To evaluate the performance and reference reliability of three large language models in neurology using a national board examination framework. A total of 803 validated multiple-choice questions from Turkish National Neurology Board Examinations (2015-2024) were administered to ChatGPT Plus, Gemini Advanced, and Microsoft Copilot Pro using a standardized prompt requiring an answer and a supporting reference. Model performance was compared with overall examinee performance and analyzed by neurological subspecialty, question type, and presence of visual content. References provided for incorrectly answered questions were independently evaluated by three board certified neurologists. Mean accuracy rates were 87% for Gemini, 86% for Copilot, and 85% for ChatGPT, significantly outperforming the human examinee average of 65% (p < 0.001), with no significant differences among models. Accuracy did not differ by question type or neurological subspecialty. All models outperformed examinees on non-visual questions, whereas no performance advantage was observed for visually based items. Reference evaluation revealed substantial limitations: ChatGPT frequently provided insufficient citations (39.6%), while fabricated references predominated in Gemini (53.0%) and Co-pilot (42.1%). Large language models demonstrate high and consistent accuracy on neurology board examination questions, with performance exceeding that of the average examinee. On visually based questions, accuracy was lower than for non-visual items, and the performance advantage over examinees disappeared. High rates of insufficient referencing indicate a clear need for expert oversight, supporting the use of LLMs as complementary tools in neurology education rather than autonomous sources of clinical or academic authority.
Operation selection in metabolic surgery is a complex decision making process led by a multidisciplinary team that integrates multiple anatomical, clinical, metabolic and psychosocial aspects. The ability of large language models (LLMs) has been proposed to provide capability to act as decision support tools, but their performance in replicating MDT level decision making in metabolic surgery remains uncertain. A retrospective pilot study was performed using anonymised data from 100 patients at a single high-volume UK NHS bariatric centre who underwent surgery up to August 2025. Preoperative demographic, anthropometric, clinical and psychosocial variables were extracted from electronic records. These were provided to ChatGPT-4 Auto using a single standardised prompt instructing the model to act collectively on behalf of the whole bariatric MDT and recommend the most appropriate metabolic operation. These recommendations were compared with formal MDT decisions and with the operation ultimately performed. Concordance was assessed using raw percentage agreement, Cohen's kappa and Stuart-Maxwell tests. ChatGPT demonstrated 70% concordance with MDT recommendations. Concordance between MDT and operation performed was 83%, while concordance between ChatGPT and operation performed was 63%. Agreement beyond chance between ChatGPT and MDT recommendations was low (Cohen's kappa 0.036) reflecting class imbalance. Stuart-Maxwell test showed no significant difference in marginal distribution between ChatGPT and MDT recommendations. Both ChatGPT and MDT recommended bypass procedures more frequently than ultimately performed. ChatGPT overall demonstrated moderate crude agreement with bariatric MDT decision making in a real world UK NHS cohort of patients. However, limited agreement beyond chance and influence of unmeasured human factors may preclude its use in an autonomous fashion. This study establishes the requirement for larger scale evaluation of LLM clinical reasoning and supports the future exploration of them as adjunctive rather than autonomous decision support tools in metabolic surgery.
Metabolic-bariatric surgery (MBS) is the most effective treatment for obesity and its associated metabolic complications. Long-term real-world outcome data from Asian populations remain limited, particularly for procedural comparisons and longitudinal metabolic trajectories. To evaluate weight loss outcomes, metabolic outcomes, and clinical predictors of treatment response following bariatric surgery in a large Asian cohort, with longitudinal follow-up up to five years in a selected subgroup of completers. This retrospective cohort study analyzed 1,478 adult patients who underwent primary bariatric surgery (laparoscopic sleeve gastrectomy [LSG], Roux-en-Y gastric bypass [RYGB], or one-anastomosis gastric bypass [OAGB]) between 2012 and 2022 at a single tertiary academic center in Malaysia. The primary outcome was percentage total weight loss (%TWL) at 1, 3, and 5 years, analyzed using an available-case approach. Pre-specified attrition analysis compared the characteristics of 5-year completers with non-completers. Secondary outcomes included longitudinal changes in glycated hemoglobin (HbA1c), blood pressure, and LDL cholesterol, with paired and cross-sectional analyses reported separately. Glycaemic outcome thresholds (HbA1c < 6.5% and < 5.7%) were derived from ADA 2021 criteria; formal remission could not be ascertained because antidiabetic medication data were not available. Procedure-specific 1-year metabolic outcomes were evaluated for LSG and RYGB. Multivariable linear regression explored predictors of 1-year %TWL. Mean %TWL was 20.80 ± 9.71% (n = 917) at 1 year, 25.14 ± 10.30% (n = 349) at 3 years, and 24.52 ± 10.98% (n = 107) at 5 years; the 5-year sample represented 7.2% of the original cohort and was enriched for older patients, those with diabetes and hypertension, and patients with greater 1-year %TWL (25.33% vs. 19.99% in non-completers, p < 0.001), indicating selective retention bias. Among diabetic patients with paired pre-/postoperative HbA1c (n = 35), mean 1-year reduction was 1.01% (p = 0.007); cross-sectional HbA1c among diabetic patients was 5.95 ± 1.08% at 1 year, with 80.9% achieving HbA1c < 6.5% and 47.8% < 5.7%. Significant blood pressure reductions occurred at 1 year in hypertensive patients across both LSG and RYGB subgroups. The regression model was weak (R² = 0.026); pre-existing diabetes (β = -1.83, p = 0.020) and certain non-Malay ethnicity were associated with lower 1-year %TWL as exploratory, hypothesis-generating signals warranting prospective validation. The overall complication rate was 1.7%; OAGB long-term outcomes could not be reliably assessed (5-year n = 3-4). Bariatric surgery in this large Malaysian cohort produced significant 1-year weight loss with a favorable safety profile. Three- and five-year data, drawn from a selected subgroup of completers and subject to attrition bias, suggest preserved weight loss in those engaged with follow-up but should not be interpreted as population-level durability claims. Diabetes and ethnicity findings from a weak predictive model are hypothesis-generating and require prospective validation.
Urbanization-driven, large-scale rural-to-urban migration in China has substantially modernized household energy use and reshaped air pollution exposure pathways. However, the magnitude, underlying drivers, and temporal evolution of the resulting health outcomes from combined indoor and outdoor air pollution remain insufficiently resolved. Here, by coupling a reconstruction of seven decades of migration (1949-2021) with nationwide household energy surveys, we quantify migration-attributable changes in integrated (indoor + outdoor) PM2.5 exposure and mortality. Relative to a no-migration counterfactual from 1949 onward, migration reduced integrated PM2.5 exposure by 27.6 μg·m-3 (15.7-39.4 as the 95% confidence interval) in 2019 and cumulatively avoided 2.3 (1.9-2.6) million deaths since 1949, with reductions in indoor exposure providing the dominant contribution. Arising from rural-urban disparities in household energy use, these health gains were pronounced even during intermediate stages of societal development and are likely to occur in other transitioning economies worldwide. Our study thus challenges the conventional view that rapid urbanization in developing economies is strictly detrimental to environmental health.
Social media has emerged as an important tool for disseminating information to medical students; however, their platform preferences and usage purposes remain unclear. We conducted a multicenter web-based survey of medical students from eight Japanese medical schools in 2025, assessing their overall social media use, platform preferences by purpose, daily usage time, and the perceived usefulness of social media for career decision-making. A total of 1,515 students participated in this study. YouTube (71.6%), Instagram (65.1%), and X (43.7%) were frequently used platforms, whereas Facebook (1.9%) was rarely used. Platform use differed by academic year: YouTube use was significantly higher among early-year students, whereas X use was significantly more common among advanced-year students (P = 0.043 and P < 0.001, respectively). Overall, 763 respondents (50.4%) perceived social media as useful for career decision-making, with X being used most frequently for this purpose, followed by YouTube and Instagram; this perception was significantly more common among the early-year students (P = 0.005). These findings provide valuable insights for educators and professional organizations seeking to engage medical students through social media. Understanding these trends is especially important as Japan faces a shortage of surgeons.
Background Nurses form the backbone of the healthcare system; their role is vital in healthcare delivery in terms of promotion, prevention, treatment, care, and rehabilitation. Nurse managers play a key role in patient care coordination and ensure quality nursing care. Apart from this, they also perform staff management, document administrative activities, and supervise hospital store activities. Efficient and effective nursing administration is essential for smooth functioning and safe patient care. Methods In an observational study in two phases, we estimated the time utilized by administrative nurses in different tasks. In the first phase, a questionnaire was developed for self-reporting various activities carried out by administrative nurses. In the second phase, activity sampling with 316 hours of observations and 15 344 nursing activities was recorded across the hospital to estimate actual time distribution and utilization of administrative nurses. Results Administrative nurses spent a major portion of their time in store management (28%), staff management (23%), and documentation (21%), and much less time for patient care supervision and teaching activities, i.e. 4%-8% of the total time available in a day. Conclusion We found that administrative nurses spent most of their time in three major activities: store management, staff management, and documentation. Their roles may need to be reviewed for them to spend more time in supervision of patient care and teaching activities.
To determine whether there is an altered risk of specific cancers among individuals with Fuchs Endothelial Corneal Dystrophy (FECD). Retrospective, case-control study using data from the Utah Population Database, Utah Cancer Registry, and associated records. Cases were defined as individuals ≥40 years with a diagnosis of FECD. Controls were matched approximately 3:1 with cases on birth year, sex, whether born in Utah, and duration of follow-up in Utah. Twenty-two types/locations of cancer that were diagnosed between 1996 and 2022 were recorded. Cancer risk models were calculated using mixed-effect logistic regression, with adjustments for obesity, diabetes, tobacco use, race, ethnicity, and sex (except for sex-specific cancers). The main outcome measure was the odds of specific cancer diagnoses among FECD cases compared with matched controls. A total of 4129 FECD cases and 12,371 controls were studied in the final analysis. A total of 885 (21.4%) FECD cases and 2514 (20.3%) controls were diagnosed with any cancer (P = 0.126). After adjusting for covariates, FECD cases did not have an altered likelihood of having a diagnosis of cancer overall (OR: 1.06; 95% CI, 0.97-1.16; P = 0.174), or according to any of the specific cancer sites/subtypes. There was a slightly higher likelihood of thyroid cancer among FECD cases (OR: 1.55; 95% CI, 1.00-2.38; P = 0.048) and prostate cancer among male FECD cases (OR: 1.20; 95% CI, 1.01-1.43; P = 0.036) that was not statistically significant after accounting for multiple comparisons. Individuals with FECD did not have a significantly altered risk of any of the studied cancers.
暂无摘要(点击查看详情)
Aluminum nitride (AlN) is an ultrawide-bandgap semiconductor whose large intrinsic band gap limits low-energy interband optical absorption. This work compares Ce substitution at the Al site, C substitution at the N site, and Ce-C co-substitution in wurtzite AlN to clarify how Ce 4f/5d states and C 2p states modify the local structure, band-edge electronic states, and optical response. Among five neutral CeAl-CN configurations considered in a 2 × 2 × 2 supercell, the nearest-neighbor Ce-C pair has the lowest total energy, with the other configurations lying 6.6-17.7 meV higher. The selected Ce-C pair also has a negative binding energy of -3.26 eV relative to the corresponding isolated single-substitution reference supercells. Structural relaxation shows lattice expansion after substitution, with the largest volume increase of 7.93% obtained for AlN:Ce-C. The calculated band gap of intrinsic AlN is 4.19 eV, whereas the effective electronic gaps of AlN:C, AlN:Ce, and AlN:Ce-C are 3.59, 2.18, and 2.04 eV, respectively. Since AlN:Ce-C is only 0.14 eV smaller in effective gap than AlN:Ce, the role of Ce-C co-substitution is interpreted mainly through defect-pair energetics, orbital redistribution, and local population changes rather than as a large additional band-gap narrowing. DOS/PDOS and population analyses show that C 2p states mainly modify the occupied valence-edge region, Ce 4f/5d states contribute near the band-edge and conduction-band regions, and Ce-C co-substitution induces finite spin population on C together with a nonzero Ce-C bond population. The scissors-corrected optical spectra show increased low-frequency dielectric and refractive responses, with ε₁(0)/n(0) changing from 3.805/1.951 for intrinsic AlN to 4.449/2.109 for AlN:Ce-C, while AlN:C gives the largest low-frequency values of 4.530/2.128. Additional low-energy interband absorption and attenuation features appear especially in AlN:C and AlN:Ce-C, associated with defect-related transitions involving C 2p and Ce-related states. These changes are interpreted as calculated interband optical-response redistribution within the present neutral-defect supercell framework, not as direct evidence of device-level optoelectronic performance. First-principles calculations were performed using the CASTEP module in Materials Studio. The exchange-correlation interaction was described using the generalized gradient approximation with the Perdew-Burke-Ernzerhof functional, and on-the-fly-generated ultrasoft pseudopotentials were employed. Intrinsic AlN, AlN:Ce, AlN:C, and AlN:Ce-C were constructed from a fully relaxed 2 × 2 × 2 wurtzite AlN supercell containing 32 atoms. Single-doped models were constructed by replacing one Al or one N atom, corresponding to 6.25% substitution on the relevant sublattice. The co-doped model contains one CeAl-CN pair, corresponding to xCe = yC = 0.0625 in Al1-xCexN1-yCy, or a combined sublattice substitution level of 12.5%. A plane-wave cutoff energy of 700 eV and a 4 × 4 × 3 k-point mesh were used. Intrinsic AlN was calculated using non-spin-polarized GGA-PBE, C-doped AlN using spin-polarized GGA-PBE, and the Ce-containing systems using spin-polarized GGA-PBE + U. A Hubbard U correction of 5 eV was applied only to the Ce 4f orbitals. Band structures, TDOS/PDOS, Mulliken and Hirshfeld population analyses, representative bond populations, and optical properties were calculated. The dielectric function, complex refractive index, absorption coefficient, and reflectivity were obtained within the linear-response framework using a 2.01 eV scissors correction only for optical-property calculations.
Despite the great success that Kohn-Sham density functional theory (KS-DFT) has achieved, the delocalization error remains a major challenge for commonly used density functional approximations, resulting in systematic errors in ionization energies, electron affinities, band structures, and charge distributions. A recently developed localized orbital scaling correction (LOSC) method, namely, linear response LOSC (lrLOSC), addresses these challenges by incorporating a functional correction that includes the screening effect and orbital localization within the LOSC framework. The method has been shown to provide accurate descriptions of bulk systems and core-level binding energies in small molecular systems. In this work, we extend the applicability of lrLOSC to a broader range of molecular systems, spanning various sizes, with a focus on the corrections to valence orbital energies and total energies. To enable the calculation of large chemical systems, we developed an efficient implementation of lrLOSC with computational costs comparable with standard KS-DFT calculations. Numerical results show that while screening provides modest improvements for small molecules, it becomes critical for achieving high accuracy in larger molecules, from linear to three-dimensional systems. With the screening effect being well captured in a unified way, lrLOSC provides accurate descriptions for a wide range of chemical systems, including organic molecular systems of varying sizes and transition-metal oxide complexes, establishing it as a powerful tool for enhancing the reliability of computational simulations of chemical systems.
Both glucagon-like peptide-1 (GLP-1) receptor agonists and bariatric surgery (BS) are interventions that result in weight loss prior to total knee arthroplasty (TKA). We sought to compare 90-day postoperative complications between TKA patients with a history of GLP-1 use versus those who underwent BS prior to TKA. Patients undergoing primary TKA for osteoarthritis from 2005 to 2025 were identified from the TriNetX Research Network. Patients who underwent BS 6 months to 1 year prior to TKA were propensity score-matched 1:1 to patients who initiated GLP-1 therapy during this same time period based on age, sex, race, chronic kidney disease, rheumatoid arthritis, hypertension, heart disease, diabetes and body mass index (BMI) measured 6 weeks prior to weight-loss intervention. A total of 1650 patients were included in each cohort with 90-day follow-up. The risk of various 90-day complications associated with each weight-loss intervention was compared utilizing univariate analyses. Rates of individual medical and surgical complications were largely comparable between cohorts; however, compared with the BS group, the GLP-1 cohort had significantly lower rates of deep vein thrombosis/pulmonary embolus (2.18% vs. 4.42%, p < 0.001), transfusion (0.30% vs. 1.03%, p = 0.016), any surgical complication (1.64% vs. 3.45%, p = 0.001), any medical complication (9.09% vs. 12.30%, p = 0.003) and readmission/emergency department visit (21.33% vs. 25.33%, p = 0.007). Compared to obese controls without intervention prior to TKA, both the BS and GLP-1 cohorts had comparable rates of all medical and surgical complications. There were largely comparable rates of complications between patients who underwent BS and those who initiated GLP-1 agonists prior to TKA, although GLP-1 use was associated with lower rates of select complications. Patients who choose GLP-1 agonists or BS can expect similar complication rates and comparable outcomes relative to obese patients who elect for no preoperative intervention prior to TKA. Level III, retrospective cohort study.
Caregivers of individuals with spinal cord injury (SCI) frequently assume demanding roles immediately after SCI, placing them at risk for burden, depression, and diminished well-being. Few interventions have been tailored to support caregivers during this critical initial transition period. This randomized clinical trial developed and tested the Transition Assistance Program (TAP), which provides structured telehealth support targeting caregiver psychosocial adjustment and care quality as they learn to be caregivers for the first time. The study recruited caregiver-care recipient dyads from two acute SCI units at an academic medical center and a Veterans Affairs hospital. Dyads (N = 31 completing follow-up) were randomized to TAP or usual care. The TAP consisted of five sessions (one in person before discharge, four via telehealth-to-home) combining structured education, a caregiver guidebook, and supportive problem-solving. Primary outcomes - caregiving quality, depression, relationship satisfaction, and positive affect and well-being - were assessed at baseline, 2, and 4 months post-discharge. Caregivers in the TAP group reported higher caregiving quality across the two follow ups (P = .041, ηp² = .141), a large effect. For all other caregiver and patient outcomes, TAP participants consistently demonstrated more favorable scores with small-to-medium effect sizes, though differences did not reach statistical significance. The TAP improved caregiver-reported caregiving quality and showed promising effect sizes across multiple psychosocial outcomes. Findings support the feasibility and potential utility of telehealth-based caregiver interventions during the transition from inpatient rehabilitation to home, warranting larger-scale trials.Trial registration: ClinicalTrials.gov identifier: NCT03244098.
The determination of quartic force fields for use in vibrational second-order perturbation (VPT2) calculations, currently available in numerous electronic structure packages, becomes very expensive as the size of the molecule increases, especially if high-level coupled-cluster theory is used. Machine-learned potentials (MLPs) for large molecules and clusters offer a viable alternative to obtaining the quartic force field (QFF). Here, we report Fortran and Python software to determine the QFF and perform VPT2 calculations of energies from the MLPs. We describe this software and then apply it to H2O and protonated oxalate as the test cases. The Fortran software is applied to 21-atom aspirin using a fast MLP reported by us. Despite the fact that there are 32,509 unique cubic force constants for aspirin, the computer time to calculate them using this MLP is trivial, i.e., around 1 min. The new software provides an efficient way to calculate quantum anharmonic energies, using the established VPT2 methodology, for machine learned potentials of large molecules.
End-of-life (EOL) costs account for a disproportionately large share of total Medicare expenditures. However, there is a lack of research examining EOL costs of patients in the US across the full spectrum of ages. To report EOL costs and the monthly health care cost trajectory in the year before death across age groups in the US. This cohort study included individuals aged 0 to 75 years or older identified from the MarketScan Mortality Detail file who died from January 2016 to December 2024. Data were linked to the MarketScan Commercial Claims and Encounters and MarketScan Medicare databases. Age at death, categorized as 18 years or younger, 19 to 49 years, 50 to 64 years, 65 to 74 years, or 75 years or older. Total and out-of-pocket (OOP) health care costs (in 2025 US dollars) in the 365 days before the date of death for patients who died in a hospital and in the month of death and the 11 months prior for patients who did not die in a hospital. Both actual and standardized (ie, Medicare pricing) costs were determined. A generalized linear model with γ family and log link and quantile regression were applied to compare mean and median costs, respectively, across age groups. Among 313 649 individuals who died from 2016 to 2024 (51.1% male; mean [SD] age at death, 77.5 [14.3] years), median actual EOL costs from the unadjusted analysis were $50 599 (IQR, $21 258-$109 121) overall; the adjusted actual median costs were highest for the youngest age group (≤18 years; $165 030 [95% CI, $160 110-$169 951]) followed by the group aged 50 to 64 years ($132 542 [95% CI, $131 698-$133 386]) and were lowest in the oldest age group (≥75 years; $48 308 [95% CI, $47 945-$48 671]). After standardizing costs to Medicare pricing, adjusted median EOL costs were $49 259 (IQR, $20 889-$103 317) overall and were highest in the youngest age group (≤18 years; $104 386 [95% CI, $100 144-$108 628]) followed by the group aged 50 to 64 years ($99 182 [95% CI, $98 454-$99 909]), with the lowest median costs observed in the oldest age group (≥75 years, $48 308 [95% CI, $47 945-$48 671]). Unadjusted OOP costs were also higher in younger age groups: 18 years or younger, $4274 (IQR, $1979-$6713); 19 to 49 years, $4063 (IQR, $1640-$6638); 50 to 64 years, $4048 (IQR, $1954-$6556); 65 to 74 years, $1793 (IQR, $1081-$2580); 75 years or older, $1509 (IQR, $793-$2299). Monthly cost trajectories showed EOL costs were persistently higher in younger age groups, with the sharpest increase within 3 months of death and peaking in the month of death (eg, mean total cost of $220 819 [95% CI, $195 716-$245 923] in the last month for patients aged ≤18 years). A much higher percentage of the youngest group (82.3%) died in a hospital compared with the oldest group (20.8%). In this cohort study of EOL costs across age groups, total and OOP costs were substantially higher in patients younger than 65 years vs 65 years or older, even after standardizing to Medicare pricing. Death in a hospital was more common in the younger age groups. The findings suggest attention to EOL care should be expanded beyond Medicare beneficiaries to patients who die young.
This study evaluates the techno-economic and environmental performance of a sequential system based on fixed-bed column adsorption using magnesium-impregnated rice husk biochar (RHB-Mg) for nutrient removal from wastewater, coupled with a prospective assessment of its reuse as a soil amendment in irrigated rice systems. Scale-up based on laboratory data resulted in a treatment capacity of 4.32 m3/day and a biochar requirement of 56.91 kg/day. The system effectively reduced nitrate and phosphate concentrations below regulatory limits under continuous operation, demonstrating high adsorption performance. The techno-economic analysis over a 20-year period revealed that operational costs are primarily driven by magnesium chloride consumption, which strongly influences overall economic feasibility. Life cycle assessment (LCA) identified biochar production as the main environmental hotspot, contributing the highest impacts across multiple categories due to energy demand. Furthermore, literature-supported and LCA-based evidence indicates that the reuse of nutrient-enriched biochar could potentially reduce fertilizer demand (prospective scenario), decrease irrigation requirements, and contribute to a potential climate change benefit through carbon storage, with an estimated reduction of -1.34-kg CO2 eq per kg of RHB-Mg applied to soil. However, this stage was evaluated as a prospective scenario and was not experimentally validated. Overall, the proposed system demonstrates strong potential within a circular-economy framework; however, process optimization-particularly in reagent consumption and energy integration-is required to enhance large-scale sustainability and economic viability.
Autism spectrum disorder (ASD) has been repeatedly linked to gut microbiota alterations, yet mechanistic insight remains limited by the scarcity of ASD-specific cultured isolates. Here, we performed large-scale gut microbial culturomics on fecal samples from 41 children with ASD and 12 typically developing (TD) controls, generating 1,724 isolates across six phyla. Longitudinal culturomics profiling was further conducted in 17 ASD children undergoing a 9-week fecal microbiota transplantation (FMT) intervention. All isolates underwent 16S rRNA sequencing and non-redundant clustering to assess species-level diversity, ASD-TD differences, and microbial dynamics associated with clinical response. ASD children harbored a distinct culturable microbiota enriched for Shigella flexneri and Shigella boydii, whereas TD children were enriched in beneficial taxa, such as Bifidobacterium catenulatum subsp. and other health-associated species. Notably, 20 species isolated from ASD children and 20 from TD children were absent from major existing gut microbiota biobanks, thereby expanding the cultivable repertoire. Among FMT participants, clinical responders exhibited increased alpha diversity, progressive enrichment of TD-associated beneficial taxa, such as Bacteroides fragilis, Anaerostipes hadrus, Parabacteroides merdae, and Turicibacter sanguinis, and a marked reduction of ASD-associated Shigella flexneri and Shigella boydii, whereas non-responders showed minimal shifts. Acquisition of TD-enriched strains at week 9 was strongly correlated with clinical improvement, suggesting that species-level ecosystem remodeling may contribute to FMT efficacy. This work establishes one of the first ASD-focused gut microbial culturomics resources, identifies Shigella as a potential ASD-associated taxon, and provides foundational evidence and testable mechanistic hypotheses for future microbiome-based interventions in ASD.IMPORTANCEMost autism spectrum disorder (ASD) microbiome studies rely on sequencing, which identifies associations but lacks live strains needed for mechanistic tests. We cultured 1,724 isolates from ASD and typically developing (TD) children, providing an ASD-focused, strain-level resource. ASD samples showed a significantly higher prevalence of Shigella flexneri. Longitudinal profiling during fecal microbiota transplantation (FMT) showed that clinical responders gained TD-enriched taxa and lost Shigella spp., and these shifts correlated with symptom improvement. This resource enables functional assays and gnotobiotic studies with ASD-relevant strains and provides a foundation for rational microbiome-based interventions.
Experimental medicine studies, small, mechanistically focused investigations, have historically driven key discoveries in human physiology and pharmacology. Despite their foundational role, these studies are increasingly marginalised in today's drug development environment due to economic pressures, regulatory conservatism, and an overemphasis on statistical endpoints from large-scale trials. This article traces the historical roots and enduring value of experimental medicine, distinguishes it from current early phase drug development studies, and explores the structural forces behind its decline. N-of-1 trials are discussed as a systematic extension of these principles, offering precision insights at the individual level. We apply this discussion to chronic kidney disease (CKD), a field where slow progression and heterogeneous pathophysiology make early mechanistic studies especially valuable. We argue that bypassing such studies in favour of speed represents a strategic gamble that may misdirect costly late-phase trials. Integrating mechanistic insights with statistical power is not superfluous, but essential, particularly in complex diseases like CKD where understanding why and how interventions work may matter as much as whether they do. We acknowledge that achieving this vision necessitates overcoming significant structural, economic, and cultural barriers within the current drug development environment; however, the costs of inaction, manifest as trial failures, patient harm, and missed therapeutic opportunities, are potentially much greater.
Comorbid insomnia and sleep apnea (COMISA) is a highly burdensome phenotype with additive effects on symptoms, quality of life, and health outcomes. Although obstructive airway diseases (OAD)-including asthma and chronic obstructive pulmonary disease-commonly coexist with sleep complaints, the association between OAD and COMISA in large sleep-clinic cohorts remains unclear. We investigated whether OAD is independently associated with COMISA in a nationwide cohort undergoing full-night polysomnography (PSG). We analyzed 12,715 adults in the Turkish Sleep Apnea Database (TURKAPNE), a prospective, multicenter registry. Demographics, comorbidities, insomnia symptoms, and PSG parameters were compared between individuals with and without OAD, defined as self-reported, physician-diagnosed asthma or COPD. COMISA was defined as the coexistence of obstructive sleep apnea (apnea-hypopnea index ≥ 5 events/h) and insomnia symptoms occurring "often" or "very often." Multivariable logistic regression was used to determine the independent association between OAD and COMISA. OAD was present in 12.2% of participants (n = 1,546). COMISA prevalence was higher in those with OAD than in those without OAD (16.1% vs 10.8%, p < 0.001). After adjustment for age, sex, body mass index, education, smoking status, and comorbidities, OAD remained independently associated with COMISA (adjusted odds ratio 1.19; 95% CI 1.06-1.35; p = 0.004). Among participants with COMISA, individuals with OAD were older, more frequently female, more obese, and demonstrated shorter total sleep time, longer wake after sleep onset, higher periodic limb movement index, and poorer nocturnal oxygenation. In this nationwide sleep-clinic cohort, OAD was independently associated with COMISA. COMISA patients with OAD exhibited more adverse clinical and polysomnographic features. These findings highlight the multidimensional sleep burden in OAD and support integrated, personalized management strategies.
Hepatitis B virus (HBV) reactivation is a potentially severe and preventable complication of immunosuppressive therapy, particularly with anti-CD20 agents such as rituximab, underscoring the need for effective screening and prophylaxis to reduce adverse outcomes. We conducted a retrospective population-based cohort study using the Clalit Health Services database in northern Israel. Adult patients aged ≥ 18 years who received rituximab between 2005 and 2022 were included. HBV screening (HBsAg and anti-HBc), antiviral prophylaxis, and HBV reactivation events were evaluated. Timely screening was defined as testing performed within 90 days before rituximab initiation. Temporal trends were analyzed using the Cochran-Armitage test, and predictors of reactivation were evaluated using Cox proportional hazards models, with the per-patient analysis considered the primary model. After exclusion of patients younger than 18 years, a total of 11,888 adult patients received rituximab during the study period (mean age 61.6 ± 15.6 years; 54.7% female). Only 10.5% of patients underwent timely HBV screening before treatment initiation, while 47.4% were never screened without time restriction and 89.5% had no HBV serology within 90 days before rituximab initiation. Among screened patients, approximately 6.0% had evidence of current or prior HBV infection. Screening rates improved significantly over time (P < 0.0001) but remained suboptimal across clinical settings. Antiviral prophylaxis increased over the study period but plateaued at approximately 50% among eligible patients. A total of 159 HBV reactivation events were identified, corresponding to an incidence of 1.06% per treatment event and approximately 1.3% per patient. Reactivation occurred predominantly among patients who had not undergone HBV screening. Multivariable analysis identified male sex as an independent predictor of HBV reactivation. Despite gradual improvement over nearly two decades, HBV screening before rituximab therapy remains insufficient in real-world clinical practice. Given that HBV reactivation is largely preventable with appropriate screening and antiviral prophylaxis, system-level interventions are needed to improve adherence to guideline-recommended care.
The need for accurate, cost-effective, non-invasive screening tests for colorectal cancer (CRC) persists. Repetitive genomic elements like Long Interspersed Nuclear Element-1 (LINE-1), due to their high copy number and cancer-associated alterations, present a promising but unexplored target in stool DNA for CRC detection. In this cross-sectional study, stool DNA was extracted from 38 patients with histologically confirmed CRC and 39 healthy controls. The concentration of human DNA was quantified via quantitative PCR targeting two short LINE-1 amplicons (55-bp and 60-bp). Diagnostic performance was evaluated using ROC curve analysis to determine sensitivity, specificity, and predictive values. The concentration of stool DNA, measured by LINE-1 qPCR cycle threshold (Ct) values, was significantly higher in CRC patients compared to controls (p < 0.001 for both amplicons). The 60-bp LINE-1 amplicon demonstrated superior diagnostic accuracy with an area under the ROC curve (AUC) of 0.92 (95%CI 0.85-0.98), compared to 0.83 (95%CI 0.73-0.93) for the 55-bp amplicon. At the optimal cut-off (Ct ≤ 24.19), the 60-bp marker achieved a sensitivity of 87.0%, specificity of 78.4%, a positive predictive value (PPV) of 71.4%, and a negative predictive value (NPV) of 90.6%. The 55-bp amplicon showed higher sensitivity (95.7%) but lower specificity (59.5%). Quantitative analysis of LINE-1 repetitive elements in stool DNA, particularly using a 60-bp amplicon, demonstrates excellent diagnostic accuracy for detecting CRC. This single-marker, qPCR-based assay represents a simple and potentially low-cost non-invasive strategy, warranting further validation in large-scale screening populations and assessment of its efficacy for detecting precancerous lesions.