Delta brush variant: A novel ictal EEG pattern in anti-NMDAR encephalitis

1 INTRODUCTION

Anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis is the most common form of autoimmune encephalitis due to specific antibody crosslinking the NMDAR, causing receptor capping and internalization.¹ The major symptoms of this disease include abnormal behavior, speech dysfunction, seizure, movement disorder, decreased consciousness level, and autonomic dysfunction.² Diagnosis can be made based on the presence of proper symptoms and IgG anti-GluN1 antibodies in cerebrospinal fluid (CSF).³ Extreme delta brush (EDB) and generalized rhythmic delta activity (RDA) are two frequent and prominent EEG patterns of anti-NMDAR encephalitis.⁴ Although not specific, occasional neuroimages described related lesions in hippocampus, basal ganglia, and white matter.² More than 80% patients with anti-NMDAR encephalitis have good outcome after receiving standardized immunotherapy.⁵

Seizure is common with prevalence of 70%-90% in all anti-NMDAR cases.^{6, 7} Hitherto, studies on electroencephalogram (EEG) of anti-NMDAR encephalitis showed that the most common epileptic discharge is regional sharp wave (71.6%), followed by periodic lateralized epileptiform discharge (19.4%), and generalized periodic epileptiform discharge(9.0%). Only 8% of the total cases were reported to have ictal patterns; among which, regional rhythmic alpha activity showed significant correlation with seizure occurrence in anti-NMDAR encephalitis.^{8, 9} EEG plays a crucial role in distinguishing seizure and paroxysmal non-epileptic events; however, the data of related ictal patterns remain limited.

Extreme delta brush is a unique EEG pattern in anti-NMDA receptor encephalitis and is characterized as a RDA at 1-3 Hz with superimposed bursts of rhythmic 20-30 Hz beta frequency activity.¹⁰ This pattern was once considered to be an ictal pattern, because the patients were more likely to undergo prolonged coma.¹¹ However, given that rare cases with EDB respond to sedative drugs electrographically, the clinical nature of this pattern remains controversial. In the present study, we introduced a unique ictal pattern evolving from EDB and discussed its significance in anti-NMDAR encephalitis.

2 SUBJECTS AND METHODS

3 RESULTS

The initial dataset of anti-NMDAR encephalitis contained 72 cases. Ictal patterns presented in 6 (8.3%) cases including four females and two males with mean age 21.3 years (range 17–30 years). DBV was performed in the following two cases.

3.1 Case 1

A 17-year-old girl experiencing progressive psychosis, agitation, and oral-facial dyskinesia for 4 days was admitted. The initial EEG results showed typical generalized EDB. Diffuse low-density lesion involving the white matter of the bilateral temporal and occipital lobes was noted by computerized tomography (CT) scan. She was diagnosed as anti-NMDAR encephalitis after NMDAR antibody was detected in her serum and CSF. Glucocorticoid with a dose of 500 mg/day was given but symptoms remained. Over the next 3 weeks, the patient developed hyperpyrexia, severe pneumonia, and pyothorax. On day 23, she began to have rhythmical jerking of bilateral limbs with cessation of choreic movement and oral-facial dyskinesia. Corresponding EEG patterns showed generalized synchronous delta rhythm with fast spike activity riding on it (Figure 1A-C). Seizure terminated after intravenous injection of midazolam at a dose of 0.2 mg/kg, but oral-facial dyskinesia reappeared. Midazolam at a dose of 0.2 mg/kg/h was continuously infused and terminated 2 days later. Subsequent immunotherapy (plasma exchange, intravenous immunoglobulin, and cyclophosphamide) was given routinely, and she recovered with a mRS score of 1 after 3 months. No seizure was reported after withdrawing AEDs (Table S1).

4 DISCUSSION

In general, we introduced a novel ictal EEG pattern in anti-NMDAR encephalitis. The presented DBV was characterized as generalized delta activity with superimposing fast spike activity. This pattern showed predominance in the florid phase of the disease. Patients with continuous EDB and prolonged hyperpyrexia tend to develop DBV. Patients with DBV were likely to spend more time in ICU and hospital, but the long-term outcome and seizure relapse rate were similar to those in patients without it.

To our knowledge, this pattern has not been described in anti-NMDAR encephalitis. As movement disorder is more common in anti-NMDAR encephalitis, ictal patterns should be carefully compared with electromyographic (EMG) signal to avoid misleading results. In general, the potentials generated in the muscles are of shorter duration than those generated in the brain. The frequency components are usually beyond 30-50 Hz.¹³ Although DBV seems similar to an EMG artifact at first look, several considerations support the epileptic nature of this pattern. First, the proposed fast component in DBV is a generalized pattern, while EMG always shows frontal or temporal predominance. Second, the burst of fast component is periodic, while EMG is always arrhythmic. Third, the fast component is locked to unequivocal seizures, while EMG is not. Finally, the movement disorder-induced artifact rarely responds to sedative, while DBV weans after intravenous injecting of midazolam.

EDB is one of the key features in anti-NMDAR encephalitis and regarded as a potential marker for the disease with a prevalence of 30% in total patients.⁴ Thus far, the ictal nature of EDB remains unclear partly because of the semiological mismatching. Previous studies revealed that EDB indicates forthcoming electroclinical seizures,¹¹ suggesting that it is an preictal state of status epilepticus in NMDAR encephalitis. By contrast, the inertness to benzodiazepines and AEDs shows that EDB itself is prone to an interictal pattern, which is related to the florid phase of encephalopathy.¹⁰ We named this pattern DBV because of its resemblance to the well-known EDB. Our DBV typically appears as a combination of clusters of fast spike activity and delta wave. Compared with EDB, the fast component of DBV shows a higher frequency and amplitude. Furthermore, the fast spike activity tends to alleviate after midazolam injection, leading DBV to be more epileptic than EDB.¹⁴ Moreover, the ictal nature of DBV was also supported by EEG terminology criterion proposed by American Clinical Neurophysiology Society.¹⁵ According to that criterion, DBV can be standardized as rhythmic delta activity superimposing spikes (RDA + S). It is noted that additional spikes which renders the delta activity make the EEG pattern more ictal-appearing than the usual term without the plus.

In fact, determining an ictal discharge is one of the most challenging issue for anti-NMDAR encephalitis. Most paroxysmal attacks of psychomotor agitation need to be recorded by EEG to avoid mistakenly diagnosing as seizures, and most of paroxysmal rhythmic activities in anti-NMDAR encephalitis are non-epileptic.¹⁶ As a result, recognition of an ictal pattern is important for clinical practice. Several recent studies have demonstrated that persistent rhythmic delta and alpha-like activities are predictors of ictal events,^{17, 18} and our study extends this area by introducing DBV. Compared with these patterns which corresponds to non-convulsive status epilepticus (NCSE), DBV is more likely to be accompanied with motor seizures. Moreover, both cases of DBV were recorded during the florid phase, which was quite different to those are common at the initial stage of anti-NMDAR encephalitis. Finally, previous ictal patterns show close relationship to teratoma,^{17, 19} but no teratoma was detected in the two cases with DBV.

The semblable morphology of DBV and EDB indicated that dysfunction of NMDAR might lead to various EEG phenotypes. In a retrospective study, Jeannin⁴ grouped the EEG abnormities in anti-NMDAR encephalitis into excessive beta activity (EBA), generalized RDA, and EDB. They found that generalized RDA was associated with abnormal movement, while EBA and EDB which contain excessive fast activity were more likely to be accompanied with seizure, psychiatrics symptoms, anxiety, and cognitive impairment. Thus, they hypothesized that the presence of EBA and EDB was related to cortical dysfunction, and generalized RDA was presented when subcortical area involved. Our findings reinforced Jeannin's hypothesis by providing the evolution of the fast component during interictal-ictal phase. Moreover, we also hypothesize that EDB is a superposition that contains signal from the cortex to form fast component (FC) and signal from subcortical area to form the delta component (DC). FC was sensitive to hyperpyrexia and benzodiazepines, and contributed to seizure, while DC showed an opposite feature. Alternation in cortical and subcortical involvement, which resulted in the composite changes in FC and DC as observed in scalp EEG, might be one of the primary causes of clinical evolution in anti-NMDAR encephalitis (Figure 2).

Another finding of this study was that the presence of DBV had a similar mRS score and seizure relapse rate during a 3-month follow-up. Our observations, combined with those of other investigators, suggest that the relationship between EEG features and outcomes is debated.²⁰ Our results supported that sedative is necessary to terminate DBV-induced seizure, but the significance of immunotherapy cannot be ignored. One retrospective study which focused on the outcome of AEDs alone in controlling the seizure of patients with autoimmune encephalitis demonstrated that only 23.5% of patients became seizure-free compared with 61.5% of patients with immunotherapy,²¹ indicating that the earlier and more intense immune treatment is undoubtedly more crucial for a better outcome. Moreover, the debate on the relationship between outcome and presence of DBV is still open due to the 3-month follow-up. Studies based on subtle clinical assessment and longer follow-up are needed to draw the final conclusion.

The main purpose of the study is to introduce a unique ictal EEG pattern to avoid misdiagnosis. However, several limitations should be noted in the interpretation of DBV. First, the relatively small sample makes it difficult to determine the overall features of DBV. This pattern tends to be accompanied with continuous EDB and prolonged hyperpyrexia. The potential necessary combination of these two factors may explain why such a pattern was not reported previously. As DBV is recorded in the acute phase of anti-NMDAR encephalitis when choreic movement is more frequent than seizure, more EEG-based studies are needed to analyze the incidence and clinical nature of DBV. Second, only 8.3% (6/72) anti-NMDAR encephalitis patients had recorded seizures in our study, which was relatively lower than other reports.⁶ This discrimination may partly result from the duration of continuous EEG monitoring. It is reported that comatose patients are more likely to have their first seizure recorded after >24 hours of monitoring.²² Thus, the presence of DBV remains unclear and needs verifying by extended continuous EEG monitoring. Third, based on the morphological and clinical relationship between DBV and EDB, we hypothesize that EDB is a kind of complex that reflects dysfunction from cortex and subcortical area, but the speculative nature of our observation still need further experiments to verify.

CONFLICTS OF INTEREST

The authors declared no conflicts of interest with respect to the research, authorship, and publication of this article. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.