搜索结果：Frontiers in endocrinology

Birth weight is a strong determinant of long-term metabolic health, with both low birth weight and macrosomia linked to increased cardiometabolic risk. Fat-soluble vitamins A, D, and E modulate pathways relevant to fetal growth, however, their trajectories and potential associations with birth weight in pregnant individuals with overweight or obesity remain scarcely characterized. To examine maternal vitamins A (retinol), D (25(OH)D), and E (&#x3b1;-tocopherol) concentrations during the second and third trimesters of pregnancy and their associations with birth weight, in individuals with pre-pregnancy overweight or obesity. This secondary analysis of the Exercise Training in Pregnancy Trial (ETIP) study included 57 mother-infant pairs with available vitamin measurements in the second and third trimesters. Plasma retinol and &#x3b1;-tocopherol were measured by high performance liquid chromatography, and serum 25(OH)D by liquid chromatography-tandem mass spectrometry. Birth weight was classified as normal (2500 - < 4000&#xa0;g) or macrosomic (&#x2265; 4000&#xa0;g). Circulating vitamin concentrations between trimesters were compared using the Wilcoxon signed-rank test and associations between vitamins and birth weight were examined using multivariate linear regression models. From the second to the third trimester, mean maternal retinol and 25(OH)D concentrations declined significantly (retinol: 1.52 (SD&#xa0;=&#xa0;0.37) to 1.32 (SD&#xa0;=&#xa0;0.33) &#xb5;mol/L; 25(OH)D: 73.7 (SD&#xa0;=&#xa0;30.0) to 63.3 (SD&#xa0;=&#xa0;25.0) nmol/L), whereas &#x3b1;-tocopherol increased from 34.6 (SD&#xa0;=&#xa0;7.1) to 46.3 (SD&#xa0;=&#xa0;10.0) &#xb5;mol/L. In the third trimester, 19.3% had vitamin A insufficiency, while vitamin D deficiency and insufficiency affected 31.6% and 33.3%, respectively. Macrosomia occurred in 43.9% of infants and 56.1% had birth weight within the normal range. Maternal vitamins A, D, and E were not associated with birth weight, and no vitamin A and D interaction was observed. In pregnant individuals with overweight or obesity, maternal vitamin A and D concentrations declined across pregnancy, while vitamin E increased. Vitamin A insufficiency, vitamin D deficiency/insufficiency, and macrosomia were common. Maternal fat-soluble vitamin levels were not independently associated with birth weight, suggesting that vitamin status during mid- to late pregnancy may not be a major determinant of fetal growth in this metabolically high-risk population.

This study aimed to evaluate the effect of a 50% reduction in preprandial bolus insulin (50%-B) on plasma glucose (PG) responses during postprandial exercise of continuous moderate intensity (CONT) and intermittent high intensity (INT) in individuals with type 1 diabetes (T1D). Sixteen adults with T1D (31% male), treated with multiple daily insulin injections (MDI), participated in a randomized crossover study comprising four experimental conditions, separated by a washout period of at least 48 hours. Participants performed two 30-minute, preceded by a 3-minute warm-up without weights:&#x2022; CONT: continuous cycling at 60% of maximal aerobic power (MAP).&#x2022; INT: 2-minute intervals alternating between 40% and 80% of MAP, repeated for 7 intervals, with the last interval adjusted so that the total exercise time is exactly 30 minutes. Each exercise modality was performed under two insulin conditions: a full preprandial bolus (100%-B) and a 50% reduction (50%-B). Plasma glucose, insulin, and cortisol were measured before, during, and after exercise. Linear mixed models were used to analyze temporal changes and condition effects. Blood glucose decreased significantly over time for both exercise types (p < 0.001). During CONT, the decline in PG was similar between doses (&#x394;100%-B: -3.01 &#xb1; 2.96 vs. &#x394;50%-B: -2.82 &#xb1; 2.28 mmol/L; p = 0.18), However, the nadir PG was higher with 50%-B compared to 100%-B (8.59 &#xb1; 4.07 vs. 5.69 &#xb1; 3.06 mmol/L, respectively; &#x3b2; = +2.91 mmol/L; p = 0.026), and hypoglycemia was less frequent (2 vs. 18 episodes; p = 0.028). During INT, PG decreased less with 50%-B than with 100%-B (&#x394;: -2.03 &#xb1; 1.63 vs. -3.62 &#xb1; 2.76 mmol/L; p = 0.022), with no hypoglycemic episodes under 50%-B compared to six with 100%-B. Mean PG remained higher with 50%-B across both exercise types (p < 0.01). Plasma insulin decreased over time (p = 0.038) regardless of bolus condition, while cortisol increased more during INT with 100%-B than with 50%-B (p = 0.02). Reducing the preprandial bolus insulin by 50% effectively attenuates exercise-induced declines in plasma glucose and substantially reduces hypoglycemia risk, particularly during intermittent high-intensity exercise. These results emphasize the clinical relevance of personalized insulin adjustments to enhance metabolic safety during exercise in individuals with T1D.

Currently, numerous studies have employed machine learning (ML) methods to develop predictive models for depression risk in patients with diabetes mellitus (DM); however, the findings remain inconsistent. Therefore, this study aims to clarify the current state of research and emerging trends in this field by systematically evaluating the performance, strengths, and limitations of existing prediction models. This systematic review evaluates the performance and clinical applicability of ML-based depression risk prediction models for patients with DM, providing reliable evidence to assist healthcare professionals in selecting and optimizing more appropriate prediction models. We conducted a systematic search of clinical studies employing ML approaches to predict depression risk in patients with DM across the PubMed, Embase, Cochrane Library, and Web of Science databases, from their inception to January 2026. The primary performance metric for the models was the area under the receiver operating characteristic curve (AUC) along with its 95% confidence interval (95% CI). Two independent researchers screened the literature, extracted data, and used PROBAST-AI to assess the risk of bias and clinical applicability of the included studies. Pooled AUC was estimated using the Der Simonian and Laird random-effects model. A total of 14 studies comprising 64 distinct ML models were included. All included studies were assessed as high risk of bias and high clinical applicability. A pooled analysis of the best-performing ML prediction models reported in each study showed a pooled AUC of 0.822 (95% CI, 0.789-0.858), indicating relatively good overall predictive performance. However, there was substantial heterogeneity among the studies (I&#xb2; =&#xa0;97.4%; P <&#xa0;0.001). Subgroup analysis based on ML model types revealed the following pooled AUC values: 0.765 (95% CI 0.706-0.829) for traditional regression models, 0.789 (95% CI 0.747-0.834) for general machine learning models, and 0.802 (95% CI 0.769-0.836) for deep learning models. Notably, logistic regression (LR) (n = 10) was the most frequently employed ML method for developing depression risk prediction models in patients with DM. To evaluate model generalizability and avoid overfitting, the included studies adopted three validation strategies: 5-fold cross-validation yielded a pooled AUC of 0.913 (95% CI 0.781-1.067), 10-fold cross-validation yielded 0.819 (95% CI 0.781-0.858), and random split validation yielded 0.747 (95% CI 0.648-0.862). The most commonly used predictors in the included models were age, sex, and body mass index (BMI), which are readily available in clinical settings and strongly associated with depression risk. ML-based depression risk prediction models for patients with DM demonstrate overall satisfactory predictive performance. However, most existing studies had relatively small sample sizes and lacked external validation. Future research should prioritize refining study design and optimizing clinical data processing to improve the generalizability and stability of these models in clinical practice. https://www.crd.york.ac.uk/PROSPERO/view/CRD420251243343, identifier CRD420251243343.

Diabetic kidney disease (DKD) remains a leading cause of end-stage renal disease despite advances in glucose-, blood pressure-, and albuminuria-lowering therapies. The glucose-responsive transcription factor carbohydrate response element-binding protein (ChREBP; encoded by MLXIPL) regulates glycolytic-lipogenic programs, yet its causal contribution to renal injury is challenging to disentangle in advanced DKD, where bulk kidney transcriptomes reflect tissue remodeling and cellular compositional shifts. We integrated two-sample Mendelian randomization (MR), kidney transcriptomic stratification, network analyses, and experimental validation. MR used blood cis-eQTL instruments for MLXIPL to estimate causal effects on type 2 diabetes (T2D) and urinary albumin-to-creatinine ratio (UACR), including a non-diabetic UACR stratum. In kidney transcriptomics (GSE30529), we evaluated remodeling-related confounding and applied within-DKD, median-based MLXIPL-high/low stratification for GSEA/GSVA and functional/network inference. Key observations were validated in db/db mice and primary proximal tubular epithelial cells (PTECs) exposed to high glucose with matched osmotic control. Genetically predicted higher MLXIPL expression was associated with increased T2D risk across multiple phenotype definitions and with higher UACR, including replication in non-diabetic individuals. Within DKD, MLXIPL heterogeneity tracked metabolic programs by GSEA and divergent pathway activity by GSVA, while signatures related to profibrotic and proliferative remodeling were concomitantly enriched in the low-MLXIPL subgroup. Network analyses positioned MLXIPL/ChREBP within a dense metabolic interaction and regulatory landscape. Experimentally, ChREBP and Mlxipl were increased in db/db kidneys and induced by high glucose in PTECs, accompanied by coordinated upregulation of lipogenic targets (Acly, Acaca, Fasn, Srebf1) and an inverse relationship with Ppargc1b. Integrating genetic inference, confounding-aware kidney transcriptomics, network biology, and experimental validation, our study supports MLXIPL/ChREBP as a pathogenic nutrient-sensing node linking diabetes susceptibility to renal injury and maladaptive metabolic remodeling in DKD, providing a mechanistic rationale for targeting this axis to mitigate residual renal risk.

Recent studies suggest that the development of prediabetes and its associated comorbidities may depend on sex and reproductive status. While the exact mechanism is unclear, differences in insulin sensitivity, body fat distribution, and glucose and lipid metabolism may play a role. In this study, we investigated how sex differences in metabolic and inflammatory parameters affect the development of prediabetic conditions in a non-obese rat model with severe dyslipidaemia. Wistar Kyoto (WKY) rats served as the control group, while age-matched Hereditary Hypertriglyceridaemic (HHTg) rats were used as a non-obese, prediabetic model with genetically determined hypertriglyceridaemia, insulin resistance and impaired glucose tolerance. Compared to WKY controls, the HHTg strain exhibited increased serum triacylglyceroles (TAG) as well as ectopic TAG accumulation in the liver, heart and skeletal muscle which was more pronounced in HHTg females. However, this higher ectopic TAG accumulation in HHTg females was not associated with increased lipotoxic diacylglyceroles. The HHTg strain showed increased visceral adiposity, which was distributed differently: HHTg females had increased perimetrial adipose tissue, while HHTg males had increased perirenal adipose tissue. Impaired insulin sensitivity was observed in both sexes of the HHTg strain in skeletal muscle and adipose tissue. Insulin resistance in the HHTg strain may be due to elevated leptin and NEFA levels, as well as decreased GLUT4 in skeletal muscle. In addition, the HHTg strain showed impaired glucose tolerance, as well as hyperinsulinaemia, which was more pronounced in HHTg males. Increased lipogenesis (mRNA Scd1), oxidative stress (decreased SOD activity) and inflammation (mRNA Tnf&#x3b1;) in the liver may contribute to the development of hepatic steatosis and hepatic lipid accumulation. In visceral adipose tissue, increased mRNA Hif1 may contribute to adipose tissue hypoxia and impair insulin sensitivity, particularly in males. Despite having more pronounced dyslipidaemia, ectopic lipid accumulation, and visceral adiposity, prediabetic females have better glucose tolerance and insulin sensitivity markers than prediabetic males. These sex differences may be due to variations in fat distribution, lipid metabolism and chronic inflammation. Our findings suggest that males are more susceptible to developing early prediabetic damage, such as insulin resistance and fatty liver, regardless of obesity.

Prader-Willi syndrome (PWS) is a neuroendocrine disorder characterized by hypothalamic dysfunction, congenital hypotonia, abnormal growth trajectories, and impaired pubertal development, all of which contribute to a markedly increased risk of scoliosis, with a cumulative prevalence reaching up to 70-80% by skeletal maturity, significantly exceeding that of idiopathic scoliosis. Unlike idiopathic scoliosis, spinal deformity in PWS follows a distinct bimodal pattern, with critical vulnerability during infancy and a second acceleration during pubertal transition. Growth hormone (GH) therapy, a cornerstone of PWS management, substantially improves linear growth, body composition, and muscle strength, yet its relationship with scoliosis onset and progression remains a clinical challenge due to the potential for accelerated growth during critical developmental windows, which may unmask or exacerbate underlying spinal instability. Current scoliosis surveillance strategies in PWS are largely extrapolated from idiopathic scoliosis and fail to account for the unique neuroendocrine and biomechanical context of this syndrome. In particular, endocrine modifiers such as GH treatment status, growth velocity, hypogonadism, pubertal stage, body composition, and genotype-specific phenotypes are rarely integrated into structured monitoring protocols. In this narrative review, we synthesize epidemiological, mechanistic, and clinical evidence to elucidate the neuroendocrine and biomechanical pathways underlying scoliosis development in PWS, including the roles of hypotonia-related instability, altered vertebral growth modulation, and delayed epiphyseal maturation. We critically examine the dualistic effects of GH therapy, the impact of pubertal maturation, and genotype-phenotype associations as key determinants of scoliosis risk and progression. Building on this evidence, we propose an endocrine-informed, risk-stratified scoliosis monitoring framework that integrates growth dynamics, hormonal status, body composition, and spinal parameters to guide surveillance intensity, imaging strategies, and multidisciplinary referral. By shifting the focus from isolated curve detection to longitudinal, endocrine-guided surveillance, this review provides a clinically actionable model to optimize early identification and management of scoliosis in children and adolescents with PWS. This framework aims to support coordinated endocrine-orthopedic care and inform future prospective studies designed to refine outcome measures and ultimately improve long-term musculoskeletal and quality-of-life outcomes in this vulnerable population.

Diabetic retinopathy (DR) remains a leading cause of blindness among working-age adults, yet scalable risk stratification tools tailored to primary care are lacking-particularly in underserved settings where specialized examinations are unavailable. We aimed to develop and externally validate a pragmatic, web-based nomogram for DR risk prediction using only routinely collected electronic health record (EHR) variables in community-dwelling individuals with type 2 diabetes (T2DM). This retrospective cohort study analyzed EHR data from two independent Chinese populations. The primary cohort comprised 1,215 T2DM patients from 45 community health centers in Shenzhen, randomly split into training (n=851) and internal validation (n=364) sets. An external validation cohort of 329 patients was obtained from a center in Nanjing. Candidate predictors were screened via univariate analysis and least absolute shrinkage and selection operator (LASSO) regression within the training set. Selected variables were entered into multivariable logistic regression to construct a nomogram, which was deployed as an interactive web application. Model performance was assessed using the area under the receiver operating characteristic curve (AUC-ROC), calibration plots, decision curve analysis (DCA), and clinical impact curves (CIC). Three predictors-diabetes duration, HbA1c, and high body mass index (BMI &#x2265;24 kg/m&#xb2;, Chinese standard)-were retained in the final model. The model demonstrated robust discrimination: AUC was 0.77 (95% CI: 0.73-0.81) in the training set, 0.79 (0.73-0.85) in internal validation, and 0.81 (0.75-0.87) in external validation. Calibration was adequate, with non-significant Hosmer-Lemeshow tests (P > 0.05) and Brier scores below 0.15 across all cohorts. DCA confirmed positive net benefit over a wide range of threshold probabilities (10-95%), and CIC revealed a 1:1 ratio between predicted and observed DR cases at risk thresholds above 40%. This three-parameter online nomogram provides a simple, readily implementable tool for DR risk stratification in primary care. Its robust external validation in an independent cohort and reliance on variables universally available in EHRs position it as a cost-effective solution to bridge the screening gap and enable timely specialist referral for high-risk T2DM patients.

With expanding applications of artificial intelligence (AI) within the research pipeline of endocrinology, it is essential that journals uphold explicit AI usage policies that maintain the rigor and integrity of published research. In this review, we aim to evaluate current AI policies of leading endocrinology journals to assess the current landscape of research and the implications of its progression. We conducted a cross-sectional review of the top endocrinology journals using the SCImago Journal Ranking (SJR) database. From November 2024 to July 2025, we reviewed AI usage guidelines from publicly available Instructions for Authors, including authorship, manuscript writing, and content/image generation. We also assessed whether journals endorsed AI-specific reporting guidelines (e.g., CONSORT-AI, SPIRIT-AI). Data were extracted independently and in duplicate using a standardized form. Reproducibility was supported through protocol registration on Open Science Framework. Of the top 100 endocrinology journals, 84.0% (84/100) mentioned AI in their Instructions for Authors and 79.0% (79/100) required disclosure of AI use during submission. Although no journals (0/100) permitted AI tools for authorship, 64.0% (64/100) allowed its use in manuscript writing, 22.0% (22/100) for content generation, and 50.0% (50/100) for image generation. Despite these guidelines, only one (1.0%; 1/100) journal required a specific reporting guideline, and very few endorsed AI statements by the IMCJE (9/100), COPE (12/100), or WAME (0/100). No statistically significant correlations were identified between AI usage policies and SJR or impact factor. Many leading endocrinology journals have addressed AI use; however, their policies remain incomprehensive. It is critical that publishers and their journals establish explicit guidelines regarding the use of AI tools to promote transparent, reproducible, and reliable research.

Polycystic ovary syndrome (PCOS) is a common endocrinopathy affecting women of reproductive age, characterized by oligo- or anovulation, hyperandrogenism, and polycystic ovarian morphology. Beyond its reproductive manifestations, PCOS is increasingly recognized as a complex endocrine-metabolic disorder frequently associated with impaired carbohydrate metabolism and insulin resistance, often independent of body mass. Despite extensive research, the molecular mechanisms underlying insulin resistance across metabolic and reproductive tissues in PCOS remain incompletely characterized. This scoping review aimed to systematically map molecular disturbances in insulin signaling and carbohydrate metabolism in PCOS, explore associations between tissue-specific mechanisms, and identify key gaps in the current evidence. We included peer-reviewed original studies published in English between January 2018 and May 2025, retrieved from PubMed, Embase, and Web of Science, that investigated molecular or cellular pathways related to insulin resistance or glucose metabolism in PCOS. The available evidence predominantly addressed granulosa cells and ovarian tissue, with additional data from endometrium, liver, adipose tissue, skeletal muscle, pancreatic beta-cells, and systemic regulatory pathways. Recurrent mechanisms underlying insulin resistance in PCOS included post-receptor defects in IRS/PI3K/AKT and MAPK signaling, impaired GLUT4 expression and trafficking, mitochondrial and glycolytic dysfunction, chronic low-grade inflammation, androgen receptor-mediated metabolic reprogramming, circadian rhythm disruption, and epigenetic or environmental modulators. Evidence from human studies remains limited, with many proposed molecular mechanisms being supported predominantly by rodent or cell line models. To translate this knowledge to clinical and therapeutic application, additional high-quality longitudinal human research with comprehensive multi-omics is necessary to validate key mechanisms in ovarian and metabolic tissues, especially those involving IRS/PI3K/AKT signaling, GLUT4 regulation, inflammation, and androgen-driven metabolic dysfunction.

Obesity is increasingly recognized as an immunometabolic disorder driven by dysregulated crosstalk between visceral adipose tissue and the liver, particularly along the liver-omentum axis, which promotes insulin resistance and hepatic steatosis. Although Chaihu-Wendan Decoction (CHWD) is effective for metabolic disorders, its molecular mechanisms of action on this inflammatory axis remains unclear. This study aimed to investigate the therapeutic mechanisms of CHWD in high-fat diet (HFD)-induced obesity model, specifically focusing on insulin signaling and immune microenvironment remodeling in the liver-omentum axis. C57BL/6J mice were fed a HFD to induce obesity and treated with CHWD. Metabolic phenotypes were assessed via biochemical and histological analyses. The molecular mechanisms were explored by evaluating the PTEN/PI3K/AKT/mTOR pathway and omental macrophage profiles using Western blot, ELISA, and immunohistochemistry. CHWD treatment significantly ameliorated HFD-induced body weight gain, dyslipidemia, and hepatic steatosis. Mechanistically, CHWD acted as a regulator of PTEN-associated signaling, which triggered a dual-regulation of PTEN/AKT/mTOR signaling, i.e., robust reactivation of upstream insulin signaling (INSR/IRS1/PI3K/AKT) coupled with the paradoxical suppression of downstream mTOR phosphorylation. This "uncoupling" process restored insulin sensitivity without promoting lipogenesis. Concurrently, CHWD remodeled the omental immune microenvironment by restoring omentin-1 secretion and promoting macrophage phenotype switching, characterized by maintenance of a CD68+ macrophage population accompanied by suppression of iNOS-mediated cytotoxic effector functions. CHWD alleviates HFD-induced obesity and metabolic inflammation by coordinately targeting the PTEN/AKT/mTOR axis and reprogramming omental immunity. These findings provide the primary evidence supporting that CHWD modulates the liver-omentum axis via distinct signaling and immune mechanisms, offering a novel therapeutic strategy for metabolic syndrome.

Idiopathic nephrotic syndrome (INS) is a glomerular disorder characterized by proteinuria, hypoalbuminemia, and edema, and relapse remains a major clinical challenge. Early prediction of relapse risk may facilitate individualized treatment and follow-up. This study aimed to develop and compare the performance of logistic regression, random forest, and deep learning models for predicting relapse in adult patients with INS using baseline clinical and laboratory data. We conducted a retrospective cohort study of 562 adult patients with idiopathic nephrotic syndrome treated between January 2022 and January 2024. The primary outcome was the first relapse within 12 months after baseline assessment. Baseline demographic characteristics, clinical history, laboratory parameters, and treatment-related variables were collected. The dataset was randomly divided into training (70%), validation (15%), and test (15%) sets. Missing data were imputed, continuous variables were standardized as appropriate, and SMOTE was applied to the training set only to address class imbalance. Three predictive models were developed: logistic regression, random forest, and a deep learning-based neural network. Model performance was evaluated using AUC, accuracy, sensitivity, specificity, and F1-score. Among the three models, the deep learning model showed the best predictive performance, with AUCs of 0.908, 0.900, and 0.883 in the training, validation, and test sets, respectively. The logistic regression model showed intermediate performance, whereas random forest showed the lowest discriminatory ability. The most influential predictors of relapse included steroid resistance, nephrotic-range proteinuria at baseline, prior relapse history/frequency, elevated ESR, and immunosuppressant use. Deep learning demonstrated better predictive performance than logistic regression and random forest for predicting 12-month relapse in adult patients with idiopathic nephrotic syndrome. These findings suggest that machine learning-based models, particularly deep learning, may serve as useful tools for relapse risk stratification. External validation in larger independent cohorts is needed before clinical implementation.

Knee osteoarthritis (KOA) is characterized by progressive cartilage degeneration and disruption of extracellular matrix (ECM) homeostasis. Chondrocytes are not a homogeneous or static population during disease progression but exhibit pronounced functional heterogeneity. Although single-cell transcriptomic studies have identified multiple chondrocyte states in osteoarthritic cartilage, how these states dynamically relate to ECM remodeling and disease progression remains incompletely understood. We integrated multiple publicly available single-cell RNA sequencing datasets of human knee cartilage to construct a unified cellular atlas and systematically compared chondrocyte states between control and KOA samples. Differential expression analysis, functional enrichment, pseudotime trajectory inference, and cell-cell communication analysis were applied to characterize ECM-related chondrocyte states and their dynamic transitions. Key signaling cues identified from single-cell analyses were further evaluated using in vitro cultured human articular chondrocytes. We observed a marked expansion of the previously described reparative chondrocyte population (RepC) in KOA cartilage. Rather than reflecting a simple increase in cell proportion, KOA-associated RepC exhibited enhanced ECM remodeling programs characterized by collagen reorganization and strengthened ECM-cell interactions. Pseudotime analysis positioned RepC downstream of proliferation chondrocytes and near a major branching region toward either regulator chondrocytes with further extension toward fibrocartilage chondrocytes or effector chondrocytes. In KOA, RepC was preferentially represented at mid-to-late pseudotime stages within the reconstructed trajectory framework. Cell-cell communication analysis suggested that RepC showed prominent inferred ECM-related interactions, particularly involving collagen and FN1-integrin pathways. Consistently, FN1 or TGF-&#x3b2;1 stimulation in vitro induced expression of multiple RepC-associated genes and enhanced SMAD2/3 phosphorylation, recapitulating key features of the RepC state observed in single-cell analyses. These findings highlight ECM remodeling features of reparative chondrocytes during KOA and support a state-centric view in which disproportionate representation of reparative states within the pseudotime trajectory framework is associated with maladaptive ECM remodeling in KOA.

This study aimed to evaluate the clinical utility of novel inflammatory and metabolic composite indices in early risk prediction of microvascular complications in patients with type 2 diabetes mellitus (T2DM), and to provide reliable evidence for early precision risk stratification. A retrospective analysis was conducted on 964 hospitalized patients with T2DM admitted to the Department of Endocrinology, First Affiliated Hospital of Xinjiang Medical University, from September 2023 to March 2025. Patients were randomly assigned to a training cohort and a validation cohort at a ratio of 7:3 using a random number table. In the training cohort, least absolute shrinkage and selection operator (LASSO) regression was applied for variable selection and to reduce multicollinearity, followed by univariate and multivariate logistic regression analyses to identify independent risk factors for T2DM related microvascular complications. Receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA) were employed to comprehensively assess the predictive performance and clinical utility of the model. Multifactorial logistic regression analysis showed that age, duration of diabetes, duration of hypertension, urine albumin-to-creatinine ratio (UACR) > 30 mg/g, as well as core indicators SIRI and TyG index, were significantly associated with the occurrence of microvascular complications in type 2 diabetes mellitus (T2DM) (P < 0.05). The predictive model constructed based on LASSO-logistic regression demonstrated an AUC of 0.869 (95% CI: 0.842-0.895) in the training set and an AUC of 0.864 (95% CI: 0.824-0.905) in the validation set, indicating stable and excellent discriminatory ability. This study confirms that SIRI and TyG index can serve as independent risk factors for microvascular complications in T2DM. The nomogram model constructed based on LASSO-logistic regression shows significantly better predictive performance than single indicators, with good model calibration, demonstrating excellent clinical net benefit. This model can accurately assess the risk of microvascular complications, providing reliable decision support for early clinical screening and risk stratification management.

Primary aldosteronism (PA) is associated with a substantially higher cardiovascular risk than essential hypertension, a disparity that cannot be fully explained by blood pressure elevation alone. Clinical studies consistently demonstrate that cardiovascular morbidity and mortality often persist in patients with PA despite adequate blood pressure control and standard therapy, underscoring the existence of residual cardiovascular risk. Accumulating experimental and clinical evidence identifies inflammation as a central mediator of aldosterone-induced cardiovascular injury. Excess aldosterone drives immune-inflammatory remodeling through coordinated activation of innate and adaptive immune responses, including macrophage- and T cell-dependent pathways, as well as downstream signaling cascades such as inflammasome activation and interleukin-6-related trans-signaling. These processes promote myocardial fibrosis, vascular dysfunction, and adverse cardiac remodeling, providing a mechanistic basis for the heightened cardiovascular risk observed in PA. Although mineralocorticoid receptor (MR) antagonists remain the cornerstone of medical therapy for PA, MR blockade alone may be insufficient to fully suppress aldosterone-driven inflammatory and non-hemodynamic effects. Persistent activation of these pathways offers a plausible explanation for the residual cardiovascular risk observed in treated patients. Emerging therapeutic strategies aim to overcome these limitations through combination approaches. Aldosterone synthase inhibitors (ASIs), by targeting aldosterone production upstream, may complement MR antagonism, while interventions directed at inflammatory pathways and non-genomic aldosterone signaling could further enhance cardiovascular protection. This review integrates current mechanistic and clinical evidence on inflammatory drivers and residual risk in PA and discusses emerging combination strategies to optimize cardiovascular risk reduction in this high-risk population.

Acquired hypothalamic obesity (aHO) is a disease characterized by rapid, clinically significant, and persistent weight gain resulting from damage to hypothalamic structures. aHO is associated with substantial morbidity, increased mortality, and marked impairment in quality of life. Etiologies include craniopharyngioma and other space-occupying lesions of the sellar/parasellar region, neurosurgical procedures, cranial irradiation, and traumatic brain injury. A multidisciplinary panel comprising ten specialists in neuroendocrinology, neurooncology, and neurosurgery from Germany, Austria, and Switzerland convened in Frankfurt am Main, Germany, on November 10, 2025, to discuss contemporary challenges and advances in this field. aHO should be conceptualized and treated within the broader clinical entity of hypothalamic syndrome, a complex disorder involving multiple neuroendocrine deficiencies, disturbances of circadian regulation, impaired control of hunger, satiety, and thirst, altered thermoregulation, and a range of cognitive, sleep-related, and psychosocial dysfunctions. Long-term outcomes for affected individuals are frequently unfavorable, largely due to increased risks of metabolic syndrome, cardiovascular disease, profound reductions in health-related quality of life, and elevated rates of premature mortality. The management of hypothalamic syndrome remains particularly challenging. Pharmacological strategies, including dextroamphetamine and glucagon-like peptide-1 receptor agonists, have demonstrated potential benefits for weight and hyperphagia-related outcomes. Recently, preliminary findings from a prospective, randomized, placebo-controlled clinical trial (TRANSCEND) provided encouraging evidence for the efficacy of setmelanotide, a melanocortin-4 receptor agonist. This perspectives report reviews clinical advances in epidemiology, diagnostics, treatment, and follow-up of patients with aHO and outlines key directions for future research aimed at improving outcomes in this vulnerable population.

Beyond its primary digestive functions, the stomach serves as an endocrine organ, secreting peptides that regulate appetite and energy balance. Among its enteroendocrine populations, X/A-like cells play a pivotal role in controlling food intake, glucose homeostasis, and lipid deposition. The secretion of X/A-like cell-derived hormones, including ghrelin and nesfatin-1, is regulated by the mechanistic target of rapamycin (mTOR) signaling pathway. However, the role of X/A-like cell mTOR signaling in skeletal metabolism remains unexplored. Using previously validated and published mouse models with X/A-like cell-specific deletion of Mtor or its upstream inhibitor Tsc1, we assessed bone phenotypes at 12 and 40 weeks of age under chow-fed conditions. Skeletal effects were also evaluated under pathological bone loss conditions, including estrogen deficiency (ovariectomy) and caloric restriction. Our findings demonstrate that mTOR signaling deficiency in X/A-like cells compromises bone health in male mice, evidenced by cortical bone loss at 12 weeks and trabecular bone reductions at 40 weeks. Furthermore, X/A-like cell-specific Mtor deletion significantly exacerbated bone loss in female mice following ovariectomy, impacting both trabecular and cortical parameters. In contrast, activation of mTOR signaling via Tsc1 deletion in X/A-like cells did not alter bone mass under either chow ad libitum or calorie-restricted conditions. Collectively, these findings identify a previously unrecognized role of gastric X/A-like cell mTOR signaling in the regulation of bone metabolism. Maintenance of intact mTOR signaling in these endocrine cells is necessary for bone homeostasis, revealing a novel gut-bone endocrine axis.

Metabolic diseases represent a significant global public health concern, imposing substantial burdens on healthcare systems, economies, and patient quality of life. Current treatments have limitations, underscoring the need for safer alternatives. Quercetin, a natural flavonoid with favorable human tolerability, shows promise for metabolic disorder management. This review critically evaluates the existing evidence on quercetin's role in metabolic disease management, summarizing its pharmacological advancements and clinical data in treating nine metabolic disorders: diabetes mellitus (DM), metabolic dysfunction-associated fatty liver disease (MAFLD), obesity, atherosclerosis, hyperuricemia, gouty arthritis, hyperlipidemia, osteoporosis, and polycystic ovary syndrome (PCOS). We systematically reviewed studies (2003-2025) from Web of Science, PubMed, Science Direct, and CNKI reporting quercetin's effects in metabolic diseases. Quercetin exhibits multifaceted pharmacological activities, including anti-inflammatory, antioxidant, antiapoptotic, hypolipidemic, and hypoglycemic effects. This underpins its therapeutic potential against nine metabolic disorders. Furthermore, emerging nanodelivery systems have demonstrated enhanced bioavailability, stability, and overall efficacy of quercetin while mitigating its dose-dependent toxicity. Quercetin shows considerable promise in the intervention of metabolic diseases. However, current research lacks mechanistic depth, bioavailability enhancement data, and clinical validation Additionally, clinical studies validating its therapeutic efficacy remain scarce. Further mechanistic investigations and randomized controlled trials are imperative to elucidate quercetin's precise mechanisms and substantiate its clinical potential in metabolic disease management.

This study aimed to develop and evaluate machine learning (ML) models for predicting non-alcoholic fatty liver disease (NAFLD) in patients with type 2 diabetes mellitus (T2DM) using readily accessible clinical and biochemical indicators. A total of 2,459 patients with T2DM were enrolled in this cross-sectional study. Eight ML algorithms, logistic regression (LG), k-nearest neighbors (k-NN), support vector machine (SVM), decision tree (DT), random forest (RF), gradient boosting machine (GBM), extreme gradient boosting (XGBoost), and na&#xef;ve Bayes (NB), were developed to construct predictive models. Feature selection was performed using Boruta, recursive feature elimination, and LASSO regression. Model performance was assessed using several metrics, including the area under the receiver operating characteristic curve (AUC), accuracy, recall, F1 score, and decision curve analysis. Among the study population, 1,309 individuals (53.23%) were diagnosed with NAFLD. Sixteen variables, including BMI, waist circumference, systolic blood pressure, triglycerides, HDL-C, ALT, GGT, bilirubin fractions, albumin, BUN, GFR, fasting insulin, RBC, and hemoglobin, were selected as key predictors. The SVM model demonstrated the best overall performance, achieving an AUC of 0.920, accuracy of 0.839, and specificity of 0.898 in the training set, and an AUC of 0.833 and accuracy of 0.733 in the validation set. Decision curve analysis confirmed superior clinical utility of the SVM model compared with other algorithms. ML-based models, particularly the SVM algorithm, effectively predicted NAFLD among patients with T2DM using easily accessible clinical and biochemical indicators. These findings highlight the potential utility of ML-assisted screening tools for improving early identification and risk stratification of NAFLD in diabetic populations.

The role of intrauterine PRP infusion in managing recurrent implantation failure (RIF) remains controversial despite its emerging clinical use. This systematic review aims to evaluate its therapeutic potential in RIF patients and further to investigate variations in outcomes based on transfer cycle type, embryo developmental stage, RIF diagnostic criteria, and endometrial thickness. We systematically searched MEDLINE, Embase, the Cochrane Central Register of Controlled Trials, Scopus, and Web of Science for randomized controlled trials (RCTs) investigating PRP treatment for RIF patients from the beginning of the database to May 2025. This meta-analysis showed that PRP administration significantly improved clinical pregnancy rate (CPR) [OR&#xa0;=&#xa0;3.18, 95%CI (2.45, 4.14), I2 = 3%], biochemical pregnancy rate (BPR) [OR&#xa0;=&#xa0;2.84, 95%CI (2.22, 3.63), I2 = 0%], ongoing pregnancy rate (OPR) [OR&#xa0;=&#xa0;3.41, 95%CI (2.08, 5.60), I2 = 30%] and live birth rate (LBR) [OR=5.10, 95%CI (1.95, 13.37), I2 = 75%] in women with RIF. However, PRP intrauterine infusion did not reduce miscarriage rate (MR). Notably, the preterm birth rate was significantly higher in the PRP group compared to controls [OR&#xa0;=&#xa0;8.24, 95%CI (2.09, 32.41), I2 = 0%]. Subgroup analysis demonstrated that PRP improved CPR, BPR and LBR in both the fresh and frozen embryo transfer cycles. Additionally, while PRP increased CPR, LBR and reduced MR in blastocyst transfers [CPR OR&#xa0;=&#xa0;3.84, 95%CI (2.82, 5.23), I2 = 0%; LBR OR&#xa0;=&#xa0;7.32, 95%CI (3.17, 16.90), I2&#xa0;=&#xa0;63%; MR OR&#xa0;=&#xa0;0.27, 95%CI (0.07, 0.96), I2&#xa0;= 54%], these effects were not observed in cleavage-stage embryo transfers. Moreover, PRP administration associated with a higher CPR [OR&#xa0;=&#xa0;3.84, 95%CI (2.82, 5.23), I2&#xa0;=&#xa0;0%], OPR[OR&#xa0;=&#xa0;4.13, 95%CI (1.79, 9.56), I2&#xa0;=&#xa0;48%], LBR [OR&#xa0;=&#xa0;7.32, 95%CI (3.17, 16.90), I2&#xa0;=&#xa0;63%] and a lower MR [OR&#xa0;=&#xa0;0.27, 95%CI (0.07, 0.96), I2&#xa0;=&#xa0;54%] in women with &#x2265;3 prior implantation failure, it did not confer the same benefit to those with a history of &#x2265;2 failed cycles. These findings suggest a possible beneficial role for PRP on pregnancy outcomes to some extent in women with RIF, particularly in cases with &#x2265;3 prior failed transfers, and blastocyst transfer may increase LBR and reduce miscarriage risk. However, further investigation is warranted to determine whether this treatment may pose an increased risk of preterm birth. https://www.crd.york.ac.uk/prospero/, identifier CRD420251061511.