Noe, Katherine H.

ABSTRACT Deep brain stimulation (DBS) of the anterior nucleus of the thalamus (ANT) can reduce seizures in patients with drug-resistant epilepsy. However, seizure freedom is rare, and there is no early response biomarker to predict long-term seizure suppression. To evaluate a short-term biomarker of therapeutic response to DBS, we studied 18-minute trials of ANT stimulation in patients with drug-resistant epilepsy undergoing intracranial electroencephalography. We compared a standard high-frequency periodic versus a non-periodic pulse pattern. We found that short-term ANT stimulation within the clinically used amplitude range preferentially altered the first 150 ms of anterograde ANT to anterior cingulate and superior frontal connectivity. Non-periodic ANT stimulation selectively modified early cortico-cortical evoked potentials more than periodic stimulation, and these changes were driven by the response phase. These pattern-specific changes in connectivity between the ANT and spatially restricted cortical areas suggest differing mechanisms and are a promising biomarker for patients being treated with ANT-DBS.

Association of Outpatient Follow-up with 30-day Readmission after Epilepsy or Seizure Discharge in Medicare Beneficiaries Aged 65 and Older Blank LJ, Van Hyfte G, Agarwal P, Mazumdar M, Jette J. Neurology . 2025;104(10):e213638. Background and Objectives: Older adults are expected to have higher readmission rates after seizure-related hospitalization. We sought to define the 30-day readmission rate for older adults after seizure hospitalization and to examine whether occurrence, timing, or specialization of follow-up with primary or neuro-related care is associated with lower readmission risk. Methods: This is a retrospective cohort study using 2016–2019 Medicare claims data, including adults aged older than 65 years hospitalized with a primary diagnosis of seizure/epilepsy. The primary outcome was readmission within 30 days. Exposure of interest was presence or absence of follow-up and specialty of the follow-up provider. Beneficiaries were followed from 90 days before admission to 30 days after discharge. We defined variables a priori based on literature/clinical judgment and used a least absolute shrinkage and selection operator (LASSO) method to determine factors that were contributing to the data's variance for inclusion in the final model. Results: Of 80,620 beneficiaries with admissions for seizure/epilepsy, 17.72% were readmitted within 30 days. Overall, 20.6% saw only primary care, 2.5% neurology only, 0.3% neurosurgery only, and 0.1% epilepsy only, and 5.4% had a combination of visits. Readmission rates differed by follow-up visit status: 22% readmission rate for those with no follow-up and only 6% with any health care visit. Among those with a visit, the readmission rates by specialty were as follows: 8% for primary care alone, 5% for neurology alone, 16% for neurosurgery alone, 1% for epileptology alone, and 2% for those who had seen a combination of these specialties. In our LASSO-selected multivariable model, outpatient follow-up was associated with lower odds of readmission: early (days 1–15) primary care visit (adjusted odds ratio [aOR] 0.49; 95% CI 0.45–0.52, p < .001); early neurology visit (aOR 0.39; 95% CI 0.33–0.46, p < .001); and later (days 16–30) neurosurgery visit (aOR 0.42; 95% CI 0.27–0.67, p < .001), later neurology visit (aOR 0.16; 95% CI 0.13–0.21, p < .001), or later primary care visit (aOR 0.16; 95% CI 0.14–0.17, p < .001), all associated with reduced odds of readmission. Discussion: We found high rates of readmission in older adults. Outpatient follow-up was associated with reduced odds of readmission. These findings reinforce the importance of discharge planning and suggest that ensuring outpatient follow-up with either primary care or neurology may be an easy intervention to reduce readmissions.

Abstract Purpose Cenobamate is approved for treating focal epilepsy and has potential for treating tumor‐related epilepsy (TRE) based on potential anti‐tumor effects, though there is little data on efficacy and safety in this patient population. Methods This was a retrospective, multi‐center study of patients with TRE who were prescribed cenobamate as adjunctive treatment between January 2020 and November 2024 at Mayo Clinic in Florida, Minnesota, and Arizona. Drug efficacy was assessed, and factors that were predictive of response to treatment were analyzed. Results A total of 17 patients were included, with 12 being drug resistant. The median seizure frequency for the entire group before starting cenobamate treatment was 12 (IQR = 4–64) per month, and after treatment at last follow‐up was 3 (IQR = 1–15) per month, with an overall 56.3% reduction. Four patients (23.5%) were seizure‐free, and 7 (41.1%) patients were responders (at least 50% seizure reduction). No adverse events were reported for patients on concurrent chemotherapy. Responders had significantly fewer seizures per month compared to non‐responders (5.72 ± 10.29 vs. 27.33 ± 30.17; p = .030). No clinical factors were predictive of responsiveness (or lack thereof) to cenobamate. Conclusions This study demonstrates that adjunctive therapy with cenobamate treatment for TRE is a feasible and potentially effective option. Clinical history, including drug resistance, did not predict drug efficacy. No adverse effects were reported in those treated with chemotherapy. Further studies are needed to validate treatment response and safety.

Risk of Perinatal and Maternal Morbidity and Mortality Among Pregnant Women With Epilepsy Razaz N, Igland J, Bjørk MH, Joseph KS, Dreier JW, Gilhus NE, Gissler M, Leinonen MK, Zoega H, Alvestad S, Christensen J, Tomson T. JAMA Neurol . 2024;81(9):985-995. Importance: Maternal epilepsy is associated with adverse pregnancy and neonatal outcomes. A better understanding of this condition and the associated risk of mortality and morbidity at the time of delivery could help reduce adverse outcomes. Objective: To determine the risk of severe maternal and perinatal morbidity and mortality among women with epilepsy. Design, Setting, Participants: This prospective population-based register study in Denmark, Finland, Iceland, Norway, and Sweden took place between January 1, 1996 and December 31, 2017. Data analysis was performed from August 2022 to November 2023. Participants included all singleton births at 22 weeks’ gestation or longer. Births with missing or invalid information on birth weight or gestational length were excluded. The study team identified 4 511 267 deliveries, of which 4 475 984 were to women without epilepsy and 35 283 to mothers with epilepsy. Exposure: Maternal epilepsy diagnosis recorded before childbirth. Prenatal exposure to antiseizure medication (ASM), defined as any maternal prescription filled from conception to childbirth, was also examined. Main Outcomes and Measures: Composite severe maternal morbidity and mortality occurring in pregnancy or within 42 days’ postpartum and composite severe neonatal morbidity (eg, neonatal convulsions) and perinatal mortality (ie, stillbirths and deaths) during the first 28 days of life. Multivariable generalized estimating equations with logit link were used to obtain adjusted odds ratio (aOR) and 95% confidence interval (CI). Results: The mean (SD) age at delivery for women in the epilepsy cohort was 29.9 (5.3) years. The rate of composite severe maternal morbidity and mortality was also higher in women with epilepsy compared with those without epilepsy (36.9 vs 25.4 per 1000 deliveries). Women with epilepsy also had a significantly higher risk of death (0.23 deaths per 1000 deliveries) compared with women without epilepsy (0.05 deaths per 1000 deliveries) with an aOR of 3.86 (95% CI, 1.48-8.10). In particular, maternal epilepsy was associated with increased odds of severe preeclampsia, embolism, disseminated intravascular coagulation or shock, cerebrovascular events, and severe mental health conditions. Fetuses and infants of women with epilepsy were at elevated odds of mortality (aOR, 1.20; 95% CI, 1.05-1.38) and severe neonatal morbidity (aOR, 1.48; 95% CI, 1.40-1.56). In analyses restricted to women with epilepsy, women exposed to ASM compared with those unexposed had higher odds of severe maternal morbidity (aOR, 1.24; 95% CI, 1.10-1.48) and their neonates had an increased odd of mortality and severe morbidity (aOR, 1.37; 95% CI, 1.23-1.52). Conclusion and Relevance: This multinational study shows that women with epilepsy were at considerably higher risk of severe maternal and perinatal outcomes and increased risk of death during pregnancy and postpartum. Maternal epilepsy and maternal use of ASM were associated with increased maternal morbidity and perinatal mortality and morbidity.

AbstractPurposeBy assessing longitudinal associations between COVID‐19 census burdens and hospital characteristics, such as bed size and critical access status, we can explore whether pandemic‐era hospital quality benchmarking requires risk‐adjustment or stratification for hospital‐level characteristics.MethodsWe used hospital‐level data from the US Department of Health and Human Services including weekly total hospital and COVID‐19 censuses from August 2020 to August 2023 and the 2021 American Hospital Association survey. We calculated weekly percentages of total adult hospital beds containing COVID‐19 patients. We then calculated the number of weeks each hospital spent at Extreme (≥20% of beds occupied by COVID‐19 patients), High (10%–19%), Moderate (5%–9%), and Low (<5%) COVID‐19 stress. We assessed longitudinal hospital‐level COVID‐19 stress, stratified by 15 hospital characteristics including joint commission accreditation, bed size, teaching status, critical access hospital status, and core‐based statistical area (CBSA) rurality.FindingsAmong n = 2582 US hospitals, the median(IQR) weekly percentage of hospital capacity occupied by COVID‐19 patients was 6.7%(3.6%–13.0%). 80,268/213,383 (38%) hospital‐weeks experienced Low COVID‐19 census stress, 28% Moderate stress, 22% High stress, and 12% Extreme stress. COVID‐19 census burdens were similar across most hospital characteristics, but were significantly greater for critical access hospitals.ConclusionsUS hospitals experienced similar COVID‐19 census burdens across multiple institutional characteristics. Evidence‐based inclusion of pandemic‐era outcomes in hospital quality reporting may not require significant hospital‐level risk‐adjustment or stratification, with the exception of rural or critical access hospitals, which experienced differentially greater COVID‐19 census burdens and may merit hospital‐level risk‐adjustment considerations.

Abstract Objective Hospital acquired infections (HAIs) are one of the top 10 leading causes of death within the United States. While current standard of HAI risk prediction utilizes only a narrow set of predefined clinical variables, we propose a graph convolutional neural network (GNN)-based model which incorporates a wide variety of clinical features. Materials and Methods Our GNN-based model defines patients’ similarity based on comprehensive clinical history and demographics and predicts all types of HAI rather than focusing on a single subtype. An HAI model was trained on 38 327 unique hospitalizations while a distinct model for surgical site infection (SSI) prediction was trained on 18 609 hospitalization. Both models were tested internally and externally on a geographically disparate site with varying infection rates. Results The proposed approach outperformed all baselines (single-modality models and length-of-stay [LoS]) with achieved area under the receiver operating characteristics of 0.86 [0.84–0.88] and 0.79 [0.75–0.83] (HAI), and 0.79 [0.75–0.83] and 0.76 [0.71–0.76] (SSI) for internal and external testing. Cost-effective analysis shows that the GNN modeling dominated the standard LoS model strategy on the basis of lower mean costs ($1651 vs $1915). Discussion The proposed HAI risk prediction model can estimate individualized risk of infection for patient by taking into account not only the patient’s clinical features, but also clinical features of similar patients as indicated by edges of the patients’ graph. Conclusions The proposed model could allow prevention or earlier detection of HAI, which in turn could decrease hospital LoS and associated mortality, and ultimately reduce the healthcare cost.