GTE score on evaluating the degree of neurological impairment in anti-NMDA receptor encephalitis

Abstract

This study mainly analyzed the EEG changes and the development of neurological impairment at different stages of the disease, and explored the clinical role of the Grand Total EEG score (GTE score) in evaluating the degree of neurological impairment function damage in anti-NMDA receptor encephalitis. The GTE scores of patients with anti-NMDA receptor encephalitis at different disease stages and the diagnostic validity of the GTE scores for evaluating the degree of neurological impairment were analyzed. 153 EEG cases were included in this study, with a course of 2 to 960 days. The abnormal EEG rates were 92% in the acute phase, 83% in the recovery phase, 76% in the short-term prognosis phase, and 30% in the long-term prognosis phase. The GTE score: acute phase=(5, 0–15); recovery phase=(2, 0–14); short-term prognosis phase= (2, 0–9); long-term prognosis phase = (0, 0–6). When the total GTE score is 5.5, the patient was more likely to have moderate to severe neurological impairment (mRS = 4), with sensitivity and specificity > 0.9. When the total GTE score was 7.5, the patient was more likely to have severe neurological impairment (mRS = 5), with sensitivity and specificity > 0.9. GTE score has a high diagnostic efficacy for the rating of neurological impairment. The total GTE score was superior to the assessment of neurological damage degree solely from background activity frequency or diffuse slow wave activity.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-97439-9.

Introduction

Autoimmune encephalitis (AE) are neurological diseases mediated by autoimmunity produced by the action of anti-neuronal antibodies^1,2. Scalp EEG has been widely used in the diagnosis and treatment process of anti-NMDAR encephalitis for its convenience, fast speed and non-invasive nature. However, at present, there are few large sample studies on anti-NMDAR encephalitis EEG in China, and there are few relevant studies on the related characteristics of EEG and the development process of neurological impairment in different disease stages of this disease^3,4. Semi-quantitative EEG GTE (Grand total EEG score) mainly from the background activity frequency, diffuse slow wave activity, background rhythm (activity) reactivity, seizure activity, focal abnormal and sharp wave activity six aspects objective and detailed evaluation EEG, can be used for epilepsy, Parkinson’s disease, cerebral infarction and other diseases assessment and differential diagnosis^8,24, but there is very few research about GTE and anti-NMDA receptor encephalitis. Analysis of neurological damage conditions and GTE scores at different disease stages were used to assess the diagnostic validity of the degree of neurological damage against NMDAR encephalitis. The objective of our study was to provide a more comprehensive understanding of the disease development characteristics and EEG manifestations of anti-NMDA receptor encephalitis, and provide certain reference value for the prognostic assessment of clinicians.

Materials and methods

Research subjects

Participants

Patients with anti-NMDA receptor encephalitis who visited the inpatient and outpatient departments of the Department of Neurology, Affiliated Brain Hospital of Guangzhou Medical University from January 2018 to December 2022 were included. All participants gave informed consent before being included in the study. This study has been approved by the Ethics Committee of the Affiliated Brain Hospital of Guangzhou Medical University.

Inclusion and exclusion criteria

1. Inclusion criteria²²: Patients were required to meet the following criteria: diagnosed with anti-NMDA receptor encephalitis (positive anti-NMDA receptor antibodies in CSF) and had at least one EEG examination during their visit. The EEG results of patients meeting the above criteria will be included in the study.

2. Exclusion criteria: (meet any one of the following: ① The medical history of neurological diseases, including cerebrovascular diseases, brain infections, brain injuries, and neurodegenerative diseases, etc. ② The clear history of mental illness. ③ No EEG examination during hospitalization.

Research methods

The GTE scores of patients with anti-NMDA receptor encephalitis at different disease stages were analyzed, the diagnostic validity of the GTE scores for evaluating the degree of neurological impairment in anti-NMDA receptor encephalitis was analyzed. The disease was divided into four stages according to the length of time from onset: acute stage (0–1 month), recovery stage (1–6 months), short-term prognosis stage (6–12 months) and long-term prognosis stage (more than 12 months). The disease stage grouping in this study was based on the progression time of clinical symptoms of anti-NMDA receptor encephalitis, which is supported by clinical experience and literature²³. The most common stage of the disease is the prodromal stage, which is characterized by non-specific symptoms such as fever, headache, dizziness, and fatigue^9–11. As the disease progresses, symptoms such as mental abnormalities, cognitive impairment, and epilepsy usually reach their peak in 2–4 weeks^12–14. Therefore, this article defines 0–1 month as the acute stage. One month after onset, most patients have been diagnosed and received treatment, and clinical improvement has gradually improved. However, due to the different severity of the disease, the recovery time varies. Generally speaking, clinical symptoms have improved significantly and the condition has stabilized six months after onset³². The above characteristics of disease stage division were also reflected in the study of long-term prognosis and recurrence of patients with anti-NMDA receptor encephalitis in western China. The study included 244 patients for long-term follow-up, with January, June, and December as important follow-up time points⁵.

All methods were performed in accordance with the relevant guidelines and regulations.

EEG data

All patients’ EEG data were recorded using a 16-channel digital EEG recording system (Nicolet, USA). Electrodes were placed according to the international 10–20 system. The band pass filter was 0.5–60 Hz. Impedance was maintained below 10 kΩ. Recordings were performed when the patient was quiet and relaxed, including evoked test methods such as eye opening/closing, intermittent light stimulation, and hyperventilation, unless clearly contraindicated. The sensitivity of the EEG recording was 10 µV, the time constant was 0.3 s, and the time window of 10 s per screen was visually assessed. Age-related physiological slow wave activity was not included in the background slowing grading assessment. The terminology and interpretation standards used in the EEG were in accordance with the terminology and definitions of the International Federation of Electroencephalography and Clinical Neurophysiology Societies Terminology Committee⁶. Each EEG was reviewed by two EEG experts who were with intermediate or higher qualifications independently. The unanimous result of the two experts is the final GTE score. If there is a significant discrepancy between their opinions, they will discuss together to reach a final GTE score. Intraclass Correlation Coefficient, ICC = 0.85.

The included EEGs were assessed using the GTE score^17,25. This scale comprehensively evaluates the patients’ EEG from six aspects ^29,30. The lowest possible score is 0 and the highest possible score is 30³¹. See supplementary file 1 for details.

Assessment of neurological impairment: modified Rankin rating scale (mRS scale)

The mRS scale is used to assess the degree of recovery of the patients’ neurological function. This rating scale has been widely used in the study of the prognosis of autoimmune encephalitis²⁸ and is divided into 5 grades. See supplementary file 2 for details.

Statistical methods

Normally distributed quantitative data were expressed as mean ± standard deviation, and non-normally distributed quantitative data were expressed as median (quartile) [M (Q1, Q3)]. The GTE scores of the four groups of patients at different disease stages were compared by analysis of variance. The ROC curve was used to analyze the diagnostic validity of the GTE score in evaluating the degree of neurological damage in anti-NMDA receptor encephalitis. The test level was α = 0.05. SPSS22.0 software was used for statistical analysis.

Result

Participants and EEG results

A total of 52 participants were included in the study, with 153 EEG cases, aged 10 to 55 years, with an average age of (30 ± 12.60) years. There were 12 children (10–17 years) (23%) and 40 adults (≥ 18 years) (77%). Among the patients, there were 18 males (35%) and 34 females (65%). The duration of illness ranged from 2 to 180 days, with a median (interquartile range) of 14 (7, 30) days. Four patients had tumor. One case was thyroid cancer, and three cases were ovarian teratomas. All patients received first-line immunotherapy (hormones and/or intravenous immunoglobulin infusion). Seven patients (14%) required second-line immunotherapy (cyclophosphamide, azathioprine, or other immunosuppressants) due to poor response to first-line immunotherapy. Among the 52 enrolled patients, 31 had normal MRI results (60%), while 21 had abnormal MRI findings (40%).

Including 52 cases (34%) in the acute phase (0 to 1 month), 42 cases (27%) in the recovery phase (1 to 6 months), 29 cases (19%) in the short-term prognosis phase (6 to 12 months), and 30 cases (20%) in the long-term prognosis phase (more than 12 months). The abnormal EEG rates were 92% in the acute phase, 83% in the recovery phase, 76% in the short-term prognosis phase, and 30% in the long-term prognosis phase. The GTE score: acute phase = (5, 0–15); recovery phase = (2, 0–14); short-term prognosis phase = (2, 0–9); long-term prognosis phase = (0, 0–6).

Diffuse slow wave activity, slowing of frequency of background activity, and decreased reactivity of rhythmic background activity were the most common abnormal EEG manifestations of anti-NMDA receptor encephalitis. The incidence of diffuse slow wave activity was 87% in the acute phase, 81% in the recovery phase, 76% in the late short-term prognosis phase, and 30% in the late long-term prognosis phase. The incidence of slowing of frequency of background activity was 73% in the acute phase, 38% in the recovery phase, 52% in the late short-term prognosis phase, and 17% in the late long-term prognosis phase. The incidence of decreased reactivity of rhythmic background activity was 58% in the acute phase, 33% in the recovery phase, 45% in the late short-term prognosis phase, and 7% in the late long-term prognosis phase. The incidence of focal slow wave activity was 2% in the acute phase, 7% in the recovery phase, and 0 in the late short-term prognosis phase and the late long-term prognosis phase. The incidence of sharp waves was 8% in the acute phase, 0 in the recovery phase, the late short-term prognosis phase, and the late long-term prognosis phase. The incidence of delta brush was 10% in the acute phase, 2% in the recovery phase, and 0 in the late short-term prognosis phase and the late long-term prognosis phase. The incidence of abnormal EEG, total GTE score and sub-item scores in anti-NMDA receptor encephalitis are shown in Tables 1 and 2.

Table 1.

The incidence of abnormal EEG in anti-NMDA receptor encephalitis at different disease stages.

The incidence of abnormal EEG in anti-NMDA receptor encephalitis
	Acute phase (n = 52), %	Recovery period (n = 42), %	Short-term prognosis (n = 29), %	Long-term prognosis (n = 30), %
Abnormal EEG	90	83	76	30
Diffuse slow wave activity	87	81	76	30
Slowing of frequency of background activity	73	38	52	17
Decreased reactivity of rhythmic background activity	58	33	45	7
Focal slow wave activity	2	7	0	0
Spike	10	0	0	0
Delta brush	10	2	0	0

Table 2.

Total score and sub-item scores of GTE in different disease stages of anti-NMDA receptor encephalitis.

Total GTE score and sub-item scores
	Acute phase (Median, 95% CI)	Recovery period (Median, 95% CI)	Short-term prognosis (Median, 95% CI)	Long-term prognosis (Median, 95% CI)
GTE total points	5, (0, 15)	2, (0, 14)	2, (0, 9)	0, (0, 6)
Diffuse slow wave activity	2, (0, 5)	0, (0, 4)	1, (0, 4)	0, (0, 2)
Frequency of background activity	2, (0, 5)	0, (0, 5)	1, (0, 4)	0, (0, 3)
Decreased reactivity of rhythmic background activity	1, (0, 5)	0, (0, 3)	0, (0, 1)	0, (0, 1)
Focal slow wave activity	0, (0, 3)	0, (0, 3)	0	0
Spike	0, (0, 3)	0	0	0

Neurological impairment results

In the acute phase, 98%(51/52) of patients had neurological impairment, including 42% (22/52) with mild neurological impairment, 23% (12/52) with moderate neurological impairment, 19% (10/52) with moderate-to-severe neurological impairment, and 14% (7/52) with severe neurological impairment. In the recovery phase, 88% (37/42) of patients had neurological impairment, including 55% (23/42) with mild neurological impairment, 14% (6/42) with moderate neurological impairment, and 19% (8/42) with moderate-to-severe neurological impairment. In the short-term prognosise phase, 66% (19/29) of patients had neurological impairment, including 52% (15/29) with mild neurological impairment, 10% (3/29) with moderate neurological impairment, and 3% (1/29) with moderate-to-severe neurological impairment. Among the patients in the stage of long-term prognosis, 40% (12/30) had neurological impairment, all of which were mild neurological impairment, and the remaining 60% had no residual neurological impairment. See Table 3 for details. The distribution of total GTE score and mRS score in different stages of anti-NMDA receptor encephalitis is shown in Fig. 1. Distribution of neurological impairment in patients with anti-NMDA receptor encephalitis is shown in Fig. 2.

Table 3.

Incidence of neurological impairment in anti-NMDA receptor encephalitis at different disease stages.

The incidence of neurological impairment
	Acute phase,% (Cases = 52)	Recovery period,% (Cases = 42)	Short-term prognosis, % (Cases = 29)	Long-term prognosis, % (Cases = 30)
No neurological impairment	2	12	35	60
Mild neurological impairment	42	55	52	40
Moderate neurological impairment	23	14	10	0
Moderate to severe neurological impairment	19	19	3	0
Severe neurological impairment	14	0	0	0

Fig. 1.

Distribution of total GTE score and mRS score in different disease stages of anti-NMDA receptor encephalitis.

Fig. 2.

Distribution of neurological impairment in patients with anti-NMDA receptor encephalitis.

Clinical role of GTE score in evaluating the degree of neurological damage in anti-NMDA receptor encephalitis

1. Moderate or above neurological impairment (mRS ≥  3).

The ROC curve area of the total GTE score was 0.90, 95%CI=(0.84, 0.95). The optimal cutoff point of 5.5 was selected when the Youden index was the largest, and the corresponding sensitivity was 0.66 and specificity was 0.62, PPV = 0.66, NPV = 0.96. The ROC curve area of the frequency of background activity was 0.85, 95%CI=(0.78, 0.92). The optimal cutoff point of 0.5 was selected when the Youden index was the largest, and the corresponding sensitivity was 0.87 and specificity was 0.71, PPV = 0.87, NPV = 0.76. The ROC curve area of diffuse slow wave activity was 0.85, 95%CI=(0.78, 0.92). The optimal cutoff point of 1.5 was selected when the Youden index was the largest, and the corresponding sensitivity was 0.87 and specificity was 0.83, PPV = 0.72, NPV = 0.83.

• (2)Moderate to severe or severe neurological impairment (mRS ≥  4).

The ROC curve area of the total GTE score was 0.97, 95%CI=(0.94, 0.99). The optimal cutoff point of 5.5 was selected when the Youden index was the largest, with a corresponding sensitivity of 0.92 and a specificity of 0.91, PPV = 0.92, NPV = 0.91. The ROC curve area of the frequency of background activity was 0.93, 95%CI=(0.87, 0.99). The optimal cutoff point of 3.5 was selected when the Youden index was the largest, with a corresponding sensitivity of 0.77 and a specificity of 0.95, PPV = 0.77, NPV = 0.95. The ROC curve area of the diffuse slow wave activity was 0.87, 95%CI=(0.80, 0.94). The optimal cutoff point of 1.5 was selected when the Youden index was the largest, with a corresponding sensitivity of 0.81 and a specificity of 0.76, PPV = 0.81, NPV = 0.99.

• (3)Severe neurological impairment (mRS = 5).

The ROC curve area of the total GTE score was 0.98, 95%CI=(0.96, 1). The optimal cutoff point of 7.5 was selected when the Youden index was the largest, with a corresponding sensitivity of 1 and a specificity of 0.91, PPV = 1, NPV = 0.91. The ROC curve area of the frequency of background activity was 0.93, 95%CI=(0.83, 1). The optimal cutoff point of 4.5 was selected when the Youden index was the largest, with a corresponding sensitivity of 0.86 and a specificity of 0.97, PPV = 0.86, NPV = 0.97. The ROC curve area of diffuse slow wave activity was 0.91, 95%CI = (0.78, 1). The optimal cutoff point of 3.5 was selected when the Youden index was the largest, with a corresponding sensitivity of 0.86 and a specificity of 0.93, PPV = 0.86, NPV = 0.92. The GTE EEG score is used to evaluate the degree of neurological damage in anti-NMDA receptor encephalitis. See Table 3 for details. The ROC curve of total GTE score, frequency of background activity and diffuse slow wave activity in evaluating the degree of neurological damage in anti-NMDA receptor encephalitis is shown in Fig. 3 for details. The effectiveness of GTE EEG score in evaluating the degree of neurological damage in anti-NMDA receptor encephalitis are showed in the Table 4.

Fig. 3.

ROC curve of total GTE score, frequency of background activity and diffuse slow wave activity in evaluating the degree of neurological damage in anti-NMDA receptor encephalitis.

Table 4.

The effectiveness of GTE EEG score in evaluating the degree of neurological damage in anti-NMDA receptor encephalitis.

	AUC	95% CI	Sensitivity	Specificity	Critical value
mRS ≥ 3
GTE Total Points	0.9	(0.84, 0.95)	0.66	0.62	5.5
Frequency of background activity	0.85	(0.78, 0.92)	0.87	0.71	0.5
Diffuse slow wave activity	0.85	(0.78, 0.92)	0.87	0.83	1.5
mRS ≥ 4
GTE Total Points	0.97	(0.94, 0.99)	0.92	0.91	5.5
Frequency of background activity	0.93	(0.87, 0.99)	0.77	0.95	3.5
Diffuse slow wave activity	0.87	(0.80, 0.94)	0.81	0.76	1.5
mRS = 5
GTE Total Points	0.98	(0.96, 1)	1	0.91	7.5
Frequency of background activity	0.93	(0.83, 1)	0.86	0.97	4.5
Diffuse slow wave activity	0.91	(0.78, 1)	0.86	0.93	3.5

Discussion

anti-NMDA receptor encephalitis is the most common type of autoimmune encephalitis. The disease has severe symptoms at its peak, is difficult to diagnose in the early stage, and has a high misdiagnosis rate. The diagnosis of anti-NMDA receptor encephalitis mainly relies on the detection of neuronal antibodies in cerebrospinal fluid. However, the feedback of antibody test results has a certain lag, and the antibody positive rate is related to the test process and detection method. All of these factors bring certain difficulties to the diagnosis of the disease⁷. EEG has been widely used in the auxiliary diagnosis, assessment of the disease and prediction of prognosis of this disease¹⁸. The EEG background rhythm can reflect the patients’ brain function status. The degree of EEG abnormality is related to the degree of brain function damage²¹. By analyzing the changes in EEG at different stages of the disease, we can understand the development process and prognosis of neurological function damage in anti-NMDA receptor encephalitis²⁷. The results showed that the EEG abnormality rate was as high as 92% in the acute phase, and gradually decreased in the recovery phase and the short-term prognosis phase, to 83% and 76%, respectively. The EEG abnormality rate in the long-term prognosis phase decreased significantly, 70% of EEGs could be completely restored to normal, and 30% were still abnormal. Diffuse slow wave activity, slowing of frequency of background activity, and decreased reactivity of background rhythm activity were the most common EEG abnormalities in anti-NMDA receptor encephalitis. Sharp waves were only seen in the acute phase. This reflects that seizures of anti-NMDA receptor encephalitis are related to the inflammatory state of the brain abnormal immune response at the peak of the disease. As the follow-up treatment and improvement of the disease, the abnormal immune response was controlled, and seizures were cured. Previous reports also had confirmed that seizures of most patients with anti-NMDA receptor encephalitis can be completely relieved²⁶. δ brushes occur in the acute phase and the recovery phase. Current studies suggest that δ brushes may be associated with the severity of the disease³³. Anti-NMDA receptor encephalitis may worsen in the early stages of the disease, some patients do not reach the peak severity at the onset of the disease. In the subsequent progression of the disease, the mRS score gradually increases. Most patients reach their peak within two weeks, while a few may reach their peak in 4 to 6 weeks. This explains why 10% of the patients in this study had δ brushes in the acute phase (0 to 1 month), while 2% were seen in the recovery phase (1 to 6 months). Focal slow wave activities are often associated with brain injury lesions. In the initial stage of the disease, antibodies attack the surface receptors of neurons, while the neurons are not significantly damaged. Therefore, in the early stage of the disease, focal injury lesions have not yet formed. Follow-up examinations show that as the abnormal inflammatory response aggravate, the neurological damages worsen, and then the focal lesions appear. MRI shows that the lesions are widely distributed and can occur in a wide range of gray matter and white matter areas including the frontal, temporal, parietal, occipital lobes and hippocampus, mainly in the non-limbic lobe²⁰; DWI is mostly isointense and diffusion is not restricted; after enhancement, it shows meningeal enhancement, thickening of blood vessels in the lesion area, and patchy enhancement in the lesion¹⁹. However, the MRI positive rate is lower than and later than EEG. In this article, the incidence of focal slow wave activity is 2% in the acute phase and 7% in the recovery phase, which further verifies the above brain injury process.

The results showed that severe neurological function damage was mainly concentrated in the acute stage (13%); in the recovery stage severe patients decreased significantly, but neurological function damage still accounted for a high proportion (88%); in the short-term prognosis period further decreased to 68%, and the incidence of moderate and moderate-to-severe neurological function damage further decreased; in the long-term prognosis period, 40% of patients still had residual neurological function damage, but damage were mild, with little impact on daily life, which was a good prognosis. Previous studies also had similar results⁵, the short-term prognosis of this disease was poor (< 3 months) and the long-term prognosis was good (> 12 months), but there were still about 20–30% of patients with poor prognosis, of which the mortality rate is about 5–10%. Poor prognostic factors include: prolonged endotracheal intubation, untimely treatment, admission to the ICU, and no tumor^15,16.

The GTE score comprehensively covers important EEG scoring points and is more accurate and detailed than the traditional grading assessment results. This article found that frequency of background activity and diffuse slow wave activity played important roles in the evaluation of the degree of neurological damage, but the total GTE score was better than based solely on the frequency of background activity or diffuse slow wave activity in the evaluation, as shown in Fig. 3. When the total GTE score cutoff value was 5.5, the patient was more likely to have moderate to severe neurological damage (mRS ≥ 4), and the sensitivity and specificity were both higher than 0.9. When the total GTE score cutoff value was 7.5, the patient was more likely to have severe neurological damage (mRS = 5), and the sensitivity and specificity were both higher than 0.9. When the frequency of background activity cutoff value was 4.5, the patient was more likely to have severe neurological damage (mRS = 5), and the sensitivity and specificity are both higher than 0.85. When the diffuse slow wave activity cutoff value was 3.5, the patient was more likely to have severe neurological damage (mRS = 5), and the sensitivity and specificity are both higher than 0.85.

The GTE score also has certain limitations. For example, in young children, the EEG background activity typically consists of more slow waves, which may result in a higher calculated GTE score. Additionally, medications that affect EEG, such as sedatives, may also impact the score.

In summary, EEG has different manifestations in different stages of anti-NMDA receptor encephalitis, and EEG can dynamically reflect the degree of neurological damage and indicate the disease process. In clinical practice, many patients are difficult to evaluate with neurological function scales due to impaired consciousness and abnormal mental symptoms. EEG examinations can provide timely feedback, at the meantime the GTE score have a high diagnostic efficacy in evaluating the degree of neurological damage.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Xiaoying Pan and Jiayi Huang wrote the main manuscript text.Guohua Chen analyzed data and prepared figures and tables. Yuefeng Zhang supportive work. All authors reviewed the manuscript.

Funding

Guangzhou Municipal Key Discipline in Medicine (2025–2027).

Data availability

The datasets generated and analysed during the current study are not publicly available due to the follow-up study but are available from the corresponding author on reasonable request.

Competing interests

The authors declare no competing interests.