Clinical Features, Immunotherapy, and Outcomes of Anti-Leucine-Rich Glioma-Inactivated-1 Encephalitis

Antileucine-rich glioma-inactivated-1 (anti-LGI1) encephalitis is a newly discovered subtype of autoimmune limbic encephalitis. The specific antibody against the LGI1 subunit of the voltage-gated potassium channels (VGKC) complex was identified in 2010 and has been associated with limbic encephalitis (1). LGI1 is a neuronal glycoprotein that is mainly expressed in the hippocampus dentate gyrus, the CA3 region, the lateral temporal cortex, and mossy fibers. As a component of VGKC complexes, LGI1 acts as a transsynaptic link between presynaptic ADAM23 and postsynaptic ADAM22, thereby enabling synaptic transmissions of neuronal signals (2, 3). Anti-LGI1 antibodies can prevent synaptic transmissions involved in the processes of learning, behavior, and memory.

Anti-LGI1 encephalitis has a specific antibody against LGI1 and frequently present medial temporal lobe inflammation (1, 4), resulting in limbic encephalitis, psychiatric symptoms, faciobrachial dystonic seizures (FBDS), and seizures (5, 6). Non-limbic encephalitis patients with anti-LGI1 encephalitis have been shown to present clinical manifestations, such as chorea and ataxia, outside the limbic system (5). Anti-LGI1 encephalitis is a less common autoimmune disease that is rarely associated with tumors. Anti-LGI1 encephalitis patients exhibited good clinical responses to immunotherapy but were more likely to relapse (5, 7, 8). These characteristics may differ from other typical immune-associated limbic encephalitis (5, 9).

The number of reported anti-LGI1 encephalitis patients in various studies is no more than 300, with most of them from Europe and the United States. The clinical course of anti-LGI1 encephalitis is complex, and its diagnosis is challenging. Additionally, therapeutic options have not been standardized, and various outcomes of different treatments have not been fully evaluated. The limited data from a small sample of Chinese patients do not fully elucidate the characteristics and therapeutic outcomes of anti-LGI1 encephalitis (6, 10, 11). This study was aimed to evaluate the clinical characteristics and prognosis of 43 Chinese patients with anti-LGI1 encephalitis.

Methods

Ethical approval for this study was obtained from the Research Ethics Committee of Xuanwu Hospital of Capital Medical University. Patients who agreed to sign an informed consent form were included. Analysis of study data were anonymized to remove patients’ identifying information.

Study Participants

We recruited patients with clinically confirmed LGI1-antibody encephalitis who had been admitted at the Xuanwu Hospital of Capital Medical University. Recruitment was performed from January 1, 2012, to December 31, 2016. Clinical inclusion criteria were as follows: patients who met the anti-LGI1 encephalitis diagnostic criteria (4), which included subacute onset (rapid progression <3 months) of cognitive impairment, seizures, altered mental status, and psychiatric symptoms; reasonable exclusion of alternate disorders; and patients who were positive for LGI1 antibody in serum or in cerebral spinal fluid (CSF).

Data Collection

The following data were recorded: demographic data (gender and age), initial symptoms, clinical symptoms during the disease course, laboratory examination findings, CSF findings, LGI1 antibody levels in serum and CSF, cranial MRI findings, electroencephalogram (EEG) findings, immunotherapeutic methods, and functional status (evaluated by the Modified Rankin Scale [mRS]).

Treatment and Prognostic Evaluation

All patients received symptomatic supportive treatment, immunotherapy, and systemic neoplasm screening. First-line immunotherapy included monotherapy or a combination of intravenous pulse glucocorticoid therapy (methylprednisolone, 1,000 mg, or 500 mg, for 3 or 5 days and tapering oral), intravenous gamma immunoglobulin (IVIg; 0.4 g/kg/day for 5 days), or plasma exchange (3–5 times) (3, 9, 12). Immunosuppressants such as rituximab, cyclophosphamide, azathioprine, or mycophenolate mofetil were used as second-line immunotherapeutic options (7, 10).

Patient follow-up started at 3 months after initial immunotherapy and continued during clinical visits or by telephone. The trained telephone interviewers were blinded to the clinical conditions. Functional status was evaluated using the mRS. On the basis of mRS findings at 6 months, patients were categorized into two groups: functionally independent and disabled. Patients with an mRS score of 0 were considered to have had a complete recovery. Good and poor outcomes were defined as an mRS score of 0–1 and 2–6, respectively, at patients’ last follow-up assessments. Relapse was defined as the appearance of a new symptom or worsening of a preexisting symptom after improvement or stabilization of the disease for at least 1 month. Unfavorable prognostic factors were also evaluated.

Statistical Approach

Categorical variables are expressed as counts (percentages). Nonnormal quantitative variables are expressed using their medians and interquartile ranges (IQRs), while normal quantitative variables are presented as means (and standard deviations). Patients’ baseline variables of demographic and clinical characteristics were analyzed using descriptive statistics and were compared between the functionally independent (mRS score 0–1) and disabled (mRS score 2–6) groups using Student’s t test or Wilcoxon rank sum test (W) for continuous variables and chi-square test for categorical variables. Univariate logistic regression analysis was performed to establish the variables that accounted for disabled outcomes. After adjusting for confounding factors with a p value <0.1 and factors that were clinically considered to be closely related to outcomes, the multivariate logistic regression model was used to assess any independent factors for the disabled outcome. A p value ≤0.05 was considered statistically significant. All statistical analyses were performed using SPSS, version 19.0 (SPSS, Chicago).

Results

Clinical Features

Clinical characteristics of the 43 patients (males, N=28 [65.1%]; females, N=15 [34.9%]) are summarized in Table 1. The median age at onset of anti-LGI1 encephalitis was 57 years (IQR: 44–65 years). The median duration of illness was 60 days (range: 37–127 days).

Seven of the 43 patients presented prodromal symptoms, such as fever, headache, dizziness, and stomach discomforts. Initial presentations (Figure 1) were epileptic seizure (N=18, 41.9%), memory impairments (N=13, 30.2%), and neuropsychiatric symptoms (N=10, 23.3%). During the disease course, the most common symptoms among the 43 patients were epileptic seizures, in patients with status epilepticus (N=10, 23.3%), FBDS (N=19, 44.2%), and cognitive dysfunction (N=41, 95.3%), mainly in memory deficits and deterioration of orientation. In addition, 33 patients (76.7%) exhibited neuropsychiatric disturbances, including behavioral dyscontrol (N=18, 54.5%), visual or auditory hallucinations (N=17, 51.5%), nonsensical speech (N=16, 48.5%), personality changes (N=8, 24.4%), anxiety (N=5, 15.2%), and irritability (N=6, 18.2%). Other symptoms in the 43 patients included sleep disturbances (N=25, 58.1%), altered level of consciousness (N=21, 48.8%), movement abnormalities (N=16, 37.2%), and autonomic dysfunction (N=14, 32.6%).

Auxiliary Investigations

Laboratory and imaging results are summarized in Table 1. A total of 22 patients (51.2%) had hyponatremia in the acute phase. Increased puncture pressure was observed in eight patients (18.6%). CSF pleocytosis (>8 cells/μL) was reported for seven patients (16.3%), and CSF protein level was elevated in nine (20.9%, mean=36.5 mg/dl [SD=21.3]). Glucose and chloride levels in the CSF were all normal. Other autoantibodies and classical paraneoplastic antibodies in serum and CSF were not detected in any patient. In addition, standard pelvic ultrasound and chest computerized tomography were normal. All patients tested positive for LGI1 antibodies in serum or CSF samples. LGI1 antibodies were retested in 22 patients (51.2%); only four patients (18.2%) presented positive CSF samples during follow-up.

Abnormal EEG results were found in 31 patients (72.1%). The most common EEG abnormalities were focal slow activity (N=10, 32.3%), diffuse slowing (N=8, 25.8%), and epileptiform discharges (N=13, 41.9%).

Among the total sample, 30 patients (69.8%) exhibited increased unilateral or bilateral signals on MRI fluid-attenuated inversion recovery or T₂ that were thought to be related to encephalitis at admission. MRI detection of one representative patient is presented in Figure 2. Among these 30 patients, involvement of the limbic system, including the hippocampus (N=23, 76.7%), temporal lobe (N=12, 53.5%), and insular cortex (N=3, 10.0%), was observed in 29 patients. Other locations involved white matter (N=8, 26.7%), basal ganglia (N=6, 20.0%), and other lesions (N=9, 30.0%) in the frontal lobe, parietal lobe, and cerebellum. Follow-up MRI scans were performed for 34 patients (79.1%); among these, hippocampal atrophy was shown in 20 (58.8%). Eight of 10 patients showed abnormal local hypoperfusion when compared with the contralateral on single-photon emission tomography (SPECT).

Treatment

Patients with seizures and FBDS were treated with single or combined oral antiepileptic drugs. All patients were treated with immunotherapy consisting of corticosteroids and/or IVIg (Table 1). All patients were initially treated with corticosteroids; 21 (48.8%) also received IVIg (0.4 g/kg daily for 5 days), whereas one patient received only IVIg. Due to relapse, four patients (9.5%) received mycophenolate mofetil. The median time from onset to treatment was 64 days (IQR: 43.0–127.0 days). All patients exhibited good clinical responses to immunotherapy as confirmed by symptomatic improvement (reduced seizures or cognitive improvement). Patients with frequent FBDS and presenting with seizures responded poorly to antiepileptic drugs and improved upon immunotherapy. Ten patients (23.3%) relapsed during corticosteroid reduction. Six of these 10 patients received an initial oral dose of corticosteroids, whereas four patients were treated with mycophenolate mofetil.

Outcomes and Analysis of Factors

The median length of hospital stay was 16 days (IQR: 13–22). The median follow-up period was 21.5 months (IQR: 7–43). The median mRS score was 3 (IQR: 3.3–4) at admission and 1 (IQR: 0–1) at the last follow-up. Forty-one patients (95.3%) exhibited good clinical outcomes; only two patients (4.7%) exhibited poor clinical outcomes at the last follow-up (Figure 3). Four patients (9.3%) relapsed during follow-up, and one patient died of pulmonary infection during subsequent treatment in the local hospital 1 month after discharge. Fifteen patients were reported to have completely recovered, and 27 exhibited mild residual symptoms. Residual symptoms included memory deficits (N=20), difficulties with calculation (N=2), mild disturbances of emotional regulation (N=3), sleep disturbances (N=3), asthenia (N=2), and mild partial motor occasional seizures (N=3) (Figure 4).

Table 2 shows the association between individual clinical, laboratory, and imaging factors and neurologic function at 6 months. Serum sodium level (p=0.018), abnormal EEG findings (p=0.026), and mRS score on admission (p=0.006) were all associated with poor functional status at 6 months. We performed regression analysis to identify independent predictors of functional status at 6 months. Univariate analysis results are summarized in Table 3. Serum sodium was associated with disabled status (odds ratio=0.81, 95% CI=0.66, 0.98, p<0.05). Abnormal EEG findings (odds ratio=0.11, 95% CI=0.01, 0.10, p<0.1) and mRS score at admission (odds ratio=5.40, 95% CI=1.00, 28.97, p<0.10) were also associated with disabled status. Other factors, such as age, clinical symptoms, time to treatment, CSF findings, MRI, and type of treatment were not significantly associated with prognosis.

Discussion

This study evaluated the clinical characteristics and therapeutic outcomes of 43 patients with anti-LGI1 encephalitis. Some of the characteristics of these patients are consistent with those of patients in previously reported studies. For example, seizure and FBDS were the prominent clinical presentation in anti-LGI1 encephalitis patients, followed by cognitive and neuropsychiatric disturbances. This disorder, for the most part, was not a paraneoplastic syndrome. Temporal and/or hippocampal atrophy developed during the later stages of this condition. Seizures responded well to immunotherapy, and cognition gradually improved. However, our findings differ from previous reports in several aspects. First, the median age at onset was 57 years (IQR: 44–65), with a male preponderance (65.1%). This stands in contrast to other autoimmune encephalidities, particularly anti-N-methyl-d-aspartate encephalitis, which has a median age of about 21 years (range of 8 months to 85 years) and a female predominance of 2:1 (4, 13). Second, hyponatremia is a common presentation of anti-LGI1 encephalitis, which is not typical of other autoimmune encephalidities. Third, most patients had more favorable outcomes and lower recurrence and mortality rates than those with anti-LGI1 encephalitis reported elsewhere (5).

The initial and most common presenting symptom in this cohort was frequent multifocal seizures; this symptom was resistant to antiepileptic drugs but responded well to immunotherapy (5, 6, 13). FBDS was common among our patients. Anti-LGI1 encephalitis is associated with FBDS, with incidence of FBDS at approximately 20%−40% (5, 7). FBDS, which is seizure-like episodes, usually presents with brief stereotyped ipsilateral face grimacing and posturing of the upper limb, which may be confounded with myoclonus (14). In 40%−71% of patients, it presents early, before other symptoms (6, 7). Studies have reported on the relationship between anti-LGI1 limbic encephalitis, seizure, and FBDS. However, it has not been established whether FBDS is a seizure or simply an extrapyramidal manifestation (6, 7, 13). Some MRI studies revealed that FBDS duration is associated with reduced pallidum volume (15, 16). Because brain positron emission tomography (17) or SPECT revealed hypermetabolism in the basal ganglia, it is possible that cortical-striatal networks are hyperactive, suggesting a possible subcortical origin for the seizures (6, 13, 17). These observations suggest a role of basal ganglia in FBDS pathophysiology. In addition, magnetoencephalography findings from anti-LGI1 encephalitis patients with FBDS showed that epileptic discharge from the temporal lobe could be transmitted to the insular cortex, thereby causing FBDS (6, 14, 17). Therefore, FBDS might be a unique type of insular seizure between myoclonus and spasm. Further research should aim at elucidating on the mechanisms involved.

There were abnormal EEG findings in 96.8% of the patients in our study. The majority of EEGs showed slow or epileptic activity that is nonspecific in identifying FBDS or establishing disease outcomes. Because of the absence of synchronized EEG abnormalities, it was difficult to recognize these brief seizures. Therefore, there is a need to increase awareness of the multiple semiology of FBDS and follow-up EEG, which may be an early clue for the diagnosis of anti-LGI1 encephalitis.

Most of our patients exhibited rapid progressive memory deficits, which affect recent memory and disorientation. In anti-LGI1 encephalitis associated with cognitive impairment, the hippocampus, temporal lobe, and basal ganglia are the major pathogenetic targets (16–19). In this study, unilateral or bilateral hippocampus or basal ganglia hyperintense signals were found in most patients. Moreover, nearly 57% of the patients were found to have developed hippocampal atrophy or sclerosis during follow-up, in tandem with previous studies (15, 16). These findings imply that hippocampal inflammation occurs in the acute phase but is not always detected by routine MRI. Hippocampal atrophy is caused by seizures or inflammation (15, 16, 18). The abnormal local hypoperfusion revealed by SPECT suggests that cortical hypometabolism is associated with synaptic dysfunction as well as hippocampal temporal atrophy and might be an important characteristic of the inflammatory process and temporal damage (17–19). There is an association between hippocampal atrophy and cognitive dysfunction in anti-LGI1 encephalitis.

In this study, however, there was no correlation between hippocampal atrophy and clinical outcomes. Instead, the extent of CNS involvement may have influenced cognitive outcomes for patients without temporal hypermetabolism and those with unilateral involvements only (16, 19). On the basis of these preliminary findings, our patients shared similar features. Patients without hypermetabolism were more likely to achieve favorable outcomes than those with mild cognitive sequelae with inflammation and hippocampal damage. Further research is needed to determine the bases for hippocampal atrophy and cognitive impairment among persons with anti-LGI1 encephalitis.

Many (76.5%) of our patients exhibited neuropsychiatric disturbances, a substantial proportion of which resolved with immunotherapy. Neuropsychiatric symptoms that occurred in the early phase led many patients to present for initial evaluation and treatment in psychiatric hospitals, which highlights the importance of early recognition of this disease.

The median time from onset to diagnosis was 2 months, suggesting that early clinical diagnosis remains a challenge in clinical practice. Early diagnosis is often delayed by atypical clinical manifestations, normal early MRI findings, and the lack of evidence of inflammatory changes in the CSF. For early diagnosis and effective immunotherapy, follow-up of suspected encephalitis cases is essential. Abnormalities in temporal lobe MRI scans suggest autoimmune encephalitis, and testing for LGI1 antibodies should be initiated as soon as there is reasonable suspicion of this condition.

Hyponatremia also was a common feature of anti-LGI1 encephalitis, similar to prior reports of this problem among persons with VGKC antibody-associated limbic encephalitis (20). Serum sodium levels were decreased in 51% of patients in the present cohort. Serum sodium levels reflect severity of disease and, accordingly, may be associated with clinical outcomes. After sodium supplementation and immunotherapy, serum sodium levels gradually returned to normal levels. However, hyponatremia is not specific, and its exact mechanisms have not been elucidated. LGI1 is highly expressed in the hypothalamus and specific renal tubules of the kidney. Hyponatremia might result from abnormal secretions of antidiuretic hormones because of hypothalamic-pituitary neuraxis inflammation (16, 21) and the antibody-mediated effect on renal tubules (22).

Although limbic encephalitis may occur as a paraneoplastic complication of small cell lung cancer, thymoma, and testicular cancer (9, 21), no evidence of tumor was found in any patient in the present cohort of patients with anti-LGI1 encephalitis. While anti-LGI1 encephalitis may be associated with tumors (23, 24), the present findings suggest that the strength of this association may be lower than with other forms of limbic encephalitis.

Optimal therapeutic options for anti-LGI1 encephalitis have not been established, and early detection and aggressive immunomodulatory therapy are needed to improve prognosis. Appropriate immunotherapy can reverse the course of this disease; in particular, it may eliminate otherwise refractory seizures (especially for FBDS and those related to electrolyte disturbances), prevent progressive brain atrophy, and mitigate cognitive impairment (24). However, there is at present no widely accepted guideline for the immunotherapy of anti-LGI1 encephalitis. Based on cohort studies and expert opinion, it is suggested that corticosteroids, IVIg, or plasmapheresis be considered as first-line therapeutic options; rituximab and cyclophosphamide be considered as second-line therapeutic options in refractory cases (5, 8).

Most (94%) of the patients in this study responded well to steroids or IVIg treatment. Few developed persistent residual neurological impairments, neuropsychiatric disturbances, or memory impairments, consistent with previous reports (24–26). Although a minority (approximately 15%) of patients with anti-LGI1 encephalitis may develop persistent cognitive impairments (26), the persistence and severity of such impairments appears to vary with the timing of immunotherapy (earlier treatment favors better cognitive outcomes). However, the timing of treatment initiation did not predict outcomes in the present cohort, suggesting that this association between treatment timing and clinical outcomes requires further investigation.

Symptomatic relapse was not common in the present cohort. Most of the patients gradually improved when administered the initial dose of oral corticosteroid or mycophenolate mofetil as secondary treatment. However, definitive evidence regarding effective treatment and maintenance of remission has not been reported and require further investigation.

Despite the limitations associated with retrospective and observational studies that are based on medical records in a single center, the characteristics of the present cohort provide useful insights into the clinical features, course, and outcomes of persons with anti-LGI1 encephalitis. However, the relatively small sample size, single center, and short-term follow-up suggest that clinicians should remain circumspect about the relative frequencies of these aspects of anti-LGI1 encephalitis and that further study of this condition is needed to better define its clinical features and natural history.

Conclusions

Anti-LGI1 encephalitis is a subtype of autoimmune limbic encephalitis with common clinical features such as seizures, FBDS, cognitive impairments, neuropsychiatric disturbances, and hyponatremia. Most cases of anti-LGI1 encephalitis are not paraneoplastic, have low recurrence rates, and are associated with relatively favorable long-term outcomes when early, aggressive, and sustained combination immunotherapies are provided. Accordingly, expeditious identification and treatment along with long-term follow-up are essential elements of the care of persons with anti-LGI1 encephalitis.