity of capturing seizures on an ambulatory EEG,
however, and possibly higher than some cancer screening methods.
Many patients suffering from abnormal nocturnal events or associated sleep complaints are left without a diagnosis after routine EEG, continuous video-EEG, or PSG alone. In the US, most inpatient epilepsy monitoring units do not perform PSGs on their patients, and most sleep centers only use the minimum EEG leads required to determine sleep state, or often home sleep studies, that do not use EEG at all. There are only rare studies evaluating combining these two diagnostic processes. Though we did not have detailed information looking at how the results changed management of the referring physician, 93% of the patients were diagnosed with a sleep disorder, and referred for treatment. Given current knowledge about the co-occurrence of sleep disorders and epilepsy, the adverse effects sleep disorders can have on seizures (i.e. the importance of treating OSA in seizure patients), and the difficulty in determining between the two clinically, we expect that this technique will become more frequently used in this patient population. Recommended indications for use of combined PSG–EEG using 18-channel EEG are delineated in Table 4. This is a technique that is under-utilized in US sleep laboratories which can assist in diagnosing paroxysmal nocturnal events accurately, and differentiate between the presence of a primary sleep disorder, seizure activity, or both.
Table 4Recommended indications for combined PSG–EEG.
Patient type/clinical situation
Purpose of the study
Known epilepsy + sleep complaints/daytime sleepiness
Determine whether the cause of the sleepiness is related to nocturnal seizures or a primary sleep disorder
Undiagnosed nocturnal events
Diagnose the events: parasomnia, seizure, or both
Daytime alterations of awareness and no definitive diagnosis
Evaluate for evidence of seizures in overnight recording vs. primary sleep disorder as cause of daytime symptoms.
☆Disclosures: Dr. Pavlova was principal investigator on a CATALYST program research project that is not related to the current study. All authors verify that there are no conflicts of interest, financial or otherwise, related to this work. No funding was provided for this study.