Prediction of Long-term Survival After Status Epilepticus Using the ACD Score

Key Points

Question  Does the duration of status epilepticus change the outcome of this condition?

Findings  In this cohort study of 261 patients with status epilepticus diagnosis in Denmark, age, duration, and nonconvulsive type of status epilepticus were associated with developing new neurological deficits at hospital discharge as predicted by a novel, 3-factor scoring system, the ACD score. New neurological deficits after discharge were a risk factor for all-cause mortality regardless of the cause of status epilepticus.

Meaning  Findings of this study suggest that the ACD score is a reliable tool for predicting long-term survival, especially in patients whose status epilepticus did not have brain-damaging causes.

Importance  Early prediction of long-term mortality in status epilepticus is important given the high fatality rate in the years after diagnosis.

Objective  To improve prognostication of long-term mortality after status epilepticus diagnosis.

Design, Settings, and Participants  This retrospective, multicenter, multinational cohort study analyzed adult patients who were diagnosed with and treated for status epilepticus at university hospitals in Odense, Denmark, between January 1, 2008, and December 31, 2017, as well as in Oslo, Norway; Marburg, Germany; and Frankfurt, Germany. They were aged 18 years or older and had first-time, nonanoxic status epilepticus. A new scoring system, called the ACD score, for predicting 2-year (long-term) mortality after hospital discharge for status epilepticus was developed in the Danish cohort and validated in the German and Norwegian cohorts. The ACD score represents age at onset, level of consciousness at admission, and duration of status epilepticus. Data analysis was performed between September 1, 2019, and March 31, 2020.

Exposures  Long-term follow-up using data from national and local civil registries in Denmark, Norway, and Germany.

Main Outcomes and Measures  The predefined end point was 2-year survival for all patients and for a subgroup of patients with status epilepticus causes that were not damaging or were less damaging to the brain. Neurological deficits before and after onset, demographic characteristics, etiological categories of status epilepticus, comorbidities, survival, time points, treatments, and prognostic scores for different measures were assessed.

Results  A total of 261 patients (mean [SD] age, 67.2 [14.8] years; 132 women [50.6%]) were included, of whom 145 patients (mean [SD] age, 66.3 [15.0] years; 78 women [53.8%]) had status epilepticus causes that were not damaging or were less damaging to the brain. The validation cohort comprised patients from Norway (n = 139) and Germany (n = 906). At hospital discharge, 29.8% of patients (n = 64 of 215) had new moderate to severe neurological deficits compared with baseline. New neurological deficits were a major predictor of 2-year survival after hospital discharge (odds ratio, 5.1; 95% CI, 2.2-11.8); this association was independent of etiological category. Nonconvulsive status epilepticus in coma and duration of status epilepticus were associated with development of new neurological deficits, and a simple 3-factor score (ACD score) combining these 2 risk factors with age at onset was developed to estimate survival after status epilepticus diagnosis. The ACD score had a linear correlation with 2-year survival (Pearson r² = 0.848), especially in the subset of patients with a low likelihood of brain damage.

Conclusions and Relevance  This study found that age, long duration, and nonconvulsive type of status epilepticus in coma were associated with the development of new neurological deficits, which were predictors of long-term mortality. Accounting for risk factors for new neurological deficits using the ACD score is a reliable method of prediction of long-term outcome in patients with status epilepticus causes that were not damaging or were less damaging to the brain.

Mortality from status epilepticus after hospital discharge substantially exceeds in-hospital mortality.^1,2 Evidence is scarce regarding risk factors associated with long-term mortality or reliable clinical scores that enable prognostication of long-term outcomes. Retrospective analyses suggest that the risk factors for postdischarge mortality differ from those for mortality in the acute phase and that established risk factors and prognostic scores are insufficient to predict survival after discharge.^3,4

Neurological deficit is a well-known complication of status epilepticus,^5,6 but the extent to which it contributes to patient outcome is poorly understood.⁷ In animal models of status epilepticus, seizure-induced cell death has been associated with the development of chronic epilepsy,⁸ and pathological studies in humans have reported hippocampal necrosis.⁹ The current classification of status epilepticus defines time points after which poor outcome may be expected owing to long-term consequences, such as neuronal death or injury and alteration of neuronal networks. However, the clinical evidence supporting this recommendation remains sparse.¹⁰ Thus, the cause of status epilepticus, and not the seizure-induced neuronal damage, is commonly considered to be the crucial factor in short- and long-term outcomes.^11-13

Indirect evidence suggests that seizure-induced neuronal damage might be a substantial factor in poor patient outcome. The time to seizure termination is associated with outcome,^12,14 and late or underdosed first-line treatment with benzodiazepines is associated with a lower rate of status epilepticus control and higher in-hospital mortality.^6,15,16 In addition, a case-control study suggested that seizure burden was associated with functional and cognitive outcomes in patients with subarachnoid hemorrhage.¹⁷ The implication of seizure duration for survival after hospital discharge is unknown, however.

To improve prognostication of outcomes after status epilepticus, we studied the risk factors associated with long-term outcome, developed a new score for predicting long-term mortality after the onset of status epilepticus with causes that were not damaging or were less damaging to the brain, and externally validated this score in 2 large, independent cohorts.

This retrospective, multicenter, multinational cohort study was approved by the Danish Health Authorities, which also waived the informed consent requirement; approval from an ethics committee was not required according to Danish legislation. Evaluation of the German patients was part of a study on status epilepticus outcomes^18,19 that was registered in the German Clinical Trials Register and approved by the local ethics committees in Frankfurt and Marburg, Germany. The Norwegian data were collected from a quality assessment study that was approved by the Data Protection Official, which provided permission to publish study results in the general public interest. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.²⁰

The exploratory cohort consisted of patients in the Danish cohort, who were diagnosed and treated at Odense University Hospital between January 1, 2008, and December 31, 2017 (eFigure 1 in the Supplement). These patients were identified retrospectively from referrals for acute electroencephalogram and/or from International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes at discharge. Details on this subset of patients have been reported.^2-4 Inclusion criteria were age 18 years or older and a diagnosis of status epilepticus, which was defined as convulsive seizures longer than 5 minutes, nonconvulsive seizures longer than 10 minutes,¹⁰ and nonconvulsive or possible nonconvulsive status epilepticus shown on electroencephalogram according to the Salzburg criteria.²¹ Exclusion criteria were anoxic or hypoglycemic brain damage, previous episodes of status epilepticus, and residency outside of Denmark.

The validation cohorts comprised patients in the German and Norwegian cohorts. Patients in the German cohort were diagnosed at Frankfurt University Hospital and at University Hospital of Giessen and Marburg (patients were from Marburg only) between January 1, 2011, and December 31, 2017. All patients with status epilepticus were included, and this subset has been described elsewhere.^18,19 For this analysis, inclusion and exclusion criteria were adapted from the criteria in the Danish cohort. Survival status of the German cohort was obtained through annual follow-up by the local registration offices in the patients’ areas of residence.

Patients in the Norwegian cohort were diagnosed at Oslo University Hospital between January 1, 2001, and December 31, 2017. Patients with status epilepticus were retrospectively identified using the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision codes at discharge from 2001 to 2017, and information about these patients has been reported.^22,23 Inclusion and exclusion criteria were slightly modified in accordance to the Danish cohort, and 92% of eligible patients (n = 139 of 151) were included in the validation cohort. Survival data were available for all patients owing to the linkage between the electronic medical records of the hospital and the Norway Civil Registration.

The electronic medical records of the exploratory cohort were analyzed retrospectively. Survival data were available for all patients in this cohort owing to the linkage with the Danish Civil Registration System. Long-term survival was defined as survival 2 years after status epilepticus onset. Data were stored using REDCap (Vanderbilt University).²⁴ The development of neurological deficits after onset was measured by the difference in the estimated National Institutes of Health Stroke Scale (NIHSS) score between baseline (before status epilepticus occurrence) and the first follow-up after or at hospital discharge.²⁵ Substantial new neurological deficits were defined as an increase in NIHSS score of 5 points or more using the definitions for acute stroke.²⁶

We also assessed demographic characteristics, duration of status epilepticus, semiological traits, treatment data, Status Epilepticus Severity Score (STESS), modified Rankin Scale (mRS) score for degree of disability, Barthel index for activities of daily living,²⁷ and Charlson Comorbidity Index.^28,29 Scores were approximated retrospectively based on information in the electronic medical records. Semiological traits were classified according to the International League Against Epilepsy recommendations.¹⁰

Nationality data were provided in the Danish civil register. The categories were Danish and other, but no data were provided for the other category.

The cause of status epilepticus in each patient was assessed, and patients were assigned to the appropriate etiological categories. For example, if the condition was caused by acute cerebral disease and presented relevant risk of or proven cerebral damage, patients were not assigned to the group of causes that were not damaging or were less damaging to the brain. Examples of cerebral disease that poses relevant risk for cerebral damage are acute ischemic stroke; cerebral hemorrhage; cerebral infection; probable autoimmune encephalitis; rapidly progressing dementia; and malignant brain tumor, including low-grade glioma. Conversely, status epilepticus causes such as remote symptomatic category, insufficient adherence to medication, alcohol abstinence, mild dementia, and drug-induced seizure were classified as not damaging or less damaging to the brain. Among the etiological categories defined by the International League Against Epilepsy³⁰ were remote symptomatic status epilepticus, acute symptomatic status epilepticus, cryptogenic status epilepticus, progressive central nervous system disorder, and defined electroclinical syndrome.

Except for patients with a previously established epilepsy diagnosis, all patients received at least 1 cerebral scan (magnetic resonance imaging or computed tomography), lumbar puncture, and/or state-of-the art autoimmune antibody screening (since 2010 in the Danish cohort), if required. Patients with absence status epilepticus were excluded from the cohort of patients with status epilepticus causes that were not damaging or were less damaging to the brain because of different pathophysiology and prognosis,^10,31 whereas patients with idiopathic epilepsy and convulsive status epilepticus that was caused by insufficient compliance remained in this group.

Statistical analyses were performed from September 1, 2019, to March 31, 2020. We used IBM SPSS, version 24.0 or 26.0 (IBM SPSS) and Stata, version 16 (StataCorp LLC).

The predefined end point was 2-year survival in the logistic regression model for all patients and for a subgroup of patients with status epilepticus causes that were not damaging or were less damaging to the brain. Survival outcome was not censored; initial analyses checked linearity of the logit. The model fit was investigated using goodness-of-fit statistics and casewise diagnostics. The overall fit of the model was assessed with Cox and Snell R². Model coefficients (β) were evaluated with the Wald statistic and accompanied by 95% CIs. Effect sizes were computed from odds ratios (ORs). All variables were included in the models by forced entry.

A linear regression model was developed to investigate predictors of new neurological deficits at first follow-up. The included characteristics were selected from preliminary analyses. Given the apparent heteroscedasticity across independent variables, a weighted least-squares regression was used to minimize bias in the model. Significance of the model was assessed with the F statistic, and the goodness of fit was assessed with R². The significance of individual coefficients was assessed using the t statistic, and effect sizes were measured with coefficient β.

To estimate short-term survival without new neurological deficits, we used a logistic regression model that applied the least absolute shrinkage and selection operator (LASSO)³² method to regularize coefficient estimates and to perform variable selection. The LASSO model was applied to the subgroup of patients with status epilepticus causes that were not damaging or were less damaging to the brain. The regularization parameter was chosen to be the smallest regularization, leading to no more than 5 predictors in the model. The following variables were used: age, convulsive status epilepticus, definite and possible nonconvulsive status epilepticus according to the Salzburg criteria, nonconvulsive status epilepticus in coma, Charlson Comorbidity Index, logarithm of the duration of status epilepticus (hours), mRS score at baseline, NIHSS score at baseline, refractory status epilepticus, time from status epilepticus onset to diagnosis, need for treatment in the intensive care unit, level of consciousness (according to STESS definition), history of epilepsy (according to STESS definition), and worst seizure type (according to STESS definition).²⁸

The models were fitted to data from the exploratory cohort, and the area under the curve (AUC) of a receiver operating characteristic curve was estimated. To address overfitting, bootstrapping was applied to estimate the bootstrap optimism.^33,34 The corrected AUC was the apparent AUC subtracted from the bootstrap optimism. Based on the LASSO model, a 5-factor risk score was derived by approximating the linear predictor (after centering and scaling); continuous parameters were approximated by step functions, which for status epilepticus duration led to asymmetric intervals because of back-transformation from the logarithmic scale.

A total of 261 patients (mean [SD] age, 67.2 [14.8] years; 132 women [50.6%] and 129 men [49.4%]) were included. Among this full cohort was a subset of 145 patients (mean [SD] age, 66.3 [15.0] years; 78 women [53.8%] and 67 men [46.2%]) with status epilepticus causes that were not damaging or were less damaging to the brain. The characteristics and outcome of the full cohort and subset of patients are shown in Table 1. eFigure 1 in the Supplement provides an overview of patient identification, screening, and inclusion.

The mean follow-up duration was 2.2 years, and 27 patients had a follow-up of less than 2 years. For all patients, mortality increased nearly 3-fold from the short-term in-hospital rate of 17.6% to the long-term (after 2 years) rate of 47.1%. None of the patients with electroclinical syndrome or absence status epilepticus developed new neurological deficits after status epilepticus onset and were therefore removed from further analyses.

The logistic regression model using 2-year survival after discharge as the binary end point showed an association between new neurological deficits and progressive central nervous system disorders in the full cohort after adjustment for known prognostic factors (adjusted OR, 5.1; 95% CI, 2.2-11.8) (Table 2). The results for new neurological deficits did not change in the subset including only remote symptomatic and unknown causes (Table 2).

Figure 1A illustrates the association between new neurological deficits and survival of patients in the full cohort who were discharged alive from the hospital. The estimated median survival for patients with an NIHSS score increase lower than 5 points was 85.2 (95% CI, 30.0-107.2) months compared with 34.2 (95% CI, 8.1-60.3) months in patients with an NIHSS score increase between 5 and 10 points and 3.4 (95% CI, 2.3-4.5) months in patients with an NIHSS score increase higher than 10 points. eFigure 2 in the Supplement shows survival according to duration, change in disability (mRS score), and performance of activities of daily living (Barthel index), which confirmed the associations. New neurological deficits at discharge remained associated with poor survival regardless of etiological category, as illustrated by the survival curves from patients with remote symptomatic status epilepticus (Figure 1B). For example, 52.9% of patients (9 of 17) with an NIHSS score increase of 5 points or higher died within 2 years after discharge, whereas this outcome was the case for only 19.0% of patients (11 of 58) with an NIHSS score increase lower than 5 points.

Among patients who were alive at discharge or first follow-up, 29.8% (64 of 215) had developed moderate to severe neurological deficits (Table 1). An increase in NIHSS score correlated with decreased Barthel index (Pearson r = 0.74; P < .001) and increased mRS score (Pearson r = 0.58; P < .001). The mean increase in NIHSS score was lowest in patients with remote symptomatic status epilepticus (2.4 points) and was similar among all other etiological categories (acute symptomatic status epilepticus: 5.4 points; cryptogenic status epilepticus: 5.5 points; and progressive central nervous system disorder: 5.3 points; P = .03, Kruskal-Wallis test). Table 3 gives an overview of clinical characteristics associated with increased NIHSS score at discharge. In the subset of patients with status epilepticus causes that were not damaging or were less damaging to the brain, we found that mean duration of status epilepticus and nonconvulsive status epilepticus in coma remained associated with change in neurological function (Table 3).

We developed a LASSO model based on patients with status epilepticus causes that were not or were less damaging to the brain. The binary outcome of the LASSO model we developed was being alive without new neurological deficits at discharge or first follow-up. The LASSO model identified 5 prognostic factors (age, duration of status epilepticus, time to diagnosis, treatment in the intensive care unit, and level of consciousness at hospital admission) that were included in a preliminary 5-factor score. Using the exploratory cohort and receiver operating characteristics, we found that the AUC of the 5-factor score was 0.80 after correcting the optimism of the model using bootstrapping (with the uncorrected AUC being 0.84).

Brief prognostic scores are preferable in clinical practice, and 2 components of the 5-factor score contributed only marginally to the total score. We therefore simplified the 5-factor score by removing the 2 weakest factors (treatment in the intensive care unit and time to diagnosis). The simplified score is the ACD score, which represents age at onset, level of consciousness at admission, and duration of status epilepticus (Figure 1C).

We tested the accuracy of the ACD score in the entire cohort including all etiological categories. The AUCs of the preliminary 5-factor score and the ACD score were similar (0.79 [95% CI, 0.73-0.85] vs 0.78 [95% CI, 0.72-0.84]). The optimal cutoff value for the ACD score was 10 points, with a sensitivity of 0.63 and a specificity of 0.82 (and a Youden index of 0.45) for being alive without new neurological deficits at first follow-up. The positive predictive value of an ACD score lower than 10 points for survival without new neurological deficits was 0.85, and the negative predictive value was 0.58.

We validated the ACD score in 2 independent cohorts of patients with available long-term follow-up data. The cohort from Germany comprised 906 patients, of whom 542 (59.8%) had status epilepticus with causes that were not damaging or were less damaging to the brain, and the cohort from Norway consisted of 139 patients, with 105 (75.5%) belonging to a subset with status epilepticus causes that were not damaging or were less damaging to the brain. In the cumulative cohort from all 3 countries (Denmark, Germany, and Norway), the ACD score had a correlation with survival as illustrated by the Kaplan-Meier survival curves in Figure 1D. The AUC of the ACD score was less accurate in patients with status epilepticus causes that had mixed etiological categories (Norway: AUC, 0.700 [95% CI, 0.601-0.799]; Germany: AUC, 0.689 [95% CI, 0.649-0.728]) than in the subset of patients with status epilepticus causes that were not damaging or were less damaging to the brain and when using receiver operating characteristic analyses and mortality 2 years after diagnosis (Norway: AUC, 0.763 [95% CI, 0.636-0.839]; Germany: AUC, 0.733 [95% CI, 0.685-0.781]). Surviving patients with a follow-up of less than 2 years were censored.

We therefore tested the ACD score in the subset of patients from all 3 countries (n = 792). An ACD score higher than 10 points was associated with mortality in the Kaplan-Meier analyses in the validation cohorts from Germany and Norway (Figure 2A). In the combined cohort, mortality at 2 years after discharge showed an almost linear correlation with the ACD score (Pearson r² = 0.848) (Figure 2B). The association was essentially unchanged when using logistic regression or risk difference regression. A nomogram based on the linear regression line in Figure 2B provides a tool for estimating 2-year survival in patients with status epilepticus causes that were not damaging or were less damaging to the brain (Figure 2C).

In this cohort study, we found that accounting for risk factors for new neurological deficits associated with status epilepticus using the ACD score is a valuable way to estimate long-term survival after status epilepticus diagnosis. We believe the ACD score complements the available scores for estimating in-hospital mortality^35-38 and orienting early treatment strategies. Furthermore, we believe that the ACD score has substantial potential to help clinicians make treatment decisions in patients with ongoing and long-standing status epilepticus attributable to remote symptomatic or unknown etiological categories because it predicts both neurological outcome and long-term survival. Although the ACD score is associated with 2-year survival in all etiological categories, it is limited by the obvious impact of, for example, progressive causes and comorbidities in patients with brain-damaging status epilepticus causes. Conversely, the ACD score appeared to be more reliable in patients with cerebral scans showing no acute changes and no indications of inflammation.

Establishing the exact duration of status epilepticus may be a challenge for the ACD score. We defined duration as the time from diagnosis of status epilepticus to termination as judged by the treating neurologist. This duration did not mean that the patients were experiencing status epilepticus for the entire period, and only few patients had continuous electroencephalogram monitoring. In addition, duration may partially reflect the severity of the underlying cause and thereby affect the expected future outcome. Although these shortcomings challenge pathophysiological conclusions, they do not impair the clinical usability of the ACD score.

In addition to developing a novel, simple prognostic score, we clarified the impact of new neurological deficits associated with status epilepticus. Regardless of how we analyzed the data, new neurological deficits were a predictor of long-term survival regardless of the status epilepticus cause and were associated with a 5-fold increase in the odds of death at 2 years after status epilepticus diagnosis. However, the data did not elucidate the pathological mechanisms associated with new neurological deficits. Additional studies are needed to address whether neuronal cell death, changes in cortical networks, or secondary mechanisms are crucial for functional decline.

This study has several strengths. All Scandinavian survival data were complete because of the linkage between patients’ medical records and the national civil registers. Although the German cohort used a different method to obtain long-term follow-up data, long-term survival was similar between the Danish and the German cohorts. The more favorable outcome of the Norwegian patients was likely attributable to a lower proportion of patients with acute symptomatic status epilepticus. To our knowledge, the NIHSS had not been used before this study to quantify neurological deficits after status epilepticus diagnosis. It has some limitations because it does not cover cognitive deficits and may provide a falsely high score in patients with impaired consciousness. However, we chose the NIHSS because it is well established, is less affected by comorbidities, and covers all important supratentorial functions. Its high correlation with the other assessments (mRS score and Barthel index) indicates that the NIHSS provides reliable estimates of the patients’ status before and after discharge.

This study also has several limitations. First, although the study is limited by its retrospective design, we do not believe that the retrospective estimation of the NIHSS score biased the results given that the same raters assessed it uniformly. Furthermore, the NIHSS assessment was based on the comprehensive description of the neurological status from treating neurologists and physiotherapists; therefore, the NIHSS scores were likely to be close to values found in a prospective study. The primary end point of 2-year survival would be little affected by the biases that are often associated with retrospective studies given the almost complete follow-up. Second, selection bias is another possible limitation of retrospective studies. However, mortality and other patient outcomes in the Danish cohort were similar to those reported in prospective studies^6,39 and in large population-based analyses.^40,41 Therefore, selection bias was unlikely to affect the main conclusions of this study.

This cohort study found that age, long duration of status epilepticus, and nonconvulsive status epilepticus in coma were associated with the development of new neurological deficits; in turn, these neurological deficits were a negative predictor of long-term survival. Accounting for risk factors for new neurological deficits using the ACD score allows for the reliable prediction of the long-term outcome in patients with status epilepticus causes that were not damaging or were less damaging to the brain.

Accepted for Publication: February 7, 2022.

Published Online: April 11, 2022. doi:10.1001/jamaneurol.2022.0609

Correction: This article was corrected on June 1, 2022, to fix a label in Figure 2C and the figure caption.

Corresponding Author: Christoph Patrick Beier, MD, Department of Neurology, Odense University Hospital, Sdr Boulevard 29, 5000 Odense C, Denmark (cbeier@health.sdu.dk).

Author Contributions: Dr C.P. Beier had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Roberg and Monsson contributed equally to the manuscript.

Concept and design: Roberg, Monsson, Taubøll, Strzelczyk, D. Beier, Beniczky, C.P. Beier.

Acquisition, analysis, or interpretation of data: Roberg, Monsson, Kristensen, Dahl, Ulvin, Heuser, Strzelczyk, Bechert, Knake, Rosenow, D. Beier, Beniczky, Krøigård, C.P. Beier.

Drafting of the manuscript: Roberg, Monsson, Kristensen, Dahl, Heuser, Strzelczyk, C.P. Beier.

Critical revision of the manuscript for important intellectual content: Roberg, Monsson, Kristensen, Ulvin, Heuser, Taubøll, Strzelczyk, Bechert, Knake, Rosenow, D. Beier, Beniczky, Krøigård, C.P. Beier.

Statistical analysis: Roberg, Monsson, Kristensen, Dahl, Taubøll, Strzelczyk, D. Beier, C.P. Beier.

Obtained funding: Roberg, Monsson, Heuser, Taubøll.

Administrative, technical, or material support: Roberg, Monsson, Dahl, Heuser, Taubøll, Bechert, Knake, Rosenow.

Supervision: Heuser, Taubøll, Strzelczyk, Beniczky, Krøigård, C.P. Beier.

Conflict of Interest Disclosures: Dr Strzelczyk reported receiving personal fees from Arvelle Therapeutics, Desitin Arzneimittel, Eisai, GW Pharmaceuticals, Marinus Pharma, UNEEG Medical, UCB (Union Chimique Belge) Pharma, and Zogenix as well as grants from Zogenix and GW Pharmaceuticals outside the submitted work. Dr Rosenow reported receiving personal fees as a speaker from Angelini Pharma Inc, Eisai GmbH, Arvelle Therapeutics, Novartis, UCB Pharma, and GW Pharmaceuticals; personal fees as an advisor from Arvelle Therapeutics and UCB Pharma; personal fees as a course organizer from Eisai GmbH; and research grants from Eisai GmbH, European Union, and LOEWE programme of the federal state of Hesse outside the submitted work. Dr Beniczky reported receiving personal fees as a speaker from Natus Neuro, personal fees as a consultant from Epihunter, and personal fees as a speaker from Eisai A/S outside the submitted work. Dr. Beier reported receiving personal fees as speaker from Angelini Pharma Inc, Eisai A/S, and UCB Nordic A/S; personal fees as an advisor from Arvelle Therapeutics; and research grants from Eisai A/S. No other disclosures were reported.

Funding/Support: Drs Roberg and Monsson were funded by a 6-month scholarship from the University of Southern Denmark.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Bertil Pederson, MD, University of Southern Denmark, assisted with identifying patients, and Claire Gudex, MD, MPH, MBChB, University of Southern Denmark, provided proofreading. These individuals received no additional compensation, outside of their usual salary, for their contributions.