Per- and polyfluoroalkyl substances (PFAS) are persistent environmental contaminants that can disrupt human hepatic metabolism both directly and through alterations of the gut microbiota. However, the contribution of microbiota-mediated mechanisms to PFAS-induced hepatic dysfunction remains poorly understood. Here, we investigated how PFAS-modified gut microbial metabolites affect human hepatocyte metabolism using an in vitro colon fermentation model, supported by an in vivo mouse and in vitro human hepatocyte exposure studies. PFAS exposure altered the fecal metabolome in human colonic fermentations, particularly affecting pathways related to fatty acid, amino acid, vitamin, and mitochondrial metabolism. Fecal metabolomics from PFOA-exposed mice showed overlapping pathway-level alterations, including effects on fatty acid, bile acid, and steroid hormone metabolism, supporting the biological relevance of the in vitro findings. Exposure of HepaRG hepatocytes to control fermentation extracts markedly altered lipid profiles, confirming that gut-derived metabolites actively regulate hepatic metabolism. Notably, PFAS-exposed fermentation extracts induced distinct hepatocyte metabolic changes compared with PFAS-spiked control extracts, indicating effects driven by PFAS-modified microbial metabolites rather than direct PFAS carry-over. These changes included decreased acyl-carnitines and increased L-carnitine, consistent with altered fatty acid transport and mitochondrial β-oxidation. PFAS-modified extracts also altered bile acids, steroid metabolites, inosine, and sialic acid derivatives, suggesting broader alteration of bile acid signaling, endocrine-related metabolism, purine metabolism, glycoprotein turnover, and lipid-glucose homeostasis. These findings from our pilot study demonstrate that PFAS exposure reshapes gut microbial metabolite profiles with downstream consequences for hepatocyte metabolism. Our findings provide new mechanistic insight into how PFAS may contribute to metabolic disorders.

INTRODUCTION The increasing recognition of peripheral GABAergic signaling has renewed interest in the vascular effects of benzodiazepines (BDZs), particularly in light of their frequent use during pregnancy. Although BDZs have been studied in various vascular beds, their exact vasorelaxant mechanisms are still unclear. This study aimed to investigate the mechanisms underlying diazepam-induced vasodilation in human umbilical arteries (HUAs). METHODS Isolated endothelium-intact HUA rings were pre-contracted with serotonin, and concentration-response curves for diazepam (0.01-100 μM) were obtained using the tissue bath method. The roles of the endothelium, potassium and calcium channels, and TSPO were assessed using specific inhibitors or blockers: L-NAME, indomethacin, sildenafil, ODQ, glibenclamide, 4-aminopyridine, tetraethylammonium, and PK11195. The effects of diazepam pre-treatment on BaCl2-and serotonin-induced contractions were also investigated. RESULTS Diazepam induced a time- and concentration-dependent relaxation of HUA, with ⁓ 96% of the maximal effect being achieved at 100 μM. This response was endothelium-independent, as it was unaffected by eNOS or COX inhibition, but included an ODQ-sensitive component compatible with involvement of the soluble guanylate cyclase (sGC)-cGMP pathway. Potassium channel involvement was confirmed, and diazepam significantly suppressed BaCl2-and serotonin-induced contractions, suggesting calcium channel antagonism. PK11195 significantly attenuated diazepam-induced relaxation, implying involvement of TSPO. CONCLUSION Diazepam elicits near-complete relaxation in serotonin-precontracted HUA through endothelium-independent mechanisms that likely involve potassium channel activation, calcium channel inhibition, a component sensitive to ODQ suggesting possible involvement of the sGC-cGMP pathway, and interaction with TSPO. These results emphasize the remarkable vasoactivity of diazepam and its importance in the regulation of umbilical vascular tone.

Edible Natural Deep Eutectic Solvents (NADES) offer a route to ready-to-use extracts without solvent removal. This study examined how rational formulation design influences physicochemical properties, extraction performance, energy efficiency, thermal behavior, and cytocompatibility during bioactive recovery from Nigella sativa seeds. Twelve formulations spanning malic acid–polyol, citric acid–polyol, binary polyol, and ternary acid–polyol families were evaluated under standardized ultrasound-assisted conditions and compared with ethanolic ultrasound-assisted extraction, Soxhlet extraction, and cold-pressed oils. Selected NADES formulations outperformed the conventional systems. E1 (malic acid:glycerol:water) achieved the highest total phenolic content and lowest specific energy consumption (40.00 kJ mg−1 GAE), below the Soxhlet benchmark (55.17 kJ mg−1 GAE), whereas E9 (glycerol:xylitol:water) exhibited the greatest ABTS activity. Neat-NADES apparent pH and viscosity were inversely associated with total phenolic content, while viscosity and density were inversely associated with ABTS activity. ATR-FTIR showed stable, solvent-dominated fingerprints over 15 days, whereas DSC better differentiated neat NADES from their extracts. Most systems remained cytocompatible at 10,000× dilution, while acid–polyol formulations reduced viability at 500×; partial neutralization of N5/E5 restored viability. NADES performance was formulation- and application-dependent, requiring joint optimization of composition, acidity, viscosity, energy efficiency, and biologically compatible concentration.

Colorectal cancer remains a leading cause of cancer-related mortality worldwide, with automated polyp classification from endoscopic images offering a promising avenue for improving early detection. Existing approaches rely on single convolutional neural network (CNN) backbones with manually designed classification heads, limiting both representational capacity and deployment flexibility. This paper presents a swarm intelligence-augmented multi-backbone deep learning framework for eight-class gastrointestinal lesion classification on the Kvasir benchmark. Four CNN backbones (ResNet50, DenseNet121, MobileNetV2, EfficientNetB3) are independently fine-tuned using a two-phase transfer learning protocol and their penultimate-layer features concatenated into a 5888-dimensional representation, reduced to 256 dimensions via PCA. Five swarm intelligence algorithms—Particle Swarm Optimization, Artificial Bee Colony, JADE, L-SHADE, and CMA-ES—are benchmarked on the classification head architecture search task; all independently converge to tanh activation, a consistent pattern across independently initialized algorithms that is suggestive of, though not conclusive evidence for, particular geometric properties of PCA-transformed deep feature spaces. The PSO-optimized single-layer head (284 units, tanh) outperforms a manually designed three-layer baseline by 0.75% while using 67% fewer parameters. SI-guided class weight optimization yields targeted F1 improvements on the two most clinically significant classes (polyps: +0.015, ulcerative-colitis: +0.013). The fixed-head classifier trained on fused four-backbone features achieves 91.08% accuracy on Kvasir v2 (multi-seed mean 91.47% ± 0.49 across nine converging seeds; one seed failed to converge and is disclosed rather than excluded), below end-to-end DenseNet121 (92.25%; Wilcoxon p = 0.31, not statistically significant), while enabling classifier updates in under 30 s; a three-backbone subset dropping the weakest backbone (EfficientNetB3) reaches 92.33%, exceeding the full four-backbone fusion. Cross-dataset evaluation on Kvasir v1-to-v2 confirms near-zero generalization gaps across dataset scales; a restricted two-class evaluation on HyperKvasir (the only two of eight classes with usable labeled data) reaches 96.28% accuracy, and dual Grad-CAM with SI minimal sufficient region analysis, validated quantitatively against Kvasir-SEG ground-truth masks, provides spatially grounded, clinically interpretable explanations.

BACKGROUND Despite a general improvement in the HIV burden over the past two decades, gendered social determinants including intimate partner violence (IPV) continue to affect women's vulnerability to HIV, which could be further exacerbated as resources dwindle over the coming years. Our study aimed to quantify recent trends in the HIV burden, the magnitude of the HIV burden associated with IPV, and the potential impact of declining financial support globally. METHODS Using the Global Burden of Diseases, Injuries, and Risk Factors Study (GBD) 2023 framework, we estimated annual HIV incidence, prevalence, and mortality for 204 countries and territories from 1990 to 2023, disaggregated by age and sex. Input data included national HIV programme reports, population-based serosurveys, clinical and vital registration data, and systematically reviewed literature. Countries and territories were grouped according to the availability and completeness of HIV prevalence and mortality data, with tailored modelling approaches applied for each group. To characterise the potential contributions of IPV to the HIV burden, time series estimates of IPV prevalence by location and effect size estimates were combined to generate population attributable fractions of the HIV burden. To forecast future trends and assess the potential impact of reductions in development assistance for health, we used a non-parametric stochastic frontier approach to estimate how funding levels could affect antiretroviral therapy (ART) coverage. Adjustments in future ART coverage were integrated into our forecasts, altering projected rates of HIV-related incidence and mortality. We compared scenarios with and without funding cuts, quantifying the potential epidemiological effects of reduced ART availability by sex. FINDINGS Between 2003 and 2023, the annual number of new HIV infections declined globally, from 2·85 million (95% uncertainty interval [UI] 2·76-2·96) to 2·06 million (1·88-2·29). Meanwhile, HIV-related deaths decreased from a peak of 1·71 million (1·61-1·83) in 2004 to 0·83 million (0·73-0·96) in 2023. The number of people living with HIV rose from 26·3 million (25·5-27·3) in 2003 to 42·4 million (40·3-44·4) in 2023, reflecting improved survival associated with ART expansion. Although females continued to have higher prevalence of HIV than males in 2023 (23·4 million [22·2-24·6] vs 19·1 million [17·6-20·5]), the gap in incidence between females and males has substantially declined. Regionally, sub-Saharan Africa had the greatest declines in both incidence (64·3%; 58·0-68·6) and mortality (74·6%; 70·4-78·5) between 2003 and 2023, while central Europe, eastern Europe, and central Asia recorded the highest increases in incidence rates (232·1%; 146·5-299·1) and mortality rates (37·8%; 31·2-45·5) during this period. IPV was associated with an estimated 10·7% (1·6-20·4) of HIV-related deaths among females aged 15 years and older globally in 2023, corresponding to 40 700 (6400-80 600) deaths, with the highest population attributable fractions observed in Oceania (17·4%; 2·8-33·1), central sub-Saharan Africa (14·2%; 2·1-29·4), and eastern sub-Saharan Africa (13·2%; 2·0-25·3). Forecasts indicate that funding reductions could decrease ART coverage by 8·1% globally between 2025 and 2030, resulting in approximately 1·6 million new HIV infections among females and 1·3 million new HIV infections among males, alongside 790 600 and 670 600 HIV-related deaths, respectively. These impacts are primarily concentrated in sub-Saharan Africa. INTERPRETATION Despite two decades of substantial progress, the global HIV response remains highly sensitive to gendered social determinants and funding stability. The potential contribution of IPV to HIV-related mortality among women underscores the need for integrated interventions that address violence prevention and post-violence care alongside HIV treatment. The projected effects of funding cuts-millions of additional infections and deaths-highlight the fragility of the gains achieved and the urgent need to protect HIV financing. Achieving and sustaining the UNAIDS 2030 targets will require renewed investment, gender-responsive programming, and resilient health systems capable of providing equitable access to care. FUNDING Gates Foundation and the US National Institute of Allergy and Infectious Diseases.

Simple Summary The biodiversity of bats in Bosnia and Herzegovina has only been assessed by methods that require capture and/or the analysis of echolocation signals, thus giving and incomplete picture of the true biodiversity. To address this, we aimed to measure bat diversity by analyzing DNA from bat droppings (guano) and comparing the results of this advanced genetic method to traditional morphological measurements of captured bats. The results revealed that while the morphological method identified only 10 bat species, the DNA analysis successfully identified 13 species—including Mehely’s horseshoe bat, a species never recorded in the country before. Furthermore, DNA testing clarified uncertain morphological identifications and discovered more species at every location. The study concludes that Bosnia and Herzegovina hosts a much richer diversity of bats than previously thought, and proves that non-invasive DNA sampling from guano is a highly effective tool for routine bat monitoring and provides an easy, harmless way for conservationists to pinpoint and protect critical natural habitats, ensuring the survival of these ecologically vital animals and maintaining healthy local ecosystems.

The rapid growth of e-commerce has intensified last-mile delivery activities in urban areas, creating challenges for city logistics systems related to operational efficiency, congestion, and environmental impacts. In this context, understanding the factors that influence customer satisfaction with logistics operators is increasingly important, as mismatches between customer expectations and delivery service characteristics may lead to operational inefficiencies such as failed delivery attempts and repeated delivery rounds. This study proposes a machine learning framework for predicting customer satisfaction with postal and logistics operators in urban delivery systems using survey data on customer characteristics, preferences, and service perceptions. Several machine learning algorithms were developed and evaluated to identify the key determinants of customer satisfaction and assess their predictive performance. Beyond predictive accuracy, the study interprets customer satisfaction as an indicator of the alignment between customer expectations and delivery service configurations. Improved alignment may support service configurations that reduce delivery mismatches and repeated delivery attempts, which are recognized as a significant source of additional transport activity in urban freight systems. By identifying customer segments whose expectations are not adequately addressed by existing delivery services, the proposed framework can support more informed service design and operational decision-making. From a city logistics perspective, the potential reduction in failed deliveries and repeated delivery rounds may contribute to lower vehicle kilometers travelled, congestion, energy consumption, and emissions associated with urban freight transport, although these operational and environmental indicators were not directly measured in this study. The proposed approach therefore provides a data-driven decision-support tool that can help operators improve service quality and serve as a basis for future integration with operational and environmental indicators in sustainable last-mile delivery planning.

This retrospective study of 98 patients assessed whether hypertensive heart disease (HHD) is associated with angiographic coronary artery disease (CAD) burden, quantified by the Duke CAD Index, and whether HHD or CAD burden influenced revascularization strategy. HHD was independently associated with higher Duke CAD Index values in linear regression (B = 10.33, p = 0.038) and with greater odds of more severe CAD categories in ordinal regression (OR = 3.19, p = 0.015). However, revascularization choice did not differ by HHD status or CAD severity (p = 0.95). A model incorporating age, sex, diabetes mellitus, and HHD demonstrated moderate discriminative ability for predicting severe CAD (AUC = 0.753). These findings suggest that HHD identifies a high-risk coronary atherosclerotic phenotype, but they are hypothesis-generating and require confirmation in larger prospective studies before clinical application.

Background/Objectives: Early risk stratification remains challenging in patients with non-ST-segment elevation myocardial infarction (NSTEMI). The present study evaluated the prognostic value of 24 h high-sensitivity cardiac troponin I (hs-Troponin I) and assessed whether combining biomarkers and echocardiographic parameters improves short-term risk prediction. Methods: This prospective observational cohort study included 170 consecutive adult patients with confirmed NSTEMI who were admitted to a Medical Intensive Care Unit and prospectively enrolled between February 2022 and January 2023. Clinical, routine biochemical, inflammatory, hematological, lipid, and echocardiographic data were collected during index hospitalization. High-sensitivity cardiac troponin I was measured at admission and again 24 h after hospitalization, with the 24 h value used as the principal marker of myocardial injury in the prediction analyses. The primary endpoint was major adverse cardiovascular events (MACEs), defined as cardiovascular death, recurrent myocardial infarction, ischemic stroke, urgent coronary revascularization, or hospitalization for worsening heart failure, within 3 months. Multivariable logistic regression, Cox regression, sequential prediction modeling, and internal bootstrap validation were performed. Results: MACEs occurred in 88 patients (51.8%). Twenty-four-hour hs-Troponin I, but not admission hs-Troponin I, was independently associated with MACEs (OR 1.57, 95% CI 1.09–2.26; p = 0.015) and a shorter time to the first MACE event (HR 1.38, 95% CI 1.07–1.78; p = 0.012). Lower left ventricular ejection fraction (LVEF) was also independently associated with adverse outcomes. The addition of 24 h hs-Troponin I, LVEF, and C-reactive protein improved discrimination from an AUC of 0.665 to 0.759 (optimism-corrected AUC, 0.717), with corresponding improvements in reclassification. A simplified multimarker score was independently associated with event-free survival (HR 2.36, 95% CI 1.53–3.64; p < 0.001). Conclusions: In patients admitted to a medical intensive care unit with NSTEMI, the integration of 24 h hs-Troponin I, LVEF, and C-reactive protein improved short-term risk prediction beyond that of clinical variables alone. A practical multimarker model based on routinely available parameters identified patients at increased risk of adverse cardiovascular outcomes during early follow-up.

Background In patients with heart failure with preserved ejection fraction (HFpEF), left atrial (LA) strain represents a new indicator of mechanical atrial dysfunction and an important prognostic marker. However, the clinical significance of serial changes in LA strain during patient follow-up is still not sufficiently elucidated. Objectives To evaluate changes between baseline and one-year follow-up in LA strain during one-year follow-up in patients hospitalized with HFpEF and to investigate their association with major adverse cardiovascular events (MACE). Methods This prospective observational cohort study included 123 consecutive patients hospitalized due to acute decompensated HFpEF. Comprehensive echocardiographic evaluation was performed during index hospitalization and after 12 months. LA reservoir strain was assessed as peak atrial longitudinal strain (PALS). The primary endpoint was MACE, defined as rehospitalization due to worsening heart failure and/or cardiovascular death during one-year follow-up. Results Patients who developed MACE had lower baseline LA strain, higher LAVI, higher E/e′, higher N-terminal pro-B-type natriuretic peptide (NT-proBNP), higher right ventricular systolic pressure (RVSP), lower tricuspid annular plane systolic excursion (TAPSE), and more impaired left ventricular global longitudinal strain (LV GLS) compared with patients without MACE. Assessment at two predefined time points showed overall improvement in LA strain during follow-up, but patients with adverse outcomes remained characterized by persistently impaired LA strain and an unfavorable hemodynamic profile. Atrial fibrillation was more frequently observed among patients who developed MACE. Conclusion Persistently impaired or worsening LA strain during one-year follow-up was associated with adverse outcomes in patients hospitalized with HFpEF. Assessment of LA strain at baseline and one-year follow-up may improve longitudinal risk stratification beyond conventional echocardiographic parameters.