Faculty Bibliography

BACKGROUND:There is practice heterogeneity in the use, type, and duration of prophylactic antiseizure medications (ASM) in patients hospitalized with acute nontraumatic intracerebral hemorrhage (ICH). METHODS:We conducted a systematic review and meta-analysis assessing ASM primary prophylaxis in adults hospitalized with acute nontraumatic ICH. The following population, intervention, comparison, and outcome (PICO) questions were assessed: (1) Should ASM versus no ASM be used in patients with acute ICH with no history of clinical or electrographic seizures? (2) If an ASM is used, should levetiracetam (LEV) or phenytoin/fosphenytoin (PHT/fPHT) be preferentially used? and (3) If an ASM is used, should a long (> 7 days) versus short (≤ 7 days) duration of prophylaxis be used? The main outcomes assessed were early seizure (≤ 14 days), late seizures (> 14 days), adverse events, mortality, and functional and cognitive outcomes. We used Grading of Recommendations Assessment, Development, and Evaluation methodology to generate recommendations. RESULTS:The initial literature search yielded 1,988 articles, and 15 formed the basis of the recommendations. PICO 1: although there was no significant impact of ASM on the outcomes of early or late seizure or mortality, meta-analyses demonstrated increased adverse events and higher relative risk of poor functional outcomes at 90 days with prophylactic ASM use. PICO 2: we did not detect any significant positive or negative effect of PHT/fPHT compared to LEV for early seizures or adverse events, although point estimates tended to favor LEV. PICO 3: based on one decision analysis, quality-adjusted life-years were increased with a shorter duration of ASM prophylaxis. CONCLUSIONS:We suggest avoidance of prophylactic ASM in hospitalized adult patients with acute nontraumatic ICH (weak recommendation, very low quality of evidence). If used, we suggest LEV over PHT/fPHT (weak recommendation, very low quality of evidence) for a short duration (≤ 7 days; weak recommendation, very low quality of evidence).

The α-synuclein seed amplification assay (αSyn-SAA) sensitively detects Lewy pathology, the amyloid state of α-synuclein, in the cerebrospinal fluid (CSF) of patients with Parkinson's disease (PD). The αSyn-SAA harnesses the physics of seeding, whereby a superconcentrated solution of recombinant α-synuclein lowers the thermodynamic threshold (nucleation barrier) for aggregated α-synuclein to act as a nucleation catalyst ("seed") to trigger the precipitation (nucleation) of monomeric α-synuclein into pathology. This laboratory setup increases the signal for identifying a catalyst if one is present in the tissue examined. The result is binary: positive, meaning precipitation occurred, and a catalyst is present, or negative, meaning no precipitation, therefore no catalyst. Since protein precipitation via seeding can only occur at a concentration many-fold higher than the human brain, laboratory-elicited seeding does not mean human brain seeding. We suggest that a positive αSyn-SAA reveals the presence of pathological α-synuclein but not the underlying etiology for the precipitation of monomeric α-synuclein into its pathological form. Thus, a positive αSyn-SAA supports a clinical diagnosis of PD but cannot inform disease pathogenesis, ascertain severity, predict the rate of progression, define biology or biological subtypes, or monitor treatment response.

The adoption of large language models (LLMs) in healthcare demands a careful analysis of their potential to spread false medical knowledge. Because LLMs ingest massive volumes of data from the open Internet during training, they are potentially exposed to unverified medical knowledge that may include deliberately planted misinformation. Here, we perform a threat assessment that simulates a data-poisoning attack against The Pile, a popular dataset used for LLM development. We find that replacement of just 0.001% of training tokens with medical misinformation results in harmful models more likely to propagate medical errors. Furthermore, we discover that corrupted models match the performance of their corruption-free counterparts on open-source benchmarks routinely used to evaluate medical LLMs. Using biomedical knowledge graphs to screen medical LLM outputs, we propose a harm mitigation strategy that captures 91.9% of harmful content (F1 = 85.7%). Our algorithm provides a unique method to validate stochastically generated LLM outputs against hard-coded relationships in knowledge graphs. In view of current calls for improved data provenance and transparent LLM development, we hope to raise awareness of emergent risks from LLMs trained indiscriminately on web-scraped data, particularly in healthcare where misinformation can potentially compromise patient safety.

Paired vagus nerve stimulation (VNS) has emerged as a promising strategy to potentiate recovery after neurological injury. This approach, which combines short bursts of electrical stimulation of the vagus nerve with rehabilitation exercises, received approval from the US Food and Drug Aministration in 2021 as the first neuromodulation-based therapy for chronic stroke. Because this treatment is increasingly implemented in clinical practice, there is a need to take stock of what we know about this approach and what we have yet to learn. Here, we provide a survey on the foundational basis of VNS therapy for stroke and offer insight into the mechanisms that underlie potentiated recovery, focusing on the principles of neuromodulatory reinforcement. We discuss the current state of observations regarding synaptic reorganization in motor networks that are enhanced by VNS, and we propose other prospective loci of neuromodulation that should be evaluated in the future. Finally, we highlight the future opportunities and challenges to be faced as this approach is increasingly translated to clinical use. Collectively, a clearer understanding of the mechanistic basis of VNS therapy may reveal ways to maximize its benefits.

BACKGROUND:α-Synuclein seed amplification assay on cerebrospinal fluid (CSF-αSyn-SAA) has shown high accuracy for Parkinson's disease (PD) diagnosis. The analysis of CSF-αSyn-SAA parameters may provide useful insight to dissect the heterogeneity of synucleinopathies. OBJECTIVE:To assess differences in CSF-αSyn-SAA amplification parameters in participants with PD stratified by rapid eye movement (REM) sleep behavior disorder (RBD), dysautonomia, GBA, and LRRK2 variants. METHODS:(T50), time to threshold (TTT), slope, and area under the curve (AUC). Sporadic PD (n = 371) was stratified according to RBD and dysautonomia (DysA) symptoms. Genetic PD included carriers of pathogenic variants of GBA (GBA-PD, n = 52) and LRRK2 (LRRK2-PD, n = 124) gene. RESULTS:CSF-αSyn-SAA was positive in 77% of LRRK2-PD, 92.3% of GBA-PD, and 93.8% of sporadic PD. The LRRK2-PD cohort showed longer T50 and TTT, and smaller AUC than GBA-PD (P = 0.029, P = 0.029, P = 0.016, respectively) and sporadic PD (P = 0.034, P = 0.033, P = 0.014, respectively). In the sporadic cohort, CSF-αSyn-SAA parameters were similar between PD with (n = 157) and without (n = 190) RBD, whereas participants with DysA (n = 193) presented shorter T50 (P = 0.026) and larger AUC (P = 0.029) than those without (n = 150). CONCLUSION/CONCLUSIONS:CSF-αSyn-SAA parameters vary across genetic and non-genetic PD subtypes at the group level. These differences are mostly driven by the presence of LRRK2 variants and DysA. Significant overlaps in the amplification parameter values exist between groups and limit their use at the individual level. Further studies are necessary to understand the mechanisms of CSF-αSyn-SAA parameter differences. © 2024 International Parkinson and Movement Disorder Society.

OBJECTIVE:This study aimed to externally validate the Memory Assessment Clinics Scale for Epilepsy (MAC-E), a brief self-report measure of subjective memory complaints in adults with epilepsy. METHODS:A cross-sectional study was conducted including adults with focal pharmacoresistant epilepsy from three Level 4 epilepsy centers in the U.S., who completed the MAC-E as part of a clinical neuropsychological evaluation. Confirmatory factor analysis was conducted, and goodness-of-fit criteria were calculated to assess model fit: comparative fit index (CFI), root mean square error of approximation (RMSEA), and standardized root mean residual (SRMR). Item response theory models were constructed, and Mokken analysis was used to assess discrimination and unidimensionality. Internal consistency was evaluated with McDonald's Omega. RESULTS:values for each of the 5 factors (0.58-0.91 and 0.34-0.82, respectively). MAC-E items demonstrated high levels of discrimination as well as the ability to evaluate across the entirety of each latent trait. Score responses were uniformly distributed across latent traits, and unidimensionality was established by factor (all H coefficients > 0.4). Internal consistency was high across factors (omega range: 0.77-0.88). CONCLUSIONS:Results of this study demonstrate good external validation of the MAC-E in an independent, multicenter cohort of adults with epilepsy. These findings provide further support that the MAC-E is a psychometrically valid, self-report instrument to assess every-day memory abilities in adults with epilepsy in both clinical and research settings.

OBJECTIVE:This study examines the American Headache Society First Contact-Headache in Primary Care program metrics to date in order to assess the program's reach and provide direction for future initiatives. BACKGROUND:Approximately 4 million primary care office visits annually are headache-specific encounters. Therefore, it is important that primary care providers are knowledgeable about headache management. Recognizing the need, the American Headache Society First Contact designed the comprehensive First Contact-Headache in Primary Care program with input from an advisory board comprised of a diverse group of physicians and advanced practice providers with backgrounds in family and internal medicine, pediatrics, obstetrics and gynecology, and neurology. This is the first study to assess the reach of the program and critically examine how to best meet the needs of clinicians and patients going forward. METHODS:We report descriptive statistics for the First Contact website metrics from October 2020 to June 2023 and grand rounds program data from May 2020 to December 2023. We also conducted a cross-sectional analysis of survey data from presentations conducted at two large national family medicine symposia, as well as a thematic analysis of the question: "Please indicate what areas of your practice could be enhanced or improved with additional education?" RESULTS:The First Contact program homepage was the second most visited page on the American Headache Society website (>100,000 views). A total of 20 podcast episodes were created for the program (>3500 plays). The First Contact program held 99 events (72 institutional grand rounds, 22 State-level meetings, and five national meetings), reaching >7000 clinicians. The institutional grand rounds and state-level meetings were held across 27 States and Washington D.C. Only 31.9% (30/94) of First Contact program events (excluding national meetings) occurred in the West census region, which has the fewest headache subspecialists and lowest headache subspecialist density in the United States. When examining survey data of participants who attended the two virtual national family medicine symposia (39.3% response rate, N = 636/1620), 85.7% (544/635) reported being "completely confident" or "very confident" in their ability to recognize and accurately diagnose patients presenting with a primary complaint of headache and 81.5% (517/634) reported being "completely confident" or "very confident" in their ability to develop evidence-based treatment plans that are tailored to the needs of individual patients. The use of diagnostic tools to recognize patients with migraine (60.4%, 384/636) and translating standards of care to the practice setting (42.5%, 270/636) were the most reported intended changes by participants. Most participants reported that program content was of clinical relevance and would improve their patients' outcomes (90.5% [571/631] and 90.6% [572/631], respectively). Over three-quarters (77.8%, 495/636) of participants reported areas of their practice that can be improved by additional education specifically regarding workflow, diagnosis, and management. CONCLUSION/CONCLUSIONS:This study evaluates one of the first national initiatives for primary care education. Data from the two First Contact Family Medicine national symposia indicate the program is generally well received with most participants reporting improved confidence and intention to implement key changes in practice to improve care for patients with headache; however, there remain areas of exploration for education that could further enhance participant experience and expand the reach of the initiatives. Areas for future programming include continued education on multifactorial approaches to headache treatment and suggestions for addressing cost, insurance, and time constraints. Also, future work may examine where the First Contact program might focus initiatives based on specific areas of need in headache care, such as geographic "desert" areas, racial and ethnic disparities, and uninsured/underinsured populations.