Frontiers in Endocrinology最新文献

Background: Diabetic kidney disease (DKD) is the leading cause of end-stage renal disease globally, yet early diagnosis remains challenging due to conventional biomarker limitations, including UACR variability and reduced eGFR sensitivity. While machine learning shows promise in diabetes prediction, its application to early DKD identification using routine parameters remains underexplored. This study aimed to develop and validate machine learning models incorporating routine blood and biochemical parameters for early DKD prediction.

Methods: This retrospective study analyzed 3,114 diabetic patients from the Second Affiliated Hospital of Wannan Medical College (EDN1) and 1,496 patients from NHANES 2005-2018 (EDN2) for external validation. Early DKD was defined as UACR 30-300 mg/g with eGFR ≥60 ml/min/1.73m². Seven machine learning algorithms were compared. Feature importance was assessed using SHAP framework, and Mendelian randomization explored causal relationships.

Results: Among 3,114 patients, 1,333 (42.8%) had early DKD. Logistic regression achieved optimal performance (AUC = 0.689, sensitivity=40.5%, specificity=81.3%). Top predictors included triglyceride-glucose index (TyG), gender, creatinine, globulin, and age. External validation confirmed significant associations for HbA1c, globulin, TyG, and neutrophil-to-albumin ratio.

Conclusions: The machine learning model successfully identified early DKD using routine parameters, with TyG index, HbA1c, and globulin as key predictors, demonstrating potential as a cost-effective screening tool.

Background: Congenital hypothyroidism (CH) is a preventable cause of growth and neurodevelopmental deficits. In India, lack of universal newborn screening (NBS) delays diagnosis. This study examines how treatment timing influences growth, nutrition, and body proportionality in children with CH.

Methods: This cross-sectional study included 66 children with CH, aged under 5 years, receiving Levothyroxine(LT4) therapy for one year. Age and sex-matched 64 healthy children were also enrolled. Children with CH were stratified into early-treated (<3months) and late-treated (>3months) groups. Anthropometric measurements were converted to Z-scores using WHO standards. Body proportionality was assessed using: height-arm span difference, upper segment-to-lower segment (US: LS) ratio, Manouvrier's Indice Skélique (MIS). MIS classifies skeletal proportions as brachyskelia (short legs, long trunk; score ≤84.9), mesatyskelia (intermediate; 85.0-89.9), and macroskelia (long legs, short trunk; ≥90).

Results: Children with CH had significantly lower height-for-age (HAZ -2.31 ± 1.32 vs. -0.41 ± 0.96), weight-for-age (WAZ -1.98 ± 1.41 vs. -0.36 ± 1.02), and head circumference-for-age (HCZ -1.52 ± 1.28 vs. -0.21 ± 0.89) Z-scores compared with healthy controls (all p < 0.001). Stunting (HAZ < -2 SD) was present in 53% of children, while underweight (WAZ < -2 SD) was observed in 36.4%. Late-treated children had a higher prevalence of severe stunting (67.7% vs. 11.4%), severe underweight (29.0% vs. 2.9%), and microcephaly (35.5% vs. 8.6%) compared with early-treated children (p < 0.05). Disproportionate stature was observed in 71.0% of late-treated children compared with 42.9% of early-treated children (p < 0.05). US: LS ratio and Manouvrier's Index further confirmed greater skeletal disproportionality in late-treated children compared with early-treated children (p < 0.05).

Conclusions: Delayed treatment in CH significantly impairs growth, increases skeletal disproportionality, and adversely affects neurocranial development. Early initiation of therapy-ideally within the first 3 months-is essential. These findings highlight the need for newborn screening in India.

Objective: To evaluate the overall and site-specific fracture risk in individuals with IBD through a meta-epidemiologic approach, synthesizing data from cohort studies and providing a comprehensive analysis of fracture risk at different anatomical sites.

Methods: Following PRISMA 2020 guidelines, we systematically searched PubMed, Embase, and the Cochrane Library (inception to April 2025) for cohort studies reporting fracture risk in IBD patients. Eligible studies provided relative risk (RR) and 95% confidence interval(CI) for all-cause or site-specific fractures. Two reviewers independently screened records, extracted data, and assessed study quality using the Newcastle-Ottawa Scale (NOS). Random-effects meta-analysis, sensitivity/subgroup analyses, and publication bias assessment (funnel plots, Egger's test) were performed.

Result: Eleven cohort studies (4 prospective, 7 retrospective) from multiple countries were included, involving 2,102 to 54,591 IBD patients. NOS scores ranged from 5 to 8, indicating moderate to high study quality. Pooled analysis showed a 13% increased risk of all-cause fractures in IBD patients (RR = 1.13, 95% CI: 1.03-1.24; I²=70.8%, p<0.001). Subgroup analysis revealed higher fracture risks in Crohn's disease (CD: RR = 1.23, 95% CI: 1.21-1.25) compared to ulcerative colitis (UC: RR = 1.16, 95% CI: 1.13-1.19). Site-specific risks were significantly higher for rib (RR = 1.24, 95% CI: 1.08-1.42; I²=0%, p=0.978), hip (RR = 1.39, 95% CI: 1.22-1.59; I²=54.2%, p=0.053), upper limb (RR = 1.46, 95% CI: 1.18-1.82; I²=94.6%, p<0.001), and lower limb fractures (RR = 1.60, 95% CI: 1.36-1.88; I²=75.4%, p<0.001). Sensitivity analyses confirmed the robustness of the results, and funnel plots/Egger's test indicated no significant publication bias (p=0.612).

Conclusion: IBD is associated with increased risks of all-cause and site-specific fractures, particularly in CD patients and lower limb fractures. These findings underscore the need for targeted bone health monitoring in IBD management.

Systematic review registration: PROSPERO, identifier: CRD420251038879.

Background: Remnant cholesterol (RC) has been recognized as a critical risk factor for vascular diseases. However, its association with cardiovascular disease (CVD) and overall mortality in individuals with diabetic kidney disease (DKD) has not been thoroughly investigated.

Methods: This retrospective single-center study enrolled 329 patients with biopsy-confirmed DKD from August 2009 to December 2018. The prognosis of patients with varying RC levels was compared. In addition, the association between RC levels and left ventricular structure and function was examined. Furthermore, the predictive capability of RC for clinical outcomes was assessed.

Results: Over the follow-up period, 76 patients (23.1%) experienced CVD events, while 44 patients (13.4%) died. Kaplan-Meier analysis demonstrated that patients in the high RC group exhibited significantly elevated rates of CVD events (p = 0.0047) and all-cause mortality (p = 0.0213). Additionally, multivariate Cox regression analysis confirmed RC as an independent risk factor for CVD events (HR = 1.323, 95% CI 1.076-1.626, p = 0.008) and overall mortality (HR = 1.359, 95% CI 1.039-1.779, p = 0.025). Finally, the AUC analysis indicated that the inclusion of RC in traditional risk factors improved their predictive accuracy.

Conclusion: RC independently contributes to the risk of CVD events and all-cause mortality in individuals with biopsy-confirmed DKD.

Diabetic neuropathy may present with a wide spectrum of sensory manifestations in the oral cavity, ranging from increased sensitivity to pain to impaired nociception. Early detection is crucial, since the changes have significant effects on dental treatment, patient safety and quality of life. We describe two contrasting pictures of patients with diabetes mellitus with signs of oral sensory neuropathy. The first case was a 14-year-old female with poorly controlled type 1 diabetes mellitus (T1DM) presenting with chronic carious lesions, gingivitis, and angular cheilitis. She also had amplified pain reactions to dental procedures despite adequate local anaesthesia and quantitative sensory testing revealed hyperaesthesia related to early diabetic neuropathy. The second patient was a 55-year-old male with a long-term history of type 2 diabetes mellitus (T2DM) with chronic periodontitis and reduced oral sensitivity. He complained of hypoesthesia and reduced pain perception during dental procedures with evidence of established neuropathic involvement. These cases illustrate the broad spectrum of clinical manifestations of diabetic sensory neuropathy in stomatology from early hyperaesthesia in a child with T1DM to late hypoesthesia in an adult with T2DM and illustrate the importance of metabolic control, interdisciplinary collaboration and individualised diagnostic and therapeutic strategies in dentistry.

Aims: Male hypogonadism is highly prevalent among patients undergoing dialysis and carries significant clinical implications, although is often overlooked. Testosterone deficiency in this population is associated with adverse clinical outcomes, mainly because of cardiovascular disease, protein energy wasting and infection complication. Therefore, identify these patients is a high clinical priority.

Methods: This is a cross-sectional study that enrolled adult men on maintenance hemodiafiltration in four dialysis centers. Hypogonadism was defined by two consecutive measurements of serum total testosterone levels. Free testosterone was calculated based on serum albumin and sex hormone-binding globulin. Symptoms of androgen deficiency were evaluated using androgen deficiency in aging males (ADAM) questionnaire. The presence of comorbidities and laboratory markers was also evaluated.

Results: Hypogonadism was identified in 59 out of 121 patients (48.7%). Patients with hypogonadism were older (62 ± 15 vs. 57 ± 15 years, p < 0.001) and had higher prolactin levels (22 [13-36] vs. 14 [10-18] ng/mL, p = 0.002). No other significant difference was observed in demographic, clinical, or laboratory features between patients with and without hypogonadism. Among the 16 patients who received testosterone supplementation, 66.7% showed improvement in ADAM scores, with the median score decreasing from 3 (2-4) to 1 (0-2) (p = 0.003).

Conclusion: More than one-third of men undergoing hemodiafiltration were diagnosed with hypogonadism. Aside from older age, no other distinguishing characteristics were identified in this population. Therefore, routine assessment of testosterone levels should be considered for all men undergoing dialysis. Further studies are needed to determine whether hormone supplementation can improve clinical outcomes.

Introduction: Growth hormone deficiency (GHD) in children and adolescents is a chronic condition requiring long-term therapy with recombinant human growth hormone (rhGH). Daily injections pose adherence challenges, prompting the development of long-acting GH (LAGH) formulations, such as once-weekly somatrogon. While phase III trials have demonstrated its efficacy, real-world data are limited.

Methods: This retrospective study evaluated all pediatric patients with GHD who initiated somatrogon between March 2023 and January 2025 at a tertiary endocrine center in Italy and completed at least 6 months of treatment.

Results: Forty patients (50% naïve; 50% switched from daily rhGH) were included. At 6 months, height SDS increased significantly in both naïve (Δ +0.19) and switch patients (Δ +0.17), with no significant difference between groups. However, by 18 and 24 months, naïve patients showed significantly greater height gains, with a median cumulative Δ of +0.81 at 18 months. IGF-1 SDS increased significantly only in the naïve group. Median gain in height SDS at 12 months in naïve patients (+0.37) was lower than reported in registration trials, likely reflecting the broader clinical heterogeneity of real-world populations. Treatment was well tolerated, with no discontinuations and few mild adverse events. Several families reported improved adherence and quality of life.

Conclusions: In this first real-world cohort, somatrogon was safe and effective in supporting linear growth, although height gains were lower than in clinical trials. Weekly administration may offer practical benefits, especially for patients with complex needs or poor adherence to daily injections.

Introduction: Despite the established link between elevated Low-density lipoprotein cholesterol (LDL-C) and diabetes, the association of LDL-C levels within the normal range with diabetes is poorly characterized. This study aimed to investigate the relationship between normal-range LDL-C levels and the incidence of new-onset diabetes.

Methods: In this retrospective cohort study, a total of 98,857 individuals with normal LDL-C levels (≤3.4 mmol/L) were enrolled between 2010 and 2016. The association was assessed using Cox proportional hazards regression models. Restricted cubic splines were utilized to examine dose-response relationships and potential nonlinearity. Subgroup analyses were pre-specified by age, sex, and Body mass index (BMI).

Results: During a median follow-up of 3.1 years, 2,107 incident diabetes cases were documented. After multivariable adjustment, participants in the highest level of LDL-C (G4: 2.6-3.4 mmol/L) had a significantly increased risk of diabetes compared to those in the lowest level (G1: 0-1.4 mmol/L), with an adjusted hazard ratios (HR) of 2.11 [95% confidence interval (CI): 1.32-3.38, P = 0.002]. A clear dose-response relationship was observed. Notably, the risk of incident diabetes increased significantly once LDL-C levels exceeded approximately 1.8 mmol/L (Adjusted HR = 1.93, 95% CI: 1.64-2.28, P<0.001). This association remained consistent across all pre-specified subgroups.

Discussion: Among individuals with normal LDL-C levels, higher LDL-C is an independent, dose-dependent risk factor for new-onset diabetes. Maintaining LDL-C below a threshold of approximately 1.8 mmol/L may be associated with a lower diabetes risk, suggesting its potential role in refining primary prevention strategies.

Context: Pretibial myxedema (PTM) is a refractory autoimmune dermopathy associated with Graves' disease. Although metabolic dysregulation has been recognized in thyroid-associated disorders, the metabolic profile and its functional role in PTM remain unclear.

Objective: To characterize the metabolic landscape of PTM lesions and explore the contribution of fatty acids to fibroblast dysfunction and inflammation.

Methods: We performed untargeted metabolomic profiling of PTM skin lesions and healthy controls using LC-MS and GC-MS, integrated with spatial metabolomics to localize metabolic changes. Functional assays were conducted by stimulating human foreskin fibroblasts (HFFs) with palmitic acid (PA) and oleic acid (OA), followed by RNA sequencing, cytokine assays, and immunohistochemistry.

Results: PTM lesions exhibited substantial metabolic dysregulation, including accumulation of fatty acids and elevated tricarboxylic acid cycle intermediates. Spatial metabolomics confirmed pronounced lipid deposition in the dermis, the primary site of PTM pathology. RNA-seq of fibroblasts stimulated with PA and OA revealed enrichment of inflammatory pathways, including IL-17 and NF-κB signaling, and marked upregulation of IL-8 (CXCL8). Fatty acid stimulation induced robust IL-8 secretion, consistent with increased IL-8 expression in PTM lesions. Moreover, PA promoted α-SMA expression in fibroblasts, suggesting induction of myofibroblast differentiation.

Conclusions: Our findings demonstrate that dermal fatty acid accumulation in PTM may contribute to fibroblast-mediated inflammation and fibrosis. This study provides novel insights into the metabolic-immunologic interface underlying PTM pathogenesis.

Background: Papillary thyroid carcinoma (PTC) is the most common endocrine malignancy, with excellent survival but substantial variation in recurrence risk. Traditional clinicopathologic risk models, while still a cornerstone of current guidelines, overlook the biological differences between patients, resulting in both overtreatment and undertreatment.

Content: Next-generation sequencing has advanced our molecular understanding of PTC by identifying recurrent driver mutations that shed light on tumor initiation. However, mutations fall short of explaining the full spectrum of clinical behavior. DNA-based mutation profiling offers a fixed snapshot of genetic alterations, while transcriptomics captures the tumor's active biological state, integrating signaling pathways, differentiation status, immune interactions, and metabolism. Large-scale efforts like The Cancer Genome Atlas, along with emerging transcriptomic classifiers, have shown that gene-expression subtypes ("BRAF-like" and "RAS-like") more accurately predict iodine avidity, tumor aggressiveness, and treatment response than histology or genotype alone. Transcriptome-based tools such as Thyroid GuidePx^® now allow for biologically informed risk stratification that goes beyond traditional clinicopathologic and mutation-only approaches.

Summary and outlook: In the preoperative setting, transcriptomic testing can inform whether patients are best suited for active surveillance, lobectomy, or total thyroidectomy. Postoperatively, it sharpens decisions around completion surgery, radioactive iodine use, and the intensity of TSH suppression. Integrating transcriptomic data into clinical decision-making enables more precise selection for treatment escalation or de-escalation. To unlock the full potential of transcriptome-guided management in PTC, prospective validation and adoption into ATA and NCCN guidelines will be critical.