Abstract

Background and Objectives Longitudinal outcome studies in leucine-rich glioma inactivated-1 (LGI-1) immunoglobulin G (IgG) autoimmune encephalitis (AE) are needed to inform clinical management and prognostication. This study aims to evaluate longitudinal predictors of disability and disease severity in LGI-1-IgG AE.

Methods This retrospective observational study of patients with LGI-1-IgG AE was conducted between 2013-2022. Disability and disease severity were defined by scores on the modified Rankin Scale (mRS) and the clinical assessment scale in AE (CASE), respectively. Demographic variables, clinical/paraclinical data, brain MRI, and Montreal Cognitive Assessment (MOCA) scores were examined as predictors of mRS and CASE scores in logistic and linear regression models, respectively.

Results Thirty patients (60% male, median age = 68.5; interquartile range (IQR) = 63.0–75.0) were included, with a median follow-up time of 19.1 months (IQR = 5.3–47.1) The majority developed seizures (29, [97%]) and/or cognitive impairment (30, [100%]) and received acute (27, [90%]) and maintenance (23 [77%]) immunotherapy. The median initial MOCA was 23/30 (IQR = 21.0–25.0). Baseline mRS (median = 2.0, IQR = 2.0–3.0) and CASE (mean = 4.3, SD = 3.7) correlated with one another (r = 0.58, p < 0.001) and with initial MOCA score (mRS r = –0.60, p = 0.012; CASE r = –0.56, p = 0.021) After 12 months from symptom onset, mRS (OR = 0.88, [95% CI = 0.82–0.94], p < 0.001) and CASE (β = −0.03, [SE = 0.01], p < 0.001) improved significantly. Lower initial MOCA score (OR = 0.68, 95% CI = 0.47–0.98, p = 0.041) and temporal lobe(s) T2 hyperintensity (OR = 16.50, 95% CI = 2.29–119.16, p = 0.006) were associated with higher mRS longitudinally. At last follow-up, most patients had persistent memory dysfunction (25, [83%]) while few had ongoing seizure activity (3, [10%]).

Discussion Overall, there was a high degree of correlation between mRS and CASE scores in patients with LGI-1-IgG AE, with both scores improving significantly after 12 months. Memory dysfunction and psychiatric disturbance were the most prevalent longitudinal symptoms. Cognitive impairment and temporal lobe T2 hyperintensity at baseline were both associated with greater disability at long-term follow-up, underscoring these as important determinants of disability outcomes in LGI-1-IgG AE.

GlossaryAE=autoimmune encephalitis; CASE=clinical assessment scale in autoimmune encephalitis; CCS=Corticosteroids; FBDS=faciobrachial dystonic seizures; FLAIR=fluid-attenuated inversion recovery; IgG=immunoglobulin G; IQR=interquartile range; IVIG=IV immunoglobulin; LGI-1=leucine-rich glioma inactivated-1; MOCA=Montreal Cognitive Assessment; mRS=modified Rankin Scale; PLEX=plasmapheresisIntroduction

Leucine-rich glioma inactivated-1 (LGI-1) immunoglobulin G (IgG) targets neuronal cell-surface antigens (LGI-1), resulting in an autoimmune limbic encephalitis that predominantly affects older male individuals.1,-,3 LGI-1-IgG is one of the most common autoantibody biomarkers of autoimmune encephalitis (AE), comprising approximately 20% of positive autoantibody biomarkers at tertiary centers.4 Individuals present with subacute cognitive decline, behavioral disturbance, and frequent focal seizures often in the form of faciobrachial dystonic seizure (FBDS).1,2

Longitudinal studies of AE have primarily measured disability using mRS, a global disability scale developed for clinical trials in stroke that is heavily weighted toward motor dysfunction.5,6 Such disability scales may be limited in their ability to capture nonmotor disability in AE, such as cognitive impairment and psychiatric disturbance. The clinical assessment scale in autoimmune encephalitis (CASE) is a disease severity scale designed specifically for AE.6 The score was initially applied broadly to AE and showed excellent interobserver and intraobserver reliability, although most patients included had NMDAR-IgG AE.6 CASE is planned to be used as a secondary outcome measure in randomized controlled trials of novel immunotherapies in autoimmune encephalitis including LGI-1-IgG AE. However, further evaluation of the applicability of this score to AE cohorts is needed.

Previous observational cohort studies of LGI-1-IgG AE demonstrated an improvement in modified Rankin Scale (mRS) scores after immunotherapy.7,-,9 These studies suggested that seizures often improved early in the disease course while cognitive and behavioral symptoms were more persistent.3 Despite treatment, patients with LGI-1-IgG AE can continue to experience long-term cognitive, psychological, and functional impairment.10 Among our anti–LGI-1 AE cohort that underwent neuropsychological assessments, we observed that all patients had persistent cognitive deficits, despite a majority of patients receiving immunotherapy.11 This study aims to evaluate longitudinal clinical outcomes of LGI-1 AE patients using both mRS and CASE and examine whether clinical factors at initial presentation can predict longitudinal disability and disease severity in LGI-1-IgG AE.

MethodsStudy Design and Population

This was a retrospective, observational study of adult patients with LGI-1-IgG AE evaluated at the Cleveland Clinic between 2013 and 2022 who met clinical criteria for AE.12 Patients with LGI-1-IgG AE were identified from the Cleveland Clinic autoimmune neurology registry and electronic medical records.

Standard Protocol Approvals, Registrations, and Patient Consents

This study was approved by the Cleveland Clinic Institutional Review Board with a waiver of consent.

Antibody Testing

All patients were positive in serum or CSF for LGI-1-IgG by cell-based assays at clinical laboratories (Athena diagnostics, Associated Regional and University Pathologists, Mayo Clinic Neuroimmunology laboratory).

Prognostic Clinical Variables

Clinical variables evaluated included demographics, hospital admission, presence of seizure activity, FBDS, malignancy, and CSF inflammation (defined as either pleocytosis and/or elevated CSF protein, elevated immunoglobulin-G index, and/or oligoclonal bands) at diagnosis. Montreal Cognitive Assessment (MOCA)13 was collected after symptom onset and within 100 days of diagnosis (eFigure 1, links.lww.com/NXI/A928). Immunosuppressive therapies were categorized into treatment groups: (1) < 1 month of corticosteroids (CCS), IV immunoglobulin (IVIG), plasmapheresis (PLEX), or combination; (2) ≥ 1 month of CCS, IVIG, PLEX, or combination; (3) rituximab and CCS, IVIG, PLEX, or combination; and (4) azathioprine or mycophenolate and CCS, IVIG, PLEX, or combination.

Brain MRI was obtained after symptom onset and <1 week of diagnosis (eFigure 1, links.lww.com/NXI/A928) on 1.5 and 3 T machines and used standard sequences including T2-weighted and T1-weighted, magnetization-prepared rapid gradient echo, fluid-attenuated inversion recovery (FLAIR), and T1-weighted postcontrast sequences. MRI signal abnormality was determined from both radiologist reports and clinical review.

Clinical Outcomes: Disease Severity and Disability Scores

Clinical outcome scores (mRS and CASE) at each clinical visit were determined retrospectively by 2 neurologists (A.A. and A.K.). CASE scores used 9 measures (score 0–3 for each category, total = 27): seizures, memory dysfunction, psychiatric symptoms, consciousness, language problems, dyskinesias/dystonia, gait instability and ataxia, brainstem dysfunction, and weakness (eFigure 1, links.lww.com/NXI/A928).6 The mRS scores used 6 categories of disability severity ranging from grade 0 “no significant disability” to grade 6 denoting death.14

Statistical Analysis

Patient demographics and baseline clinical characteristics were summarized by mean and SD or median and interquartile range for continuous variables and count with percentage for categorical variables. Univariate linear regression models with baseline MOCA scores as a dependent variable were fit.

Longitudinal mRS scores and CASE scores were plotted against months since symptom onset. Mixed-effect models were fit with moderate mRS (≥3 vs 0–2) and continuous CASE scores as dependent variables and months since symptom onset as an independent variable. Models were fit using all available patient data. To assess clinical outcomes short and long term, models were created with scores <12 months from symptom onset and compared with >12 months. In the bivariate mixed-effect models, each baseline variable in Table 1 was added one at a time as an independent variable in addition to months since symptom onset. All computations were performed in SAS version 9.4 (SAS Institute Inc, Cary NC). All tests were two-sided, and p-values less than 0.05 were considered statistically significant.

Table 1

Patient Characteristics

Data Availability

Anonymized data not published within this article will be made available by reasonable request from any qualified investigator on approval by the institutional review board.

ResultsStudy Cohort

Thirty patients with LGI-1-IgG AE were included, of whom 18 (60%) were male, with a median age of 68.5 (IQR = 63.0–75.0) years at disease onset. The median time from symptom onset to diagnosis was 16.3 (IQR = 6.3–29.1) weeks. The median follow-up time was 19.1 (IQR = 5.3–47.1) months, with a median number of follow-up assessments of 10 (IQR = 3–14). Hospital admission was required for 24 patients (80%), with a median of 2 (IQR = 1–3) admissions. The most frequent symptoms included seizures (29, [97%]) and cognitive impairment (30, [100%]). Status epilepticus was diagnosed in 6 patients (20%), and 15 patients (50%) had FBDS. No patient had concurrent contactin-associated protein-like 2 IgG. Eight patients (27%) had a history of malignancy including breast (2), pancreatic (1), colorectal (1), bladder (1), prostate (1), thyroid (1), and non-Hodgkin lymphoma (1) (Table 1).

MRI was available for 28 patients after symptom onset and <1 week of diagnosis and showed temporal lobe T2 hyperintensity in 15 patients (54%). The median time from diagnosis to MRI was −1.3 (IQR = −4.1 to −0.59) months. Of the 25 patients with available CSF samples, 10 patients had inflammatory CSF (40%). Most patients received acute immunotherapy (CCS and/or IVIG and/or PLEX; 27, [90%]), in addition to maintenance immunotherapy, (23 [77%]). Maintenance therapy included rituximab (13, [43%]), IVIG (5, [16.7%), mycophenolate (3, [10%]), oral prednisone slow taper (3 [10%]), and azathioprine (1, [3%]).

Baseline Disease Severity (CASE) and Disability (mRS) Scores: Distribution and Correlation

The median time from symptom onset to initial mRS and CASE was 1.1 months (IQR = 0.49–4.8). The mean baseline CASE score was 4.3 (SD = 3.7), with the most frequently affected clinical domains being seizure, memory dysfunction, and psychiatric disturbance (Figure 1). The median baseline mRS was 2.0 [IQR = 2.0–3.0] (Table 1). Baseline mRS and CASE scores were moderately correlated (r = 0.58, p < 0.001).

Figure 1 CASE Scores at Baseline and Last Follow-up, Stratified by Symptoms

*Bars with percentage less than 10% are not labeled in the figure. CASE = clinical assessment scale in autoimmune encephalitis.

Cognitive Impairment on Initial MOCA Was Associated With Higher Baseline Disease Severity (CASE) and Disability (mRS) Scores

MOCA score obtained within 100 days of diagnosis was available for 17 of 30 patients. The median time from diagnosis to MOCA assessment was 0.13 (IQR = −0.07–0.30) months. The median initial MOCA score was 23/30 (IQR = 21.0–25.0) (Table 1). Both baseline mRS and CASE had moderate correlation with initial MOCA score (mRS r = −0.60, p = 0.012; CASE r = −0.56, p = 0.021). Higher baseline mRS score (parameter estimate [SE] = −2.68 (0.85), p = 0.007) and higher baseline CASE score (parameter estimate [SE] = −1.45 (0.50), p = 0.011) were associated with lower baseline MOCA score (eTable 1, links.lww.com/NXI/A928).

Longitudinal Trajectory of Disease Severity (CASE) and Disability (mRS) Scores

Twenty-nine patients had longitudinal mRS and CASE scores and 17 patients had follow-up ≥ 12 months. Figure 2 demonstrates that overall, patients improved in their CASE and mRS scores over time. Last visit mRS and CASE scores correlated well (r = 0.83, p < 0.001). To examine the impact of short-term vs longer-term follow-up, we examined outcomes at 12 months and >12 months. At 12 months, CASE had improved, but mRS had not (Table 2). However, both mRS (OR = 0.88, 95% CI 0.82–0.94, p < 0.001) and CASE (β = −0.03, [SE] = 0.01, p < 0.001) improved after 12 months from symptom onset (Table 2 and Figure 2).

Figure 2 Longitudinal Trajectory of (A) CASE and (B) mRS Scores Over Time

*The red bolded lines are fitted LOESS curves. **17 patients had follow-up >12 months. CASE = clinical assessment scale in autoimmune encephalitis; mRS = modified Rankin Score.

Table 2

Mixed-Effect Models for CASE and Moderate mRS

Longitudinal Encephalitis Symptoms Measured by the CASE Score

At initial evaluation, most patients had acute refractory seizures (21, [70%] with CASE subscore ≥2), whereas at final clinical encounter only 3 (10%) patients had refractory seizures, although 90% remained on antiseizure medications (Figure 1). Moderate-to-severe memory dysfunction (CASE subscore ≥2) was present in 14 (47%) patients at initial evaluation, improving to 7 (23%) patients reporting moderate-to-severe memory dysfunction at final clinical encounter. However, 25 (83%) patients continued to experience mild (CASE subscore ≥1) memory dysfunction (Figure 1). Moderate-to-severe psychiatric symptoms were present (CASE subscore ≥2) at initial evaluation in 3 (10%) and remained constant at last follow-up, but the proportion of patients with mild (CASE subscore ≥1) psychiatric symptoms increased from 9 (30%) to 15 (50%) at last clinical visit (Figure 1).

Subgroup Analyses of Longitudinal Outcomes in Patients With Severe Disability at Onset

In the subgroup patients with a moderately severe CASE score of ≥6 at baseline (n = 5), all improved to a CASE score <6 within 20 months and remained at a CASE score <6 at last follow-up (eFigure 2, links.lww.com/NXI/A928). In the subgroup patients with a moderately severe mRS ≥2 at baseline (n = 24), 8 (33.3%) improved to an mRS of 0 or 1 within 20 months and remained at an mRS of 0 or 1 for the duration of follow-up while the remaining 16 remained at an mRS ≥2 (eFigure 2).

Factors Associated With Higher Long-term Disease Severity (CASE) and Disability (mRS) Scores

Lower initial MOCA (OR = 0.68, 95% CI 0.47–0.98, p = 0.041) and abnormal brain MRI temporal lobe T2/FLAIR hyperintensity (OR = 16.50, 95% CI 2.29–119.16, p = 0.006) were associated with higher mRS longitudinally (Figure 3, eTable 2, links.lww.com/NXI/A928). Female sex was associated with higher longitudinal CASE score, but this was driven by a single outlier. No other variables including inflammatory CSF, presence of seizure activity, FBDS, malignancy, hospital admission, or acute vs long-term treatment at onset were associated with longitudinal mRS or CASE (eTables 2 and 3).

Figure 3 Clinical Predictors of (A) Moderate mRS and (B) CASE at Last Follow-up

*Significant predictor of CASE; †Significant predictor of mRS; Error bars: 95% confidence interval. CASE = clinical assessment scale in autoimmune encephalitis; FBDS = faciobrachial dystonic seizure; MOCA = Montreal Cognitive Assessment; mRS = modified Rankin Score; Treatment group 2 = > 1-month corticosteroids, IV immunoglobulin, plasmapheresis, or combination; Treatment group 3 = acute therapy and rituximab; Treatment group 4 = acute therapy and azathioprine or mycophenolate.

Discussion

This study examined predictors of longitudinal disability and disease severity outcomes in an observational cohort of LGI-1-IgG AE. CASE and mRS correlated throughout the disease course and improved in most patients over time. Memory dysfunction and psychiatric disturbance were the predominant long-term symptoms while seizures improved in the majority. Lower MOCA score and temporal lobe hyperintensity at diagnosis were associated with greater disability at long-term follow up.

Over time, disease severity and disability in the LGI-1-IgG AE cohort improved, including in the subgroup with high mRS and CASE at onset, although the improvement in mRS was not significant until after 12 months of monitoring. These findings indicate that disability measurements documented within the first year of monitoring may not be truly indicative of long-term disability. This is important for observational studies and clinical trial design, highlighting the need for longitudinal follow-up > 12 months to accurately determine disability outcomes.

Our study demonstrated a correlation between mRS and CASE at baseline and longitudinally, consistent with previous reports in other forms of AE,6 suggesting a close relationship between disease severity and overall disability level. CASE subscores offered additional insights, beyond the global mRS score, into specific AE-associated symptoms at onset, allowing for the evaluation of these symptoms longitudinally. A high proportion of patients had seizures and memory impairment at baseline, similar to previously described cohorts.3 Longitudinal follow-up demonstrated that seizure burden improved over time while cognitive and psychiatric symptoms persisted. Although prevalent at onset, the presence of seizures was not associated with disability or disease severity at baseline or last follow-up. This may be due to the focal and brief nature of seizures in LGI-1-IgG AE3 which may lessen their contribution to overall disability.

Psychiatric symptoms, most commonly involving anxiety and depression, became more prevalent with time. It is possible that these symptoms are better recognized in longitudinal clinical visits both by patients and their providers, as opposed to the acute setting where patients are more impaired in their ability to recognize and communicate these symptoms and where other symptoms may take precedent. Furthermore, persistent cognitive impairment and its impact on an individual's capacity to return to previous activities may further contribute to mood disturbance. These finding suggest that psychiatric symptoms represent an important facet of this disease, requiring increased attention during longitudinal clinical visits.

In contrast to seizure activity which improved longitudinally, cognitive impairment persisted, as measured by both CASE memory dysfunction and MOCA scores, although did not interfere with daily activities in most patients at last follow-up. This reflects our findings that most (if not all) patients have cognitive deficits on neuropsychological testing.11 Furthermore, both baseline mRS and baseline CASE were significantly associated with initial MOCA score, suggesting that patients with a lower initial MOCA score had more severe disease and a higher overall level of disability at presentation. A lower initial MOCA score was also associated with a higher mRS longitudinally. Together these findings highlight that cognitive impairment is a significant contributor to long-term disability. It may be useful for clinicians to screen patients early in the disease process for cognitive impairment in order to aid prediction of longitudinal disability.

T2 hyperintensity of the temporal lobe(s) at initial evaluation was associated with higher mRS longitudinally. This association could represent more severe disease and/or longer disease duration prior to diagnosis, prolonged seizure activity, greater acute inflammation, and possibly greater risk of longitudinal hippocampal atrophy.7,15,-,17

Limitations of this observational study include its retrospective nature. Clinical assessments were completed ad hoc, rather than at specific time points; however, this reflects real-world practice. We had limited longitudinal MOCA testing, limiting our ability to evaluate the application of this as a longitudinal outcome measure at this time. This study demonstrates that the application of CASE score appears to offer a more detailed symptom-based score for AE than mRS and could be used longitudinally for outcome studies. However, as the CASE score was developed in a cohort enriched with anti-NMDAR AE patients, we observed some limitations of the CASE score for LGI-1-IgG AE including the lack of assessment of sleep symptoms, which are frequent in this population,18 and the lack of precision for measuring severity of specific psychiatric symptoms which may be better assessed using validated psychiatric measures such as Patient Health Questionnaire-9 and Generalized Anxiety Disorder-7.19 Furthermore, CASE score for memory dysfunction does not provide the granular domain specific findings that can be obtained from detailed neuropsychological testing. This was a highly treated cohort, and we were limited by power and retrospective design to evaluate different treatment effects on disability outcomes.

In this longitudinal study of LGI-1-IgG AE, we have found that the CASE severity score correlates well with disability measured by mRS, and CASE provided the additional value of greater granularity of AE symptoms and their severity. Cognitive impairment at onset is an important predictor of longitudinal functional disability, suggesting a role for early cognitive evaluations for prognostication.

Study Funding

The authors report no targeted funding.

Disclosure

A. Aboseif, Y. Li, B. Lapin, M. Amin, R. Galioto and A. Rae-Grant report no disclosures. A. Milinovich has received research support from Merck, Pfizer, Twin Health, NFLPA, and the National Institute on Aging. J.R. Abbatemarco has received personal compensation for scientific advisory boards for EMD Serono, Genentech, and Horizon, and has received research support from Horizon. J.A. Cohen has received personal compensation for consulting for Biogen, Convelo, EMD Serono, Gossamer Bio, Mylan and PSI; and is serving as an Editor of Multiple Sclerosis Journal. V. Punia has received research support from Eisai INc., American Epilepsy Society, Ohio Department of Higher Education. A. Kunchok has received personal compensation for scientific advisory boards for Genentech and Horizon therapeutics, and consulting for EMD Serono. Go to Neurology.org/NN for full disclosures.

Appendix AuthorsFootnotes

Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

The Article Processing Charge was funded by Cleveland Clinic.

Submitted and externally peer reviewed. The handling editor was Editor Josep O. Dalmau, MD, PhD, FAAN.

Received June 7, 2023.Accepted in final form September 5, 2023.Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American Academy of Neurology.