Tag: statistics

Front Dent Med. 2026 Aug 11;7:1895392. doi: 10.3389/fdmed.2026.1895392. eCollection 2026.

ABSTRACT

BACKGROUND: Platelet-rich fibrin has shown regenerative potential in periodontal treatments; nonetheless, patients with Type 2 diabetes mellitus (T2DM) may experience altered biological activities. Therefore, this study aimed to evaluate the regenerative potential and growth factor release profiles of advanced platelet-rich fibrin (A PRF) and injectable platelet-rich fibrin (i-PRF) derived from participants with controlled and uncontrolled T2DM.

METHODS: This study compared n = 87 (29 per group) participants-controlled T2DM, uncontrolled T2DM, and Healthy non-diabetic groups-analysing blood samples for HbA1c, platelet count, and growth factors like Vascular endothelial growth factor (VEGF), Platelet-derived growth factor-BB (PDGF-BB), Insulin-like growth factor-1 (IGF-1), Transforming growth factor-beta 1 (TGF-β1), Transforming growth factor-beta 2 (TGF-β2) via Enzyme-linked immunosorbent assay (ELISA), assessing cytocompatibility with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assays, and employing statistical methods like Analysis of variance (ANOVA) correlation, and regression.

RESULTS: PRF derived from participants with uncontrolled T2DM demonstrated significantly reduced PDGF-BB and TGF-β2 release compared with healthy participants in both A-PRF (P = 0.016), i-PRF (P = 0.033), and A-PRF (P = 0.002) and i-PRF (P = 0.0034. VEGF release was significantly lower in uncontrolled T2DM than in well-controlled T2DM in both A-PRF (P = 0.0096) and i-PRF (P = 0.0023). Compared with the uncontrolled T2DM group, well-controlled T2DM exhibited significantly higher VEGF release in both A-PRF (P = 0.0096) and i-PRF (P = 0.0023), and significantly higher IGF-1 (P < 0.05) and TGF-β2 (P = 0.033) levels in i-PRF. Platelet counts were comparable across the study groups, whereas cell viability was highest in PRF from healthy participants. HbA1c was not independently associated with the evaluated growth factor concentrations in multivariable analyses.

CONCLUSIONS: A-PRF and i-PRF derived from participants with well-controlled T2DM exhibited biological properties comparable to those of healthy non-diabetic individuals, whereas uncontrolled T2DM was associated with reduced release of key growth factors, indicating compromised healing potential. These findings suggest that achieving good glycaemic control may optimize the biological performance of PRF for periodontal regenerative therapy. However, the present findings are based on in vitro analyses; confirmation through well-designed prospective clinical studies remains essential before these findings can be translated into routine clinical practice.

PMID:42643293 | PMC:PMC13503254 | DOI:10.3389/fdmed.2026.1895392

Front Med (Lausanne). 2026 Aug 11;13:1895176. doi: 10.3389/fmed.2026.1895176. eCollection 2026.

ABSTRACT

OBJECTIVES: Artificial intelligence (AI) systems are increasingly adopted in scientific visualization, clinical sciences, medical education, and health sciences research. Despite its impressive generative capabilities, the AI role remains poorly established in some areas. This study aims to investigate whether AI systems, ChatGPT-4 and Google Gemini 1.5 Pro, can identify their own generated images.

METHODS: In this study, AI systems, OpenAI’s ChatGPT-4 and Google Gemini 1.5 Pro, were employed to generate and identify medical science images. AI-generated images in medical science span 16 clinical illustrations: 8 anatomical and 8 pathophysiological mechanisms. The images were subsequently reintroduced into the same system for identification. The descriptive statistics, including numbers, percentages, and accuracy rates, across the images were analyzed. For a correct answer or an incorrect answer, a score of 1 was allocated; a p-value less than 0.05 was considered significant.

RESULTS: The results revealed that Artificial Intelligence models, ChatGPT and Google Gemini, were significantly less able to identify AI-generated images correctly. There were significantly higher incorrect rates than correct identification rates for AI-generated images overall (27/32 incorrect vs. 5/32 correct; 84.37% vs. 15.62%; p = 0.001). ChatGPT showed an overall accuracy of 18.75% (p = 0.021), while Google Gemini demonstrated an overall accuracy of 12.5% (p = 0.004). The poorest performance was observed in Google Gemini anatomy images, with no correct identifications (0%, p = 0.008).

CONCLUSION: Artificial Intelligence models, ChatGPT and Google Gemini, have limited ability to identify AI-generated images correctly. The results demonstrate structured, architecture-dependent patterns of image identification dysfunction. The study emphasizes implications for AI governance and validation protocols in research, medicine, and medical education.

PMID:42643292 | PMC:PMC13503190 | DOI:10.3389/fmed.2026.1895176

Front Cardiovasc Med. 2026 Aug 11;13:1852845. doi: 10.3389/fcvm.2026.1852845. eCollection 2026.

ABSTRACT

INTRODUCTION: The gamma-glutamyl transferase to high-density lipoprotein cholesterol ratio (GGT/HDL-C) integrates oxidative stress and lipid dysregulation, but its relationship with subclinical arterial stiffness elevation (SASE) and potential sex- and age-specific differences remains unclear. This study examined the association between GGT/HDL-C and SASE, assessed using brachial-ankle pulse wave velocity (baPWV), in an asymptomatic adult population.

METHODS: This cross-sectional study included 10,518 adults who underwent routine self-paid health examinations at Xiamen Chang Gung Hospital between 2013 and 2016. Clinical, biochemical, and vascular parameters were assessed. SASE was defined as a mean baPWV >1,700 cm/s. Participants were stratified by sex and age (≤50 and >50 years) and categorized into sex- and age-specific tertiles of GGT/HDL-C. Hierarchical logistic regression models were adjusted for age, fasting glucose, triglycerides, low-density lipoprotein cholesterol, alcohol intake, and lipid-lowering medication use. Receiver operating characteristic analyses were performed to evaluate discriminatory performance.

RESULTS: The study included 5,877 men and 4,641 women, with a median age of 46 years; 799 participants (7.6%) had SASE. Higher GGT/HDL-C tertiles were associated with less favorable cardiometabolic profiles and progressively higher baPWV. In the fully adjusted model, the highest versus lowest GGT/HDL-C tertile was associated with increased odds of SASE in men >50 years (odds ratio [OR], 1.823; 95% confidence interval [CI], 1.303-2.550), women ≤50 years (OR, 4.620; 95% CI, 1.012-21.101), and women >50 years (OR, 1.435; 95% CI, 1.009-2.040). In men ≤50 years, the association was positive but of borderline statistical significance (OR, 1.877; 95% CI, 0.994-3.544). Sensitivity analyses yielded generally consistent findings. GGT/HDL-C alone demonstrated modest discriminatory performance, which was better in women than in men.

DISCUSSION: Higher GGT/HDL-C ratios were associated with SASE, with variations in the magnitude and statistical precision of the associations across sex and age groups. The GGT/HDL-C ratio may serve as an adjunctive biomarker for preliminary vascular risk stratification. However, its modest stand-alone discriminatory performance and the cross-sectional, single-center study design warrant cautious interpretation and external validation.

PMID:42643267 | PMC:PMC13503325 | DOI:10.3389/fcvm.2026.1852845

Front Psychol. 2026 Aug 11;17:1825908. doi: 10.3389/fpsyg.2026.1825908. eCollection 2026.

ABSTRACT

Considering the prevalence of bullying in Türkiye, there are relatively few studies focusing on bullying victimization among gifted and typically developing students. This study examined the relationship between life satisfaction and bullying victimization among these two groups, considering the mediating role of dysfunctional attitudes. The participants comprised 486 students: 241 gifted students (99 female, 142 male; M _age = 11.29; SD _age = 1.35) and 245 typically developing students (118 female, 127 male; M _age = 11.68; SD _age = 1.36). Data were collected using the Satisfaction with Life Scale for Children (SWLS-C), the Dysfunctional Attitudes Scale for Children (8-14 years), and the Child-Adolescent Bullying Scale-9 Short Form. Data were analyzed using IBM SPSS Statistics 25 and AMOS, including multi-group structural equation modeling. The results revealed no significant difference in life satisfaction scores between gifted and typically developing students. However, compared with gifted students, typically developing students had higher scores for dysfunctional attitudes and bullying victimization. In addition, dysfunctional attitudes significantly mediated the association between life satisfaction and bullying victimization in both groups, and the structural paths did not differ significantly across groups. These findings highlight the importance of cognitive and socio-emotional factors in understanding bullying victimization among students.

PMID:42643258 | PMC:PMC13503155 | DOI:10.3389/fpsyg.2026.1825908

Front Public Health. 2026 Aug 11;14:1820948. doi: 10.3389/fpubh.2026.1820948. eCollection 2026.

ABSTRACT

OBJECTIVE: This Bayesian network meta-analysis evaluates the comparative efficacy of cognitive interventions on cognitive function in elderly individuals with cognitive impairment.

METHODS: Randomized controlled trials examining the effects of cognitive interventions on cognitive function in older adults aged 60 years and above with cognitive impairment were searched in the Cochrane Library, PubMed, Embase, Web of Science, China National Knowledge Infrastructure (CNKI), China Biomedical Literature Service System (SinoMed), VIP Database, and Wanfang Data. The search covered the period from database inception to August 2025. Two researchers independently performed study selection, data extraction, and risk-of-bias assessment using the Cochrane risk-of-bias tool. RevMan 5.4 was used for risk-of-bias visualization, Stata 16.0 was used to construct the evidence network plots, and Bayesian network meta-analysis was conducted using the gemtc package in R version 4.5.2.

RESULTS: This meta-analysis incorporated 35 randomized controlled trials (RCTs), encompassing a total of 2,316 patients. Eight distinct cognitive intervention modalities were evaluated. Regarding the Mini-Mental State Examination (MMSE) scores, surface under the cumulative ranking curve (SUCRA) analysis revealed the following probability ranking for efficacy: computerized cognitive training (CCT; SUCRA = 81.08%), non-invasive brain stimulation (NIBS; SUCRA = 73.87%), cognitive stimulation therapy (CST; SUCRA = 67.24%), serious games (SUCRA = 65.64%), memory training (SUCRA = 61.79%), integrated cognitive training (SUCRA = 51.87%), virtual reality (VR; SUCRA = 27.02%), usual care (SUCRA = 14.38%), and olfactory training (OT; SUCRA = 7.10%). League table comparisons further demonstrated statistically significant improvements in MMSE scores compared to control groups for several interventions: CCT (mean difference [MD] = 2.14, 95% CI: 0.90 to 3.39), integrated cognitive training (MD = 1.28, 95% CI: 0.24 to 2.31), CST (MD = 1.70, 95% CI: 0.80 to 2.65), and serious games (MD = 1.67, 95% CI: 0.23 to 3.13). For the Montreal Cognitive Assessment (MoCA) scores, the SUCRA ranking was: memory training (SUCRA = 70.92%), CST (SUCRA = 70.63%), CCT (SUCRA = 69.71%), integrated cognitive training (SUCRA = 60.27%), virtual reality (VR; SUCRA = 48.24%), serious games (SUCRA = 46.16%), NIBS (SUCRA = 29.93%), and usual care (SUCRA = 4.15%). Significant enhancements in MoCA scores versus control were observed for: CCT (MD = 3.13, 95% CI: 1.79 to 4.50), integrated cognitive training (MD = 2.81, 95% CI: 1.41 to 4.28), and CST (MD = 3.26, 95% CI: 1.16 to 5.50).

CONCLUSION: Current evidence suggests that CCT, CST, memory training, and NIBS may be more effective interventions for improving cognitive function in elderly individuals with cognitive impairment.

PMID:42643257 | PMC:PMC13503207 | DOI:10.3389/fpubh.2026.1820948

Neuroimage Rep. 2026 Aug 14;6(3):100400. doi: 10.1016/j.ynirp.2026.100400. eCollection 2026 Sep.

ABSTRACT

•Leakage-free nested CV-RFE identifies a reproducible connectivity signature that characterizes the PTE-diseased state in our cohort.•Data-driven feature selection is essential for classification performance, with only four stable functional connections (from an original feature space of 4950) able to distinguish PTE from TBI with an AUC of 0.95.•We identify a specific connectivity phenotype in our PTE cohort which includes the Visual network as a hub of PTE-related changes, left-hemisphere overrepresentation and hyper-coupling, and right-hemisphere un-coupling.•The resting-state network disruptions identified by our ML paradigm went unrecognized by direct comparison between groups and traditional statistics, supporting the use of advanced data-driven methodologies for exploring functional connectivity in PTE.

PMID:42643256 | PMC:PMC13503136 | DOI:10.1016/j.ynirp.2026.100400

Front Nutr. 2026 Aug 11;13:1878741. doi: 10.3389/fnut.2026.1878741. eCollection 2026.

ABSTRACT

BACKGROUND: Preserved ratio impaired spirometry (PRISm) is a clinically relevant lung-function phenotype associated with adverse respiratory and systemic outcomes. Oxidative imbalance and metabolic dysfunction may contribute to impaired pulmonary reserve, but evidence from Chinese health examination populations remains limited. We examined the associations of oxidative balance score (OBS) and metabolic dysfunction indices with PRISm in Chinese adults.

METHODS: This hospital-based cross-sectional study included 1,311 adults undergoing routine health examinations between June 2024 and December 2025. PRISm was defined as FEV1/FVC ≥ 0.70 and FEV1 < 80% predicted. The standardized weighted OBS was constructed from dietary and lifestyle antioxidant/pro-oxidant components, and TyG, LAP, and AIP were analyzed as standardized continuous variables and quartiles. Multivariable logistic regression, restricted cubic splines, exploratory mediation analysis, and five-fold cross-validated machine-learning models were applied.

RESULTS: PRISm was identified in 92 participants (7.02%). The cohort mean (SD) values were 22.69 (4.98) for OBS, 8.69 (0.40) for TyG, 33.01 (16.73) for LAP, and 0.05 (0.36) for AIP. In fully adjusted models, higher OBS was associated with lower odds of PRISm (OR per 1-SD increase, 0.61; 95% CI, 0.48-0.77), whereas TyG, LAP, and AIP were positively associated with PRISm, with ORs of 1.52, 1.23, and 1.36, respectively. Restricted cubic splines showed an approximately linear inverse OBS-PRISm association. Higher OBS was also associated with more favorable metabolic profiles. Exploratory pathway analyses showed directionally consistent but statistically non-definitive indirect-effect estimates through metabolic indices. Machine-learning analyses showed modest discrimination and ranked TyG, OBS, LAP, and AIP among the leading predictors.

CONCLUSION: Favorable oxidative balance was independently associated with lower PRISm probability, whereas metabolic dysfunction indices were associated with higher PRISm probability. These findings support an oxidative-metabolic profile associated with early lung-function impairment, while longitudinal studies are needed to clarify temporality and causality.

PMID:42643239 | PMC:PMC13503214 | DOI:10.3389/fnut.2026.1878741

Front Nutr. 2026 Aug 11;13:1789238. doi: 10.3389/fnut.2026.1789238. eCollection 2026.

ABSTRACT

BACKGROUND: Dietary habits are important factors of metabolic and hepatobiliary health. However, there is still a lack of research on the relationships between gallstone disease, nonalcoholic fatty liver disease (NAFLD), and regular eating habits.

OBJECTIVE: To investigate the associations between dietary patterns and hepatobiliary outcomes, including NAFLD, gallstone disease, and related biochemical markers, in adults who visit tertiary-care clinics.

METHODS: This retrospective observational association study involved 600 adults aged 20-70 years with a principal component analysis of a validated food frequency questionnaire and categorized them into healthy (n = 285) and Western (n = 315) diets. Hepatobiliary health was assessed by anthropometric measurements, fasting biochemical (alanine aminotransferase, fasting glucose, lipid profile, malondialdehyde) measurements, and abdominal ultrasonography. The odds ratios (ORs) of NAFLD and gallstone disease were estimated using logistic regression to control age, sex and BMI. The effect modification of overall and central adiposity was studied through stratified analyses.

RESULTS: Higher BMI, a wider waist circumference, more energy intake, and less physical activity were all related to following a Western diet (all p < 0.001). In crude and partially adjusted models, Western dietary patterns were associated to increased odds of NAFLD and gallstone disease; however, these relationships were not statistically significant after additional adjustment for waist circumference, total energy intake, and physical inactivity (NAFLD: OR = 1.37, 95% CI: 0.68-2.75, p = 0.384; gallstone disease: OR = 1.83, 95% CI: 0.93-3.61, p = 0.080). Participants following a Western dietary pattern exhibited higher levels of ALT, triglycerides, total cholesterol, fasting glucose, and malondialdehyde, along with lower HDL-C concentrations (all p < 0.001). Most of the participants with overweight, obesity, or abdominal obesity showed the strongest associations.

CONCLUSION: These results shed more light on the association between dietary patterns and hepatobiliary health. More research on dietary-pattern-oriented approaches is necessary to fully comprehend their potential impact in hepatobiliary health.

PMID:42643230 | PMC:PMC13503224 | DOI:10.3389/fnut.2026.1789238

Front Endocrinol (Lausanne). 2026 Aug 11;17:1888652. doi: 10.3389/fendo.2026.1888652. eCollection 2026.

ABSTRACT

BACKGROUND: Metabolic dysfunction-associated steatotic liver disease (MASLD; formerly termed non-alcoholic fatty liver disease) and hypertension frequently coexist within the same cardiometabolic context. The metabolic score for insulin resistance (METS-IR), which is derived from body mass index, fasting blood glucose, triglycerides, and high-density lipoprotein cholesterol without requiring insulin measurement, may serve as a practical auxiliary indicator of metabolic dysfunction.

OBJECTIVE: This study investigated the association between METS-IR and MASLD among adults with hypertension and evaluated the dose-response pattern, non-linear features, discrimination, and effect modification.

METHODS: This secondary cross-sectional analysis used publicly available data from a Chinese single-center health check-up cohort. A total of 943 adults with hypertension were included. MASLD was operationally defined as ultrasonographically detected hepatic steatosis in adults with hypertension after exclusion of excessive alcohol intake and other known causes of liver disease. The primary multivariable model was adjusted for age group, sex, tobacco use, alcohol use, and diabetes; an extended model additionally included uric acid, total cholesterol, low-density lipoprotein cholesterol, and fat-to-muscle ratio. Logistic regression, quartile analysis, restricted cubic spline analysis, sensitivity analyses, receiver operating characteristic analysis, and subgroup interaction analyses were performed.

RESULTS: Among the 943 participants, 640 had MASLD, corresponding to a prevalence of 67.9%. The prevalence of MASLD increased across METS-IR quartiles from 40.7% to 67.2%, 74.6%, and 89.0%. After adjustment for age group, sex, tobacco use, alcohol use, and diabetes, each 1-standard deviation increase in METS-IR was associated with higher odds of prevalent MASLD (odds ratio [OR] = 2.78, 95% confidence interval [CI]: 2.24-3.44, P < 0.001). Compared with quartile 1, the ORs for quartiles 2, 3, and 4 were 2.99, 4.01, and 9.72, respectively. In the extended model, the corresponding OR per 1-standard deviation increase was 2.47 (95% CI: 1.91-3.18, P < 0.001). Restricted cubic spline analysis suggested mild non-linearity. Exploratory ROC analysis yielded an area under the curve of 0.738 (95% CI: 0.705-0.772); the data-derived METS-IR cut-point of 42.40 had a sensitivity of 53.0% and a specificity of 81.8%. A borderline interaction by sex was observed (P for interaction = 0.046), whereas interactions by age group and diabetes status were not statistically significant.

CONCLUSION: Among adults with hypertension in a Chinese single-center health check-up population, higher METS-IR was associated with a higher likelihood of prevalent MASLD and showed a graded, mildly non-linear pattern. METS-IR may serve as an accessible auxiliary metabolic indicator, but its data-derived cut-point and possible sex-specific association require independent validation.

PMID:42643228 | PMC:PMC13503243 | DOI:10.3389/fendo.2026.1888652

Front Nutr. 2026 Aug 11;13:1806413. doi: 10.3389/fnut.2026.1806413. eCollection 2026.

ABSTRACT

BACKGROUND: Knee osteoarthritis (OA) is a common cause of pain and disability without a cure. Glucosamine is widely used as a symptomatic treatment owing to its safety, yet its clinical efficacy remains uncertain amid mixed trial results and conflicting guidelines. This umbrella review synthesized evidence from published systematic reviews and meta-analyses on the effectiveness and safety of glucosamine in knee OA.

METHODS: We searched MEDLINE (via PubMed), Embase, Web of Science, and the Cochrane Library up to November 2025. Eligible studies were systematic reviews or meta-analyses of randomized trials in adults with knee OA comparing glucosamine (any formulation) with placebo or other non-surgical comparators and reporting pain, physical function, or joint structure. Two reviewers independently screened, extracted data, and assessed quality (AMSTAR-2). Key outcomes were pain (visual analog scale [VAS] and WOMAC [Western Ontario and McMaster Universities Osteoarthritis Index]), functional measures, joint space narrowing (JSN), and adverse events. Overlap among shared primary trials was quantified using the corrected covered area, and the resulting dependence was addressed using robust variance estimation to pool standardized mean differences (SMDs), with p < 0.05 considered significant. Effect sizes were interpreted against minimal clinically important difference (MCID) thresholds.

RESULTS: Nineteen reviews were included (most low or critically low quality). Glucosamine produced a small but statistically significant reduction in VAS pain (SMD – 0.36, p = 0.02), whereas WOMAC pain showed no significant difference. Functional outcomes (WOMAC total and function) showed no significant improvement. Glucosamine was associated with significant slowing of JSN progression (SMD – 0.32, p = 0.01), though this signal derived mainly from long-duration sulfate trials. The magnitude of the significant effects was at or below accepted MCID thresholds. Adverse event rates did not differ from placebo.

CONCLUSION: Glucosamine demonstrated at most minimal-to-modest efficacy for knee OA, yielding a small pain reduction and possible slowing of radiographic progression but no clear functional benefit, with effects at or below the threshold of clinical importance. Given the low quality of evidence, extensive overlap among reviews, and the predominantly out-of-pocket nature of glucosamine use, these findings should be interpreted with caution and weighed against cost in shared decision-making.

SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/prospero/, identifier (CRD420251207513).

PMID:42643227 | PMC:PMC13503159 | DOI:10.3389/fnut.2026.1806413