Search Results (858)

Background/Objectives: Diabetic retinopathy (DR) is a leading cause of vision loss and is triggered by chronic inflammation and genetic predisposition. The aim of this study was to examine the connection between specific gene variations of interleukin IL-18 and IL-8 and the likelihood of developing DR, their part in the clinical severity of the condition, and their possible effect on kidney function in patients with diabetes. Methods: This study included 176 participants, 88 of whom were patients with DR and 88 of whom were healthy controls. Genotyping for IL-18 (−607 C/A, −137 G/C) was performed using allele-specific PCR, while IL-8 (−251 A/T) analysis was conducted via the PCR-restriction fragment length polymorphism (PCR-RFLP) method. A comprehensive documentation of demographic and clinical parameters was undertaken. Renal function was the subject of evaluation using the estimated glomerular filtration rate (eGFR). The statistical analyses included the independent samples t-test for continuous variables, the chi-squared test for categorical data, and binary logistic regression to assess genotype distributions. Adjusted multivariate logistic regression analysis was performed to account for potential confounding factors, and significance was determined using the Bonferroni correction (adjusted α = 0.005) for all genetic and clinical risk assessments. Results: A multivariate logistic regression analysis, adjusted for clinical variables, identified the IL-18 (−137) GC genotype (adjusted odds ratio (AOR) = 2.25, p = 0.002) and the IL-8 (−251) AT genotype (AOR = 1.98, p = 0.003) as significant independent risk factors for DR. Combined genotype analysis revealed that the IL-18/IL-8 (CA-AT) and (GC-AT) combinations were the most potent risk factors, with an adjusted odds ratio (AOR) of 3.10 (p = 0.002) for each. Additionally, specific IL-18 (CC-GC and CC-GG) haplotype combinations were identified as significant predictors of DR risk (AOR = 2.15, p = 0.004 and AOR = 0.38, p = 0.003, respectively). Exploratory subgroup analyses within the DR cohort revealed that certain genotypes, notably IL-18 (−607) CA, IL-18 (−137) GC, and IL-8 (−251) AT, remained significantly associated with the presence of vision-threatening diabetic retinopathy (VTDR) and ME following Bonferroni correction (p < 0.005). Finally, no statistically significant differences in eGFR levels were observed across the various genotype distributions (p > 0.05). Conclusions: Both IL-18 (−137 G/C) and IL-8 (−251 A/T) gene variations, and their synergistic combinations (e.g., CA-AT and GC-AT), significantly contribute to the risk and clinical severity of DR. Our adjusted multivariate analysis confirms that these variations are independent risk factors for DR, with high odds ratios observed in advanced clinical stages, such as VTDR and ME. Additionally, while hypertension, a family history of diabetes mellitus, and CAD were identified as significant clinical predictors, the lack of significant variance in eGFR across genotype distributions suggests that these genetic variations may influence DR pathology independently of gross renal impairment in our study population. In light of the observed associations, further investigation of these genetic markers in larger prospective cohorts is warranted to clarify the underlying inflammatory mechanisms and their broader clinical implications. Full article

Objective: To predict pixel-wise retinal thickness maps from near-infrared (NIR) fundus images using deep learning (DL), and to identify image features in NIR fundus photographs serving as surrogate markers of retinal thickness, with implications for decentralized diabetic macular edema (DME) screening, progression monitoring, and treatment assessment. Methods: A DL model based on a U-Net architecture was trained on paired NIR fundus and OCT images from 531 eyes across three groups: healthy controls, diabetic retinopathy (DR) without DME, and DME. Model performance was evaluated using mean absolute error (MAE), root mean squared error (RMSE), structural similarity index (SSIM), and center-involved DME (ci-DME) classification at a central subfield thickness threshold of 300 µm. Controlled image manipulation experiments, including spatial disruption of vascular patterns, relocation of hard exudates, and contrast enhancement, were performed to identify image-level features serving as surrogate markers of retinal thickness. Results: The model achieved an MAE of 30.41 ± 18.68 µm, RMSE of 36.14 ± 21.05 µm, and SSIM of 0.87 ± 0.04 across the macula, with consistent performance across ETDRS subfields. For ci-DME classification, it achieved an accuracy of 84.1%, sensitivity of 69.1%, and specificity of 88.7%. Interpretability analyses were performed as qualitative assessments to visualize image regions contributing to model predictions. These analyses highlighted retinal vascular structures, hard exudates, and local contrast variations as visual features observed in relation to model outputs. Conclusions: NIR fundus images contain sufficient structural information to support pixel-wise retinal thickness estimation, with vascular architecture, hard exudates, and local contrast variations identified as image features potentially associated with model predictions. These findings suggest that NIR-based deep learning approaches may have potential applications in the assessment of diabetic macular edema and warrant further prospective and external validation to determine their role in screening, triage support, longitudinal monitoring, and treatment-related assessment, particularly in decentralized and re-source-limited care environments. Full article

Early diagnosis of Diabetic Retinopathy (DR) is critical for preventing irreversible vision loss, but precise lesion annotation by ophthalmologists is the dominant cost in building any clinical-grade DR detection model. The structural problem in real hospital settings is not labeling cost per se, but expert availability: ophthalmologists’ time is bounded by clinical duties, so the active-learning (AL) cycle can iterate only a handful of times in practice. We frame this constraint explicitly and ask which AL designs work best under a tight expert budget. We propose Virtuous Cycle, a Human-in-the-Loop (HITL) pipeline that integrates (i) a YOLOv8x-based object detector for microaneurysms, hemorrhages, and exudates, (ii) four AL sampling strategies (Average Confidence, Random, Hybrid-Diversity, Monte Carlo Dropout), and (iii) an in-hospital annotation platform (Diavision Studio) in which clinicians refine AI pre-labels rather than draw from scratch. We evaluate Virtuous Cycle on a real-world fundus dataset from the National Medical Center (NMC) across eight AL rounds, expanding the labeled pool from 81 images (R0) to 481 images (R8) within the actual expert-time budget of two ophthalmologists. Across three independent random seeds, random sampling dominates at cold start (mean mAP@50 $0.14\to 0.25$ over R0–R1), whereas Hybrid-Diversity converges to the highest mAP@50, Precision, and Recall by R7 (431 images; mAP@50 $0.40$, Precision $0.55$, Recall $0.41$), with MC Dropout close behind; by R8, the labeled pool is exhausted and all strategies converge to the same final model. A clinician crossover analysis of 36 paired clinical images, controlling for per-clinician speed bias and per-image difficulty bias, shows no statistically significant difference in overall per-image labeling time between AI-assisted and manual annotation ($p=0.52$), but a statistically significant increase in confirmed lesion detections under AI assistance ($p=0.0058$), driven predominantly (84–$100\%$ of the net increase) by microaneurysms, the lesion type most prone to being missed unaided. The results indicate that, under expert-budget constraints, AL strategy choice should be staged: random sampling for cold start, uncertainty-and-diversity sampling once the model has matured, and that AI assistance trades a modest, lesion-burden-dependent time cost for a measurable gain in the sensitivity of microaneurysm detection. Full article

Diabetic retinopathy (DR) is a leading cause of preventable blindness among working-age adults. Systematic screening programmes in the United Kingdom, Ireland, and the Nordic countries have reduced diabetes-related visual loss, yet many European countries, including Croatia, lack organised screening. This narrative review examines European DR screening programmes, evaluates telemedicine and artificial intelligence (AI) as enabling technologies, and proposes a phased roadmap for a national programme in Croatia. Croatia has nearly 400,000 registered persons with diabetes, a national diabetes registry (CroDiab) linked to the Central Health Information System (CEZIH), and pilot screening data showing 40% DR prevalence among screened patients with type 2 diabetes. The proposed programme combines decentralised telemedicine-based imaging at primary care sites with centralised grading, a staged rollout across three phases targeting 400,000 annual screenings, and stepwise AI integration for triage. Successful European programmes share standardised digital imaging, trained grading workforces, embedded quality assurance, and registry linkage. With dedicated funding and sustained political commitment, Croatia could adopt a hybrid telemedicine–AI model and, under favourable implementation conditions, reduce the burden of advanced DR over the coming decade. Full article

Background/Objectives: Diabetic retinopathy (DR) is a leading cause of preventable visual impairment, and systematic screening is essential to detect sight-threatening stages before symptoms occur. However, long-term real-world evidence on visual outcomes from structured screening programmes remains limited. This study evaluates, in a retrospective analysis of routinely collected clinical data, the real-life visual outcomes of screening for and treating sight-threatening DR, based on 32 years of data collected in a purpose-built centre implementing the 1990 European Working Party recommendations for screening. Methods: Screening was performed by retinal photography between 1991 and 2022 in 18,161 patients (63,289 screening episodes). Diabetes specialists graded photographs, referring patients to ophthalmologists when needed. The 10-year trajectories of visual acuity (VA) were assessed in patients with different stages of retinopathy and macular involvement at first screening. Results: At first screening, two-thirds of patients had no DR, 15% had mild DR, and the remainder had referable DR or a non-assessable fundus. There was no difference by sex. Patients with more severe DR at first screening had lower VA, but this did not worsen over 10 years. Median VA in 514 patients treated with panretinal photocoagulation for pre-proliferative or proliferative DR changed from 0.10 logMAR (7–8/10) to 0.15 (6–8/10), 1874 ± 1252 days after treatment. In 823 patients photocoagulated for diabetic macular edema, median VA remained 0.10 logMAR (7–8/10) before treatment and 1998 ± 1288 days after treatment. Conclusions: Screening for sight-threatening DR using European Working Party recommendations was feasible in everyday practice and was associated with long-term preservation of visual acuity and a low incidence of severe visual loss. Full article

Retinal diseases such as age-related macular degeneration (AMD) and diabetic retinopathy (DR) are major causes of visual impairment and are closely associated with oxidative stress, inflammation, vascular dysfunction, and metabolic imbalance. Increasing evidence suggests that gut microbiota also contributes to retinal homeostasis, supporting the emerging concept of the gut–eye axis. In this context, dietary anthocyanins—with blueberry anthocyanins serving as a primary representative model—have attracted attention as potential adjunctive nutritional modulators of ocular health. However, their biological effects are strongly influenced by their limited bioavailability and extensive gastrointestinal metabolism. The objective of this review is to analyze the gastrointestinal fate of dietary anthocyanins and to discuss how their absorption, enzymatic transformation, and microbial biotransformation may influence ocular protection through the gut–eye axis. The review summarizes current knowledge regarding anthocyanin stability in the oral cavity, stomach, small intestine, and colon, as well as the formation of circulating phenolic metabolites generated by the host and through microbial metabolism. In addition, the molecular mechanisms through which anthocyanins and their metabolites may support retinal health are examined, including antioxidant, anti-inflammatory, vasoprotective, and neuroprotective actions. Overall, dietary anthocyanins, illustrated through the rich profile of blueberries, represent promising adjunctive compounds for supporting ocular health, although further clinical and mechanistic studies are still required. Full article

Background/Objectives: Diabetic retinopathy (DR) is a major microvascular complication of type 2 diabetes (T2D) and a leading cause of visual impairment. The relationships among Mediterranean diet adherence, dietary components, DR, and vision-related quality of life remain incompletely defined. This cross-sectional study evaluated Mediterranean Diet Score (MDS) as the primary dietary endpoint, individual MDS components as secondary endpoints, and micronutrient intakes as exploratory endpoints. Methods: In this single-centre study, 129 subjects with long-standing T2D were classified as no DR (NDR; n = 85), non-proliferative DR (NPDR; n = 36), or proliferative DR (PDR; n = 8). Dietary intake was assessed using a food frequency questionnaire and vision-related quality of life using the NEI-VFQ-25. Results: Subjects with DR had longer diabetes duration than those without DR (18 vs. 16 years, p < 0.01). Overall MDS did not differ by DR status, indicating a null finding for the primary dietary endpoint. In secondary analyses, lower legume consumption was observed among participants with DR and was associated with higher odds of DR in multivariable models. Participants with PDR showed poorer vision-related quality of life, although this finding was limited by the small PDR subgroup and high NEI-VFQ-25 scores in other groups. Exploratory analyses suggested associations between selected micronutrient intakes and NEI-VFQ-25 domains. Conclusions: Overall Mediterranean diet adherence was not associated with DR status. Secondary and exploratory findings should be considered hypothesis-generating and require confirmation in prospective studies. Full article

Background: Diabetic retinopathy (DR) is one of the major complications of diabetes mellitus. EJPs (Erjing Pills) are believed in Traditional Chinese Medicine to have the effects of a nourishing essence and a brightening of the eyes, but the specific effect on DR remains unclear. This study aims to investigate the therapeutic effects and underlying mechanisms of EJPs on DR. Methods: The chemical profile of EJPs was characterized by UHPLC-MS. Network pharmacology and molecular docking were employed to predict its active ingredients and potential targets. A DR rat model was induced by streptozotocin. Retinal morphology and function were assessed by OCT, FFA, and H&E staining. The expressions of proteins and mRNAs related to the AGE-RAGE pathway, oxidative stress, inflammation, and tight junctions were detected by Western blot, qPCR, and ELISA. Results: LC-MS and network pharmacology analysis identified 638 common targets between EJPs and DR, with core targets including SRC, AKT1, and MAPK1, primarily enriched in the AGE-RAGE signaling pathway. Molecular docking confirmed strong binding (binding energy < −5.0 kcal/mol) between key EJP constituents and core targets. In vivo, EJP treatment significantly alleviated retinal vascular leakage, improved retinal thickness, and alleviated histopathological damage. In addition, EJPs downregulated the AGEs-RAGE/NF-κB axis and pro-inflammatory cytokines while enhancing antioxidant defenses and tight junction proteins in the retinas of DR rats. Conclusions: EJPs ameliorate DR by protecting the blood–retinal barrier and modulating the AGE-RAGE/oxidative stress/inflammation network, demonstrating a multi-component, multi-target, and multi-pathway mechanism. This study provides a mechanistic basis for the potential application of EJPs in DR management. Full article

Background and Objectives: Type 2 diabetes mellitus (T2DM) causes retinal microvascular changes that precede clinically apparent diabetic retinopathy (DR). We aimed to identify which optical coherence tomography angiography (OCTA) biomarkers best distinguish eyes with T2DM without clinical DR from healthy controls and to evaluate machine learning classifiers trained on a comprehensive 68-parameter OCTA panel. Materials and Methods: In this prospective case–control study, 80 patients with T2DM without clinical DR and 33 controls underwent 3 × 3 mm macular OCTA using an Optovue RTVue Avanti System. After outlier screening, 221 eyes (155 T2DM, 66 controls) were analyzed. Sixty-eight OCTA parameters were extracted, covering FAZ morphometry (including foveal density FD-300), SCP and DCP vessel density and layer thickness, outer-retina and choriocapillaris flow, and a full retinal-thickness map. Between-group comparisons used the Mann–Whitney U test with Benjamini–Hochberg FDR correction. Logistic regression, random forest, and XGBoost classifiers were evaluated with patient-grouped 10-fold cross-validation; feature importance was quantified via SHAP. Results: Forty-two of 68 parameters reached FDR significance (q < 0.05). Deep capillary plexus vessel density was the most discriminative family (whole image rb = −0.66, q = 2.5 × 10⁻¹³; parafovea rb = −0.64). FD-300 was reduced in T2DM (median 47.55% vs. 51.86%; rb = −0.57; q = 1.0 × 10⁻¹⁰) and emerged as the top SHAP feature (mean |SHAP| = 0.81). FAZ circularity decreased without FAZ-area enlargement, and outer-retina flow was paradoxically elevated (rb = +0.39), consistent with a projection artifact. XGBoost using all 68 features achieved a patient-grouped cross-validated AUC of approximately 0.91, compared with 0.85 for conventional SCP + DCP whole-image density. No parameter correlated with current HbA1c in T2DM (all q > 0.98), and the well-controlled (<7%) and poorly controlled (≥7%) subgroups were indistinguishable across five of six principal biomarkers, consistent with metabolic memory. FD-300 remained independent after adjustment for hypertension, hyperlipidemia, and age (OR = 0.76; 95% CI 0.69–0.84; p < 0.001). Conclusions: A multi-compartment OCTA panel outperforms conventional two-layer vessel-density metrics in detecting preclinical diabetic microvasculopathy, although external validation is required before clinical use. FD-300 is the single most informative biomarker, while choriocapillaris and retinal thickness measures provide complementary, compartment-specific signals. Because the OCTA signature is decoupled from the current HbA1c, screening should not be deferred in well-controlled T2DM. Full article

This study proposes OrdPrune-KD, an ordinal-consistency-driven model compression framework that integrates grade-aware structured pruning with Earth Mover’s Distance (EMD)-based knowledge distillation for diabetic retinopathy (DR) grading. Unlike conventional approaches that only consider ordinal relationships at the loss level, the proposed method incorporates ordinal priors into both model compression and knowledge transfer stages. Extensive experiments on APTOS 2019, Messidor-2, and IDRiD demonstrate that the proposed framework achieves a favorable balance between model compactness and predictive performance. In particular, under a 77% parameter reduction, the student model achieves competitive performance relative to the teacher model in terms of QWK while maintaining strong high-risk sensitivity. Additional ablation studies and fairness-controlled comparisons confirm that the performance gains are primarily attributed to the proposed ordinal-aware design rather than output formulation differences. These results indicate that OrdPrune-KD provides an effective and deployable solution for lightweight DR grading systems. Full article

Research on diabetic retinopathy (DR) usually emphasizes hyperglycemia and other causes like dyslipidemia, which are still not well understood. This study examined the effects of palmitic acid (PA) exposure, alone and combined with high glucose (G25), on Müller Glial Cell (MGC) dysfunction and angiogenic signaling. Primary MGC cultures were treated with G25 (25 mM), PA (250 µM), or PA + G25 for 24 and 48 h, followed by assessments of cell viability and analysis of the Vascular Endothelial Growth Factor (VEGFA)/VEGFA receptor 2 (VEGFR2) pathway through immunofluorescence, Western blot, and ELISA. Additionally, Gaussian mixture models (GMMs) were used to identify phenotypic subpopulations based on fluorescence intensity. The results showed that while hyperglycemia did not cause significant changes, PA and PA + G25 induced apoptosis-related cell death and significantly increased the expression of VEGFA, VEGFR2, HIF-α, and SP1. Although broad phenotypic diversity was observed at 24 h, by 48 h, a distinct shift towards an angiogenic phenotype was noted, with significantly elevated VEGFA/VEGFR2 levels. In summary, this research demonstrates that PA acts as a critical inducer of an angiogenic secretory phenotype in MGCs, indicating that lipid-mediated signaling plays a vital role in neovascularization in DR, possibly independent of glucose levels. Full article

Optical coherence tomography angiography (OCTA) is a depth-resolved, label-free optical imaging modality that uses motion contrast from repeated B-scans to reconstruct retinal microvasculature and provide co-registered en face views of the superficial and deep vascular plexuses (SVP and DVP). This capability is valuable for early diabetic retinopathy (DR) assessment, where deep-plexus perfusion deficits may precede clinically evident disease. However, microvascular differences among healthy controls, diabetic eyes without clinically apparent retinopathy, and mild DR are subtle and unevenly distributed across the two vascular slabs, while most deep learning methods prematurely fuse the plexuses and weaken depth-specific evidence provided by OCTA. To address this, we propose Class-Path Specific Representation Distillation and Reasoning (CPS-RDR), an interpretable framework that aligns model reasoning with the layered organization of OCTA. A frozen DINOv2-initialized dual-branch Vision Transformer preserves separate SVP and DVP representations, while class- and path-conditioned diagnostic queries instantiate four reasoning pathways for layer-specific evidence extraction and directional cross-plexus interaction. A lightweight EvidenceFusion head linearly integrates pathway-wise evidence, enabling final predictions to be decomposed into pathway-specific contributions. On 99 eyes from 55 participants, CPS-RDR achieved 97.29% accuracy, 0.9932 macro-AUC, and 0.9829 macro-F1 under five-fold patient-level cross-validation, outperforming seven representative baselines, while producing path-resolved maps that reveal how superficial- and deep-layer optical signals jointly support early DR stratification. Full article

Diabetic retinopathy is a leading cause of vision loss worldwide, particularly impacting individuals in low- and middle-income countries with limited healthcare access. Early detection through automated screening systems is essential for improving outcomes, as timely intervention can prevent severe vision impairment. However, most of the available AI models have not been evaluated in low-resource settings. Hence, this study presents an evaluation of the efficacy of advanced deep learning architectures for detecting rDR across diverse population datasets. A dual-phase validation approach was employed to assess model performance. Internal validation utilised the BrSET dataset to establish baseline performance metrics, while external validation was conducted on the MoDRIA dataset, which encompasses various conditions and demographics, to evaluate model robustness. Key performance metrics, including accuracy, specificity, sensitivity, F1-score, and calibration scores, were systematically recorded and analysed. Internal validation revealed high accuracy across all models, EfficientNetB0 achieved the highest classification accuracy (0.9561; 95% CI 0.9490–0.9630), EfficientNetB3 demonstrated superior overall discriminative performance, achieving the highest AUROC (0.9892; 95% CI 0.9841–0.9934) highest sensitivity (0.9573), and lowest Brier score (0.0168). Meanwhile, DenseNet exhibited the most balanced clinical screening performance, achieving the highest F1-score (0.7259; 95% CI 0.6797–0.7669) and Youden Index (0.2381), indicating improved balance between sensitivity and specificity. In contrast, external validation revealed substantial deterioration in model performance across all architectures, highlighting major limitations in cross-population generalisability. Although EfficientNetB0 achieved the highest external accuracy (0.8821; 95% CI 0.8746–0.8898), AUROC values declined markedly across models (0.5140–0.6104), accompanied by poor sensitivity, reduced F1-scores, and substantial calibration instability. EfficientNetB3 achieved the highest external sensitivity (0.5939), whereas calibration analyses demonstrated unreliable probability estimation under domain-shift conditions. These findings suggest that AI models trained on geographically homogeneous retinal imaging datasets may not generalise reliably across underrepresented populations. Population differences and imaging variability substantially affected external model performance, highlighting the need for diverse datasets, rigorous external validation, and adaptive recalibration before clinical deployment of AI-driven DR screening systems. Full article

Background/Objectives: Diabetic retinopathy (DR) is one of the most common microvascular complications in type 2 diabetes mellitus (T2DM), in which oxidative stress, inflammation and angiogenic pathways are associated with the development and progression beyond glycemic control. Serum trace element levels (Cu, Zn, Fe, Mg, Cr, Mn, Cd, and Se), antioxidant enzyme activities (superoxide dismutase (SOD) and glutathione peroxidase (GSH-Px)) were measured in patients with T2DM, with and without DR, as well as in healthy controls, and their associations with the presence and severity of DR were evaluated. Methods: 61 T2DM patients, 27 healthy controls. Patients with T2DM were classified into T2DM without DR (n = 30) and T2DM with DR (n = 31). Non-proliferative DR (NPDR, n = 19) and proliferative DR (PDR, n = 12) were classified as the T2DM with DR group. Inductively coupled plasma–mass spectrometry (ICP-MS) was used to quantify serum trace elements. SOD and GSH-Px activities were measured using colorimetric assays. Results: Significant differences were observed in trace element levels and antioxidant enzyme activities among the study groups (p < 0.001 to 0.05). The DR subgroup had lower levels of Cr, Cu and Se compared to the T2DM without DR group; Cd, Zn and Mn were also higher in the T2DM with DR than in the T2DM without DR group. Fe levels were significantly higher in the PDR subgroup than in the T2DM without DR group (p < 0.001). The PDR group showed greater declines of Cr, Cu and GSH-Px compared to NPDR while higher values for Mn, Fe, and Zn were obtained (p < 0.001). Several biomarkers remained significantly associated with DR after adjustment for metabolic variables. Correlation analysis between trace elements, and antioxidant enzymes showed significant associations. Conclusions: Trace element imbalance, and reduced antioxidant enzyme activities may contribute to the development and progression of DR in T2DM. These findings suggest that oxidative stress and micronutrient imbalance may be linked to DR-related biochemical alterations. Full article

Background/Objectives: Observational studies have reported comorbidity between diabetic retinopathy (DR) and physical frailty, but their genetic interplay remains incompletely understood. This study evaluated shared genetic architecture and potential causal relationships between DR severity and frailty-related phenotypes (FRPs). Methods: GWAS summary statistics were analyzed for four DR phenotypes (broad DR, background DR [BDR], severe non-proliferative DR, and proliferative DR [PDR]) and six FRPs, including frailty index (FI), appendicular lean mass, handgrip strength (HGS), and walking pace (UWP). Global and local genetic correlations were estimated using LDSC, HDL, and LAVA. Causality was assessed using bidirectional Mendelian randomization (MR) and latent causal variable (LCV) analyses. Biological mechanisms were investigated using partitioned heritability, cross-trait meta-analysis, Bayesian colocalization, tissue and cell enrichment, prioritization (MAGMA/TWAS), and 3D chromatin annotation. Results: BDR and PDR showed positive genetic correlations with FI and negative correlations with UWP. Local genetic correlation analyses identified 82 significant regions, including signals on chromosome 6. MR supported a directional effect in which genetic liability to DR was associated with higher FI and lower HGS, with no evidence of reverse causation. LCV indicated partial genetic causality within a shared polygenic architecture. Cross-trait meta-analysis and colocalization highlighted the MHC region, prioritizing C2, AIF1, NOTCH4, and EHMT2. Additional non-MHC loci included the BCL2L15 gene cluster and TERF1. Conclusions: DR and frailty share genetic determinants involving neurovascular, metabolic, and immune-inflammatory pathways, supporting an association between DR liability and frailty-related decline. Future longitudinal and functional studies are needed to validate these findings and assess candidate pleiotropic genes. Full article