Sleep Disorders in Leucine-Rich Glioma-Inactivated Protein 1 and Contactin Protein-Like 2 Antibody-Associated Diseases

Objective: Sleep disorders are common in voltage-gated potassium channel complex antibody (VGKC-Ab) diseases. The aim was to investigate the sleep disturbances and polysomnography (PSG) characteristics in patients with VGKC-Ab-associated diseases. Methods: Twenty-seven patients with leucine-rich glioma-inactivated protein 1 antibody (LGI1-Ab) encephalitis, seven patients with contactin protein-like 2 antibody (Caspr2-Ab)-associated diseases, and 14 healthy controls with at least one PSG or actigraphy recording were recruited at Peking Union Medical College Hospital from January 2014 to July 2019. Results: Sleep disorders including insomnia, hypersomnia, rapid eye movement (REM) sleep behavior disorder (RBD), periodic limb movements in sleep (PLMS), agrypnia excitata, and obstructive sleep apnea syndrome were observed. Twenty-one PSG recordings from patients with LGI1-Ab encephalitis demonstrated a decrease in total sleep time (TST) (median 365.5, range 184.5–495.5 min), sleep efficiency (70.0%, 47–92%), N3 sleep (9.7%, 0–32.9%), and REM sleep (9.9%, 0.4–27.9%). Of five patients with Caspr2-Ab-associated diseases, TST was found to be 329.5 (167.0–377.5 min), and sleep efficiency was found to be 61.7% (34.6–71.7%). The percentage for N3 and REM sleep was found to be 15.0% (0–34.6%) and 12.7% (0–22.2%), respectively. Both RBD and PLMS were observed more frequently in patients with LGI1-Ab encephalitis. We identified status dissociatus (SD) in five (23.8%) patients with LGI1-Ab encephalitis and two (40%) patients with Caspr2-Ab diseases. The former is more likely to have simple limb movements rather than complex movements, which mimic the contents of their dreams. Continuous insomnia was more common in patients with Caspr2-Ab diseases than patients with LGI1-Ab encephalitis. Patients reported clinical and PSG improvements following immunotherapy. Conclusion: Sleep disorders in patients with VGKC-Ab-associated diseases include decreased TST and poor sleep efficiency. Our studies provide evidence of SD in patients with LGI1-Ab encephalitis.


Patients
The study enrolled 48 participants with at least one PSG or actigraphy recording at the Peking Union Medical College Hospital (PUMCH) from January 2014 to July 2019. Twentyseven subjects were diagnosed with LGI1-Ab encephalitis and seven with Caspr2-Ab diseases. We included 14 age-matched healthy controls who had no known history of sleep disorders or neurological illness and no known factors that may affect sleep such as shift work. All had a Pittsburgh sleep quality index ≤5. The demographic information, clinical features, and laboratory test data were reviewed. The diagnosis of autoimmune limbic encephalitis was based on the criteria established by Graus et al. (14): (i) subacute onset of memory deficits, seizures, or psychiatric symptoms suggesting involvement of the limbic system; (ii) T2-weighted hyperintensities of medial temporal lobes; (iii) at least one of the following: cerebrospinal fluid (CSF) pleocytosis, electroencephalogram (EEG) epileptic, or slow-wave activity; (iv) exclusion of other causes. Morvan's syndrome was defined as a combination of cognitive symptoms or seizures, peripheral nerve hyperexcitability, and dysautonomia or insomnia (15).

Determination of Antibodies
We conducted a fixed cell-based indirect immunofluorescence test (IIFT) for LGI1 and Caspr2 Abs using Biochips (Euroimmun AG, Lübeck, Germany).

Sleep Assessments
Sleep was evaluated by video PSG and actigraphies. Overnight video PSG monitoring was performed using the 76-channel EEG/PSG recording system (Compumedics Grael series; Australia), at the sleep laboratory of PUMCH. The minimal recording time was 7 h. The following signals were recorded: at least six EEG channels (two frontal, two occipital, and two central), two electro-oculography channels, two electromyography channels (mentalis/submentalis and tibialis), nasal airflow, respiratory efforts (thoracic and abdominal), pulse oximetry, body position, and a single-led electrocardiogram. The results of PSGs were evaluated by experienced sleep specialists blinded to clinical diagnosis, on the basis of standard scoring practices (16). We analyzed the following sleep parameters: total sleep time (TST), sleep onset latency, sleep efficiency, the percentages of respective sleep stages, limb movement in sleep (LMS) index, micro-arousal index, and apnea-hypopnea index (AHI).
Sleep monitor by actigraphy was recorded using the ActiGraph (Compumedics, Australia), placed on the nondominant wrist, for seven consecutive days at PUMCH. Data were analyzed by experienced evaluators. For analysis of sleepwake differentiation, we used the ActiLife 6 software provided by the manufacturer (Compumedics ActiGraph). The software uses the algorithm by Cole et al., which was validated in an adult population (17). The 24-h TST was measured. All sleep assessments were performed during the acute phase of the diseases. Most PSG and actigraphy recordings were performed during treatment. All assessments were performed before marked clinical improvements.
Before sleep assessments, patients and bed partners were specifically asked about sleep complaints including insomnia, excessive daytime sleepiness, limb movements, dream enactment behavior, or snoring. Sleep specialists also evaluated the symptoms by video during PSG monitoring.
Patients who had trouble staying awake while driving, eating, or engaging in social activities were considered to have excessive daytime sleepiness. Daytime sleepiness was assessed by free interview. Dream enactment behavior was defined as nocturnal vocalizations and limbs or trunk movements with dream ideation reported by a bed partner. Status dissociatus (SD) was reported as a polysomnographic trait, which was characterized by (i) lack of the conventional features of non-REM (NREM) sleep with the disappearance of spindle and delta activities; and (ii) unstable REM sleep appearing in short recurrent episodes, isolated, or mixed with NREM potentials (18). The most severe type of SD, associated with nearly continuous motor overactivation and peculiar dream-like behavior, was known as agrypnia excitata. Agrypnia excitata had been reported in Morvan's syndrome, familial insomnia, and alcohol withdrawal syndrome (18,19). Patients who had difficulty in initiating sleep, difficulty maintaining sleep, waking up too early, or TST of <360 min resulting in daytime sleep impairment were considered to have insomnia. Continuous insomnia was defined as no sleep for more than 2 days. Clinical improvements were based on patient subjective feelings, family observations, and neurological examinations by clinicians.

Statistical Analysis
Continuous non-normal distribution data, such as PSG data, were described as medians with range. Mann-Whitney U test was performed for comparisons between two groups. The frequency data between two groups were compared by Fisher's exact test. All the tests were two-tailed, and statistical significance was set at p = 0.05.

RESULTS
A total of 120 patients were diagnosed with LGI1-Ab encephalitis at PUMCH. Fifty-one (42.5%) of these patients presented with sleep disorders. Twenty-four patients were diagnosed with Caspr2-Ab-associated diseases, of whom 11 (45.8%) patients had sleep disturbances. Data from all participants with detailed sleep assessments were reviewed as below.
LGI1 Antibody Encephalitis Twenty-seven patients with LGI1-Ab encephalitis underwent PSG or actigraphy recordings. All patients had positive serum/CSF LGI1-Ab, but none had positive Caspr2-Ab in serum or CSF. Eighteen (66.7%) patients were male. Their median age was 63 years (range 36-77 years). All patients met the criteria of limbic encephalitis. Almost all patients developed  Table 1). CSF studies showed mildly elevated protein in five patients (0.47-0.63 g/L) and elevated white cells (8 cells/µl) in one case. One patient had gastric cancer. One patient had a hyper-metabolism mass between the left subscapularis and anterior serratus muscles in PET/CT. This patient refused biopsy and was lost to follow-up. Nineteen (70.4%) participants showed abnormal brain MRI finding. Eighteen cases had mesial temporal lobe lesions, and two had basal ganglion lesions. 18 F-FDG PET/CTs were available for 17 patients. Among these, 82.4% patients demonstrated an abnormal metabolic pattern involving medial temporal lobes (n = 12), basal ganglia (n = 11), cortex (n = 2), and thalamus (n = 1). Twenty patients completed EEG recordings. Among these, 70% had abnormal patterns including 11 patients with epileptiform discharge, seven with focal or multifocal slow waves, and three with diffused slow waves.
PSG demonstrated REM sleep without atonia (RWA) in eight patients. Five of the patients had dream enactment behaviors and were diagnosed with RBD. Other abnormal findings included PLMS in 12 (57.1%) patients and OSAS (9.5%) in two patients (AHI ≥ 5/h) according to the International Classification of Sleep Disorders (3rd edition). The frequencies of FBDSs, RWA, PLMS, and epileptiform discharges were not significant based on PSG parameters including sleep efficiency, TST, and the percentages of the respective sleep stages (Supplement 1).
We found a complete breakdown of state-determining boundaries and severe sleep fragmentation in five patients. Their PSG results showed frequent shifts between wake, N1 and N2 sleep stages, and REMs frequently intermixed with slow eye movements (Figures 1A, 2), which were consistent with the features of SD. Three (60%) patients lacked N3 sleep with scanty or absent sleep spindles. Four (80%) patients lacked stable REM sleep. Of these five patients with SD, TST was 356.0 (267-495.5) min, and sleep efficiency was 68% (50-91%). None had complex movements mimicking the contents of their dreams. Instead, we found markedly increased limb movements in three cases (LMS index 241.6-320.8/h). Three patients had PLMS, two had RWA, and four had epileptiform discharges. None had OSAS. Patients with SD had a significant decrease of REM sleep (p = 0.011) and an increase of N1 sleep (p = 0.032), than had patients without SD ( Table 3). We also observed a trend of N3 sleep decline (p = 0.050). Other PSG parameters and frequencies of clinical symptoms showed no significant differences.
Ten patients underwent actigraphy studies. Continuous insomnia was observed in one patient (10%). A 24-h day-night activity rhythmicity was observed in nine patients, including one patient with SD ( Figure 3A). Actigraphies detected a long time sleeping (TST > 10 h) in one patient, with an average TST for 7 days of 1,025 min.
All 27 patients showed good responses to immunomodulatory therapies including steroids, intravenous immunoglobulin (IVIG), and/or immunosuppressants. Their average hospital days stay was 21 (8-47) days. Two cases were followed up on PSGs without relapses. One of the two patients had SD during the acute phase. Her repeated PSG recording revealed identified sleep architecture and stable REM sleep ( Figure 1B). She also had an increase of TST from 267 to 448.5 min, sleep efficiency from 49.8% to 56.9%, and REM sleep from 1.5 to 10.9%. An improvement in PSG recording was also observed in the other patient (TST from 430 to 462.5 min; sleep efficiency from 80.3% to 84.8%; and REM sleep from 13.6% to 15.0%).
Cerebral MRIs were available in six patients with normal finding. Of four patients who underwent PET/CTs, two patients had abnormal metabolic patterns in the basal ganglia and medial temporal lobe. Three patients had EEG recordings. One was normal, one revealed epileptiform discharge, one had mild excessive slow waves, and none had tumors.
All seven (100%) patients had insomnia. Five patients had new concerns of limb movements. Four of these patients had dream enactment behavior during the acute phase. Excessive daytime sleepiness was presented in three patients ( Table 1).
PSG recording demonstrated SD in two patients ( Figure 1C). The range of TST was from 167 to 354 min. The range of sleep efficiency was from 34.6 to 71.7%. One patient lacked typical REM sleep, whereas another patient had no N3 sleep. Both cases were diagnosed with Morvan's syndrome and had increased LMS (114-120.5/h). Agrypnia excitata was identified in one patient who had complex movements mimicking the contents of dreams and severe insomnia. PLMS were detected in the other patients. Actigraphies were performed in five patients. Four (80%) artigraphies displayed continuous insomnia with marked diurnal and nocturnal motor activities ( Figure 3B). None had a long time sleeping.
All seven patients with Caspr2-Ab diseases had improvements on immunotherapy. Their hospital days were 29 (27-45) days. One patient died of pneumonia during relapse. None had repeated PSG evaluations.

DISCUSSION
Sleep disorders are common in VGKC-Ab diseases presenting in 20-65% patients with LGI1-Ab encephalitis (21-24) and 22-68% patients with Caspr2-Ab-associated diseases (15,25). In our study, nearly half of the patients had sleep disturbances, which was consistent with previous studies. Different forms of sleep disorders were observed in our study including insomnia, hypersomnia, RBD, PLMS, agrypnia excitata, and OSAS.
PSGs in patients with LGI1-Ab encephalitis demonstrated a decrease in TST, sleep efficiency, and the percentages of REM and N3 sleep stages but a marked increase in LMS index. Consistent with previous reports (8,11,26), we found that RBD and PLMS were more common in LGI1-Ab encephalitis than in Caspr2-Ab diseases. LGI1-Ab encephalitis is characterized by limbic system involvements. Previous studies demonstrated that the limbic system is related to dreams during REM sleep and that its connections with the brainstem are responsible for REM sleep atonia. The dysfunction of the limbic system could account for RBD (26,27). All patients with LGI1-Ab encephalitis had limbic encephalitis.
We found PSG parameters did not change with frequencies of RWA/RBD, PLMS, FBDSs, or epileptiform discharges in patients with LGI1-Ab encephalitis. We suggest that the declines of TST, sleep efficiency, REM, and N3 sleep were caused by encephalitis but no other sleep disorders.
SD was detected in VGKC-Ab diseases and was characterized by sleep fragments and state-determining boundaries breakdown. PSG showed frequent shifts between wake, N1, and N2 sleep stages, and REMs intermixed with slow eye movements. Agrypnia excitata was identified in one patient with Morvan's syndrome. We also provided evidence for SD in LGI1-Ab encephalitis. Studies about SD in LGI1-Ab encephalitis were very limited. Only one case report described agrypnia excitata associated with LGI1-Ab encephalitis (28). We found that 23.8% of patients with LGI1-Ab encephalitis presented SD. These patients were more likely to have simple limb movements rather than complex movements. We also found decreased REM sleep, increased N1 sleep, and a trend of N3 sleep decline in SD patients. Other PSG data and clinical features did not change with SD.
Actigraphies results revealed continuous insomnia in VGKC-Ab-associated diseases. Animal experiments demonstrated that VGKC complexes are related to sleep regulation. VGKC dysfunction may be responsible for the increased motor drive and reduction in sleep time (29,30). We found that continuous insomnia was more frequent in patients with Caspr2-Ab-associated diseases than LGI1-Ab encephalitis, which suggested that Caspr2-Ab cause more severe sleep disturbances. Actigraphy is an alternative sleep assessment for patients who could not complete PSG owing to severe insomnia. However, actigraphy could not recognize the sleep stages or identify sleep structure.
Hypersomnia was not rare in VGKC-Ab-associated diseases. Long sleeping time was detected in actigraphy recordings. The basal ganglia play an important role in sleep-wake regulation. Bilateral lesions were made in the striatum, which resulted in a significant reduction in time spent in wakefulness (31). PET/CT revealed that basal ganglia were commonly involved in patients with VGKC-Ab diseases. Hypothalamus dysfunction, resulting in orexin-A reduction, could also be a reason for hypersomnia and other sleep disorders (32).
All patients showed improvements with immunotherapy. Follow-up PSG evaluations demonstrated an increase in TST and sleep efficiency and re-identified sleep structure. Actigraphy in a patient with Caspr2-antibody-associated diseases showed continuous insomnia. LGI1, leucine-rich glioma-inactivated protein 1.

CONCLUSIONS
We found that sleep disorders including insomnia, hypersomnia, RBD, PLMS, agrypnia excitata, and OSAS were commonly encountered in VGKC-Ab diseases. A decrease in TST and sleep efficiency and changes in the respective sleep stages were observed. Evidence of SD in patients with LGI1-Ab encephalitis showed that these patients are more likely to have simple limb movements. However, the small sample size of patients with LGI1-Ab encephalitis who had SD may have a bias on results of statistical analysis. Patients with Morvan's syndrome who could not complete PSG monitor due to severe insomnia were not included. There is a need for more works to better understand sleep disorders and SD in VGKC-Ab-associated diseases.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Ethics Committee of Peking Union Medical College Hospital. The patients/participants provided their written informed consent to participate in this study.