Role of machine learning in the management of epilepsy: a systematic review protocol.

INTRODUCTION: Machine learning is a rapidly expanding field and is already incorporated into many aspects of medicine including diagnostics, prognostication and clinical decision-support tools. Epilepsy is a common and disabling neurological disorder, however, management remains challenging in many cases, despite expanding therapeutic options. We present a systematic review protocol to explore the role of machine learning in the management of epilepsy. METHODS AND ANALYSIS: This protocol has been drafted with reference to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) for Protocols. A literature search will be conducted in databases including MEDLINE, Embase, Scopus and Web of Science. A PRISMA flow chart will be constructed to summarise the study workflow. As the scope of this review is the clinical application of machine learning, the selection of papers will be focused on studies directly related to clinical decision-making in management of epilepsy, specifically the prediction of response to antiseizure medications, development of drug-resistant epilepsy, and epilepsy surgery and neuromodulation outcomes. Data will be extracted following the CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies checklist. Prediction model Risk Of Bias ASsessment Tool will be used for the quality assessment of the included studies. Syntheses of quantitative data will be presented in narrative format. ETHICS AND DISSEMINATION: As this study is a systematic review which does not involve patients or animals, ethics approval is not required. The results of the systematic review will be submitted to peer-review journals for publication and presented in academic conferences. PROSPERO REGISTRATION NUMBER: CRD42023442156.

Prevalence and predictors of depressive symptoms among caregivers of patients with epilepsy in Hong Kong.

STUDY AIM: To estimate the prevalence of high caregiving burden and depressive symptoms among caregivers (CG) of patients with epilepsy (PWEs) in Hong Kong and identify risk and protective factors for both outcomes after the Model of Stress and Carer Burden (MSCB). METHODS: This cross-sectional study recruited participants from local epilepsy clinics to complete a 15-minute survey on a tablet. Caregiving burden (CB) was assessed using the 4-item Zarit Caregiver Burden Interview. Depressive symptoms were assessed using the 2-item Patient Health Questionnaire. Family functioning was assessed using the Short-Form Family Assessment Device General Functioning Subscale. Sociodemographic data of the caregivers and clinical data of the PWE they cared for were described. Hierarchical logistic regression models were used to analyze the factors associated with the outcomes. RESULTS: A hundred and fifty-one CGs of PWEs were recruited for this study. The prevalence of high caregiving burden (ZBI-4 > 7) for CGs of PWEs was 58.9% (n = 89), whereas the prevalence of high depressive symptoms (PHQ2 > 2) was 23.8% (n = 36). Hierarchical logistic regression analysis revealed that entering patient characteristics and care situations did not enhance the model's predictability. In the full model, a high perceived CB was a risk factor for elevated depressive symptoms. Good physical health protects against depressive symptoms. CONCLUSIONS: Among caregivers of PWE in Hong Kong, a high perceived caregiving burden was a risk factor for elevated depressive symptoms; however, the clinical characteristics of the PWEs were not. Self-reported physical health is a protective factor against increased depressive symptoms.

Inhibitory dysfunction may cause prospective memory impairment in temporal lobe epilepsy (TLE) patients: an event-related potential study.

Introduction: Prospective memory (PM) is the ability to remember future intentions, and PM function is closely related to independence in daily life, particularly in patients with temporal lobe epilepsy (TLE). As PM involves various cognitive components of attention, working memory, inhibition and other executive functions, this study investigated how TLE may affect PM components and the underlying neural mechanisms. Methods: Sixty-four subjects were recruited, including 20 refractory TLE patients, 18 well-controlled TLE patients and 26 age-matched healthy controls. A set of neuropsychological tests was administered to assess specific brain functions. An event-related potential (ERP) task was used to further explore how PM and its components would be differentially affected in the two TLE types. Results: Our findings revealed that: (1) refractory TLE patients scored lower than the healthy controls in the digit span, Verbal Fluency Test and Symbol Digit Modalities Test; (2) refractory TLE patients exhibited impaired PM performance and reduced prospective positivity amplitudes over the frontal, central and parietal regions in ERP experiments when compared to the healthy controls; and (3) decreased P3 amplitudes in the nogo trials were observed over the frontal-central sites in refractory but not in well-controlled TLE patients. Discussion: To our knowledge, this is the first ERP study on PM that has specifically identified PM impairment in refractory but not in well-controlled TLE patients. Our finding of double dissociation in PM components suggests that inhibition dysfunction may be the main reason for PM deficit in refractory TLE patients. The present results have clinical implications for neuropsychological rehabilitation in TLE patients.

Association between the risk of seizure and COVID-19 vaccinations: A self-controlled case-series study.

OBJECTIVE: The risk of seizure following BNT162b2 and CoronaVac vaccinations has been sparsely investigated. This study aimed to evaluate this association. METHOD: Patients who had their first seizure-related hospitalization between February 23, 2021 and January 31, 2022, were identified in Hong Kong. All seizure episodes happening on the day of vaccination (day 0) were excluded, since clinicians validated that most of the cases on day 0 were syncopal episodes. Within-individual comparison using a modified self-controlled case series analysis was applied to estimate the incidence rate ratio (IRR) with 95% confidence intervals (CIs) of seizure using conditional Poisson regression. RESULTS: We identified 1656 individuals who had their first seizure-related hospitalization (BNT162b2: 426; CoronaVac: 263; unvaccinated: 967) within the observation period. The incidence of seizure was 1.04 (95% CI .80-1.33) and 1.11 (95% CI .80-1.50) per 100 000 doses of BNT162b2 and CoronaVac administered, respectively. Sixteen and 17 individuals, respectively, received a second dose after having a first seizure within 28 days after the first dose of BNT162b2 and CoronaVac vaccinations. None had recurrent seizures after the second dose. There was no increased risk during day 1-6 after the first (BNT162b2: IRR = 1.39, 95% CI = .75-2.58; CoronaVac: IRR = 1.19, 95% CI = .50-2.83) and second doses (BNT162b2: IRR = 1.36, 95% CI = .72-2.57; CoronaVac: IRR = .71, 95% CI = .22-2.30) of vaccinations. During 7-13, 14-20, and 21-27 days post-vaccination, no association was observed for either vaccine. SIGNIFICANCE: The findings demonstrated no increased risk of seizure following BNT162b2 and CoronaVac vaccinations. Future studies will be warranted to evaluate the risk of seizure following COVID-19 vaccinations in different populations, with subsequent doses to ensure the generalizability.

Employment among people with epilepsy in Hong Kong.

BACKGROUND: People with epilepsy (PWE) face difficulties in employment. Hong Kong depends heavily on tertiary industry and enjoys a low unemployment rate. However, there have been rare reports on employment of PWE in Hong Kong. We aimed at (1) investigating the employment status among PWE; (2) correlating demographic and clinical factors with employment status of PWE; and (3) describing the self-perceived impact of epilepsy on employment and their correlations with employment status. METHOD: This was a single center cross-sectional study conducted in 2019. Adult with epilepsy but without intellectual disability of year age 16-65 were recruited. Homemakers and retired persons were excluded. A questionnaire with two parts was given to each patient. The first part focused on objective data about employment. The second part focused on self-perception on the impact of epilepsy on employment. Responders expressed their opinions in 5-point Likert scale. Clinical data were retrieved from the computerized medical record system for interpretation. RESULTS: A total of 138 PWE were recruited. Unemployment rate among the PWE was 33%, which was much higher than the general population. Low education levels, drug-resistant epilepsy, psychiatric comorbidities, and high Charlson Comorbidity Index were correlated to unemployment in PWE. Unemployed respondents significantly more tend to regard that lack of education, stigma of epilepsy, and seizure frequency were main hurdles in employment. CONCLUSIONS: Unemployment is a severe social problem among PWE in Hong Kong. Various objective clinical and demographic factors correlated with unemployment. Work beliefs of a patient may also correlate with the employment status.

Seasonality in the incidence of anti-GQ1b antibody syndrome-A territory-wide study.

AIMS: To investigate any seasonality in the incidence of anti-GQ1b antibody syndrome (AGS). METHODS: We conducted a retrospective observational study in all hospitalized patients in local public hospitals from January 2013 to December 2018. AGS was defined by hospitalized patients with positive serum anti-GQ1b IgG, presumably encompassing Miller-Fisher syndrome, Bickerstaff brainstem encephalitis and Guillain-Barré syndrome (GBS) variants. GBS cases were retrieved from the computerized database by diagnostic label. Campylobacter jejuni infection (CJI) injection was identified by positive stool culture. Monthly incidence rates of AGS, GBS and CJI were calculated. Poisson and negative binomial regression models with long-term time trend were fitted to characterize the seasonal pattern. RESULTS: A total of 237, 572 and 2434 cases of AGS, GBS and CJI were identified, respectively, in a population of 7.3 million. The annual incidence rate of AGS was 0.54 per 100,000 person-years. AGS was demonstrated to have an annual peak in the spring season, from March to April, which was congruent with that of GBS and slightly lagged the annual peak of CJI from February to March (likelihood ratio tests all p < .001 for the seasonal terms). CONCLUSION: The incidence of AGS peaks in springtime, which is congruent with that of GBS and lags around one month after that of CJI. We demonstrated that AGS has a clear seasonality in occurrence.

Prevalence and healthcare utilization of rare neurological diseases in Hong Kong: 2014-2018.

BACKGROUND: There has been increasing attention focused on the epidemiology of rare diseases (RDs) in recent years. Rare neurological diseases (RNDs) constitute a significant proportion of RDs; however, relevant research is still lacking. METHODS: A list of ICD-10 codes corresponding to RNDs was compiled using adaptations from the Orphanet Classification of Rare Diseases, and classified into rare epilepsy, movement-related, neurocutaneous, neuroimmune, neurometabolic and neurodegenerative, neuromuscular and other RNDs. Using the Clinical Data Analysis and Reporting System, which holds public hospital healthcare records of Hong Kong anonymously, we calculated the prevalence and healthcare utilization of RND patients between 2014 and 2018. The list of RNDs was also used to review relevant pharmacological trials within the International Clinical Trials Registry Platform between 2009 and 2018. RESULTS: The prevalence of RNDs in Hong Kong is 3.6 in 1,000 individuals. Patients with RNDs had frequent emergency department, outpatient and inpatient healthcare utilization. The average annual cost per patient is estimated at HKD 182,075 (€ 19,688). Different categories of RNDs showed different patterns of healthcare utilization. Moreover, there were only 677 RND-related pharmacological trials during the study period, and no trial was found for 78% of RNDs. CONCLUSIONS: This is one of the first population studies on the prevalence and healthcare utilization patterns of RNDs, with comprehensive reviews of RND-related pharmacological research. It shows high healthcare utilization rates among patients with RNDs, as well as a wide research gap in many RNDs. We call for better attention and tailored healthcare for these patients.

Changes in pediatric seizure-related emergency department attendances during COVID-19 - A territory-wide observational study.

A territory-wide retrospective observational study was conducted in Hong Kong between January 23 to April 22, 2020 to demonstrate changes in pediatric seizure-related accident and emergency department (A&E) visits during the COVID-19 pandemic. Parallel periods from 2015 to 2019 were used as control. All-cause A&E attendances in all paediatric age groups decreased significantly during the study period. Seizure-related attendances decreased across all pediatric age-groups in 2020 (RR 0.379, 95% CI 0.245-0.588), with a disproportionately large decrease in the 0-6 years age group (RR 0.303, 95% CI 0.174-0.526) compared with the 7-18 years age group (RR 0.534, 95% CI 0.393-0.719). Decrease in RTI-related A&E attendances was also more drastic in the 0-6 age group. The two time trends are congruent in the 0-6 years but not the 7-18 years age group. Such a trend is suggestive of the usefulness of infection control measures in seizure prevention, especially amongst young children.

Impact of COVID-19 on seizure-related emergency attendances and hospital admissions - A territory-wide observational study.

This is a territory-wide study to investigate the impact of coronavirus disease 2019 (COVID-19) pandemic on Accident and Emergency Department (A&E) attendances and acute ward admissions for seizures. Adult patients who presented to the A&E with seizures from January 23, 2020 to March 24, 2020 (study period) were included and compared with parallel intervals from 2015 to 2019 (control periods). Preexisting time trend in control periods and potential changes during COVID-19 were analyzed by Poisson, negative and logistic regression models. Accident and Emergency Department attendances and ward admissions for seizures decreased significantly during the COVID-19 pandemic. A total of 319 and 230 recorded ward admissions and A&E attendances for seizures were identified during the study period in 2020, compared with 494 and 343 per annum, respectively in the control periods. The ratio of acute ward admission per A&E attendance for seizures did not change significantly. Intensive care utility and mortality rates remained stable. For some patients, delaying medical attention due to fear of nosocomial COVID-19 cross-infection may lead to severe or even life-threatening consequences. This change in medical help-seeking behavior calls for new medical care models to meet the service gap. Education to patients with epilepsy and their caregivers is of utmost importance during this pandemic.

Efficacy upon 12-weeks after achievement of maximal dose and tolerability of lacosamide as an adjunctive therapy in epilepsy: Real world clinical experience.

Lacosamide (LCM) is a new generation antiepileptic drug. It has only been available in Asia in recent years. A retrospective study at two hospitals in Hong Kong was performed to investigate the post-marketing efficacy and tolerability of the drug. A total of 81 subjects were recruited, among which 88% had drug-resistant epilepsy. The most common type of epilepsy was focal with unknown etiology. All patients used LCM as adjunctive therapy. The 50% responder rate was 42% at 12 weeks after achievement of maximal dose of LCM. No specific factor correlated with responsiveness including concomitant enzyme-inducing or sodium channel blocking anticonvulsants. Withdrawal rate within first 12 weeks after drug initiation was 14% while that at any time upon follow-up was 23%. Two cases of uncommon adverse reaction of myoclonus were also reported. The mechanism was postulated to be the sodium channel inhibiting action of LCM. Our study has shown LCM to have comparable efficacy and tolerability in post-marketing experience when compared with the landmark randomized controlled trials.

"Triggers" for referral to neurology palliative care service.

In neurology, many diseases are still incurable and have a fatal outcome. Palliative care (PC) definitely has a role in neurology. We discuss the screening criteria for PC referral, known as 'triggers' in different neurological diseases. Different clinical settings including in-patient and out-patient are covered. We hope this review can remind clinicians to involve PC in the appropriate patient group. Further clinical studies are expected to validate the triggers and trajectories of various neurological diseases.

Prognostic indicators of neuromuscular disorders for palliative care referral.

BACKGROUND: Most of the neuromuscular disorders (NMDs) have poor prognosis and lead to various symptoms amendable to palliative care. However, the suitable time of referral is uncertain. METHODS: A retrospective study was conducted to describe the trajectories of NMDs. Early death within one year after NMD diagnosis was set as the outcome. Total of 86 adult NMD patients were recruited in a university hospital. Demographic variable including gender, age at diagnosis and early-onset symptoms including dyspnea, dysphagia, loss of mobility, constipation, mood and sleep disorders, and pain were correlated with the outcome. Prediction models for early death were tested. RESULTS: Age at NMD diagnosis, early-onsets dyspnea, dysphagia, constipation and impaired mobility were found to have statistically significant correlation with early death. A prediction model consisted of these four factors had area under receiver operating characteristic (ROC) curve of 0.919. CONCLUSIONS: Elder age at NMD diagnosis, early-onset dyspnea, dysphagia, constipation and impaired mobility within the first year after NMD diagnosis may predict mortality within the first year after diagnosis. It may provide guidance to clinicians for early palliative care referral in this patient group.

The role of gene variants in the pathogenesis of neurodegenerative disorders as revealed by next generation sequencing studies: a review.

The clinical diagnosis of neurodegenerative disorders based on phenotype is difficult in heterogeneous conditions with overlapping symptoms. It does not take into account the disease etiology or the highly variable clinical course even amongst patients diagnosed with the same disorder. The advent of next generation sequencing (NGS) has allowed for a system-wide, unbiased approach to identify all gene variants in the genome simultaneously. With the plethora of new genes being identified, genetic rather than phenotype-based classification of Mendelian diseases such as spinocerebellar ataxia (SCA), hereditary spastic paraplegia (HSP) and Charcot-Marie-Tooth disease (CMT) has become widely accepted. It has also become clear that gene variants play a role in common and predominantly sporadic neurodegenerative diseases such as Parkinson's disease (PD) and amyotrophic lateral sclerosis (ALS). The observation of pleiotropy has emerged, with mutations in the same gene giving rise to diverse phenotypes, which further increases the complexity of phenotype-genotype correlation. Possible mechanisms of pleiotropy include different downstream effects of different mutations in the same gene, presence of modifier genes, and oligogenic inheritance. Future directions include development of bioinformatics tools and establishment of more extensive public genotype/phenotype databases to better distinguish deleterious gene variants from benign polymorphisms, translation of genetic findings into pathogenic mechanisms through in-vitro and in-vivo studies, and ultimately finding disease-modifying therapies for neurodegenerative disorders.

Directed Differentiation of Human Bone Marrow Stromal Cells to Fate-Committed Schwann Cells.

Our ultimate goal of in vitro derivation of Schwann cells (SCs) from adult bone marrow stromal cells (BMSCs) is such that they may be used autologously to assist post-traumatic nerve regeneration. Existing protocols for derivation of SC-like cells from BMSCs fall short in the stability of the acquired phenotype and the functional capacity to myelinate axons. Our experiments indicated that neuro-ectodermal progenitor cells among the human hBMSCs could be selectively expanded and then induced to differentiate into SC-like cells. Co-culture of the SC-like cells with embryonic dorsal root ganglion neurons facilitated contact-mediated signaling that accomplished the switch to fate-committed SCs. Microarray analysis and in vitro myelination provided evidence that the human BMSC-derived SCs were functionally mature. This was reinforced by repair and myelination phenotypes observable in vivo with the derived SCs seeded into a nerve guide as an implant across a critical gap in a rat model of sciatic nerve injury.