Diabetologia

Background Glycated haemoglobin (HbA1c) underpins type 2 diabetes (T2D) and prediabetes management worldwide and reflects both glycaemia and erythrocyte biology. A missense variant in PIEZO1 (rs563555492T), carried by 1 in 12 South Asians, has been associated with a nonglycaemic reduction in HbA1c. We aimed to further characterise this association and evaluate its clinical consequences. Methods We undertook genetic and linked health data analyses across two cohorts: 19,898 (37.4% female) South Indians from the Madras Diabetes Research Foundation (MDRF) and 43,011 (54.4% female) British Bangladeshis and British Pakistanis in Genes & Health. In MDRF, we tested associations with glycaemic and erythrocytic traits using additive genetic models. In Genes & Health we modelled diagnosis of prediabetes, T2D, and diabetic eye disease using flexible parametric survival models. Ten-year absolute risks were estimated for a population aged 40-50 years. Findings PIEZO1 rs563555492T was associated with erythrocytic traits and lower HbA1c, but not with fasting glucose, postprandial glucose, or C-peptide. This variant reduced risk of prediabetes (HR 0.63, 95% CI 0.58-0.69) and T2D (0.85, 0.78-0.93) diagnosis, and increased risk of diabetic eye disease among individuals with T2D (1.20, 1.01-1.43). Modelling suggested approximately 1,019 missed prediabetes and 303 missed T2D diagnoses per 100,000 adults over 10 years. Interpretation An ancestry-enriched PIEZO1 variant is associated with lower HbA1c independent of glycaemia, reduced prediabetes and T2D diagnosis suggesting delayed detection, and increased complication risk. Reliance on HbA1c may systematically underestimate glycaemic risk in a substantial minority of South Asians. Funding The Wellcome Trust; NIHR

Background. Type 2 diabetes mellitus (T2D) is defined by progressive pancreatic {beta}-cell dysfunction whose molecular underpinnings remain incompletely understood. Single-cohort transcriptomic analyses of donor islets have yielded heterogeneous gene lists of limited cross-study reproducibility, constraining both mechanistic interpretation and biomarker development. Methods. We combined two complementary analytical strategies applied to four public human islet transcriptomic cohorts (GSE25724, GSE20966, GSE38642, and GSE164416; n = 7-57 donors per contrast). For the integrative arm, three microarray datasets and one bulk RNA-seq dataset were processed independently and unified through gene-level random-effects meta-analysis, hallmark pathway scoring (GSVA/MSigDB), and iterative module refinement, yielding a two-axis disease framework. For the diagnostic arm, a consensus multi-method machine learning pipeline, combining LASSO penalized logistic regression, Support Vector Machine Recursive Feature Elimination (SVM-RFE), and Random Forest importance scoring, was applied to 184 differentially expressed genes from the RNA-seq cohort, with all normalization steps performed within leave-one-out cross-validation (LOOCV) folds to prevent data leakage. Machine learning classification of the RNA-seq cohort was additionally subjected to external transportability testing in the independent bulk human islet RNA-seq cohort GSE50244 using an overlap-restricted reduced score and a threshold fixed in the discovery cohort. Results. Meta-analysis across all four cohorts identified 337 high-confidence T2D-associated genes (96.1% directional concordance in beta-cell-enriched tissue). These were distilled into two refined 14-gene modules: ImmuneStress (MICB, HLA-DRA, HLA-DPA1, IL1R2, and others) and BetaCellIdentitySecretion (RASGRP1, PPP1R1A, SLC2A2, and others), whose composite IsletDysfunctionScore provided the most stable cross-platform separation of non-diabetic from T2D islets (Hedges' g = 1.80, p = 9.83 x $10^-17$, $\text{I}^2$= 0%). Consistent with progressive disease, IsletDysfunctionScore increased monotonically from non-diabetic to impaired glucose tolerance to T2D. Separately, the machine learning pipeline derived a 10-gene diagnostic panel: GABRA2, SLC2A2, ARG2, DKK3, PRIMA1, TAFA4, HHATL, PARVG, RNU1-70P, and the novel lncRNA ENSG00000284653, that achieved perfect discrimination in LOOCV (AUC = 1.000, sensitivity = 1.000, specificity = 1.000, zero misclassifications across all 57 donors). A leakage-verification experiment confirmed that this performance reflected genuine biological signal: global quantile normalization prior to cross-validation collapsed AUC to 0.380. External testing showed that 8 of the 10 panel genes were measurable in GSE50244. The frozen 8-gene reduced score retained strong discrimination (external AUC = 0.907), with 6 of 8 genes preserving directional concordance, but the discovery-derived threshold did not transfer because the external score distribution was shifted upward and compressed, yielding complete sensitivity but zero specificity at the frozen cutoff Conclusions. Integrating pathway-level meta-analysis with machine learning classification, we present a coherent two-axis model: immune/stress activation and loss of beta-cell identity/secretory competence, together with a compact, biologically interpretable 10-gene diagnostic signature. Panel genes converge on GABA signaling, glucose transport, arginine metabolism, WNT pathway inhibition, and a novel lncRNA, providing both mechanistic hypotheses and high-priority targets for external validation. These findings offer a reproducible transcriptomic scaffold for future mechanistic, biomarker, and clinical translation studies of human islet dysfunction. They also support external transportability of the core biological signal, while indicating that absolute operating thresholds are cohort-dependent and would require recalibration before deployment in independent datasets.

ObjectivesIt is not known which clinical features optimally differentiate type 1 and 2 diabetes at diagnosis. We aimed to determine which clinical features differentiate adult-onset type 1 and 2 diabetes at diagnosis and develop classification models combining these features with and without islet-autoantibodies. DesignA prospective cohort study with prediction model development and validation. SettingUK primary and secondary care. Participants1800 adults ([≥]18 years) diagnosed with diabetes in the previous 12 months, excluding known secondary or monogenic diabetes. Main outcome measuresType 1 and 2 diabetes defined by a combination of insulin treatment and endogenous insulin production (measured using C-peptide) assessed [≥]three years after diabetes diagnosis. ResultsEleven clinical features and routinely measured biomarkers discriminated type 1 from type 2 diabetes independently of diagnosis age and BMI. Lower age-at-diagnosis, BMI and waist-hip ratio, unintentional weight-loss, and higher presentation HbA1c or glucose were the most discriminative features, with other features only weakly discriminative. Models integrating clinical features with and without islet-autoantibodies, developed in those age 18-50 years at diabetes diagnosis, had high performance in internal validation (clinical features only: AUCROC (95% CI) 0.94 (0.93, 0.96), clinical features and islet-autoantibodies: AUCROC 0.97 (0.96, 0.98)), and maintained high discrimination in older adults (age >50 at diagnosis; clinical features only: AUCROC 0.93 (0.90, 0.96), clinical features and islet-autoantibodies: AUCROC 0.97 (0.94, 0.99)). Simplifying the models to a point-based score (the StartRight Score) resulted in similar performance. These models had higher performance than current clinical guidance. In UK primary care data models were strongly predictive of outcomes associated with type 1 diabetes, including in those initially treated as type 2 diabetes. ConclusionsLower age-at-diagnosis, BMI, and wait-hip ratio, unintentional weight-loss and high presentation glycaemia are the most discriminative features for diagnosis of type 1 diabetes in adults. Models combining routine clinical features, with or without islet-autoantibodies, have high accuracy and could assist clinical classification and prioritisation of classification biomarker testing. Study registrationhttps://clinicaltrials.gov/study/NCT03737799 Summary boxesO_ST_ABSSection 1: What is already known on this topicC_ST_ABSO_LIMost type 1 diabetes occurs in adults, but differentiating it from type 2 diabetes, which is much more common, is challenging, and misclassification is common. C_LIO_LIAge-at-diagnosis and BMI are currently the only clinical features robustly shown to distinguish between type 1 and type 2 diabetes at diagnosis; many other features included in textbooks and guidelines have little supporting evidence. C_LIO_LIGuideline bodies, including the UK National Institute for Health and Care Excellence (NICE), have identified a need for evidence on what features discriminate type 1 and 2 diabetes in adults, and how these features can be combined to improve diagnosis. C_LI Section 2: What this study addsO_LIThis is the first study to prospectively assess the utility of clinical features for diabetes subtype at diagnosis. C_LIO_LIThe five most discriminative routine clinical features for distinguishing type 1 from type 2 diabetes at diagnosis are age-at-diagnosis, BMI, waist-hip ratio, pre-diagnosis unintentional weight-loss, and presentation glycaemia (HbA1c or glucose). C_LIO_LIMany features included in current guidelines were only very weakly discriminative of subtype, and no single clinical feature was able to adequately differentiate between type 1 and type 2 diabetes alone. C_LIO_LIA clinical prediction model combining ten routinely available clinical features, with or without islet-autoantibodies, as both a prototype calculator and a points-based score (the StartRight Score), had high accuracy in differentiating type 1 from type 2 diabetes and outperforms current clinical guidance and islet-autoantibody assessment alone. C_LI

BackgroundDiabetic kidney disease (DKD) is a leading cause of kidney failure in individuals with type 2 diabetes (T2D), yet risk identification in routine clinical practice remains incomplete. A critical and often overlooked barrier is risk observability: how much of a patients underlying risk is actually captured in their clinical record at the time of screening. Existing prediction models evaluate performance using model-specific thresholds, making it difficult to understand how additional data sources alter real-world screening behavior or which individuals benefit when models are expanded. MethodsWe developed a series of five nested machine learning models evaluated at a one-year landmark following T2D diagnosis using data from the All of Us Research Program (N = 39,431; cases = 16,193). Each successive model added a distinct information layer -- intrinsic risk, laboratory snapshots, medication exposure, longitudinal care trajectories, and social determinants of health (SDOH) -- while retaining all prior features. All models were evaluated under a fixed screening policy targeting 90% specificity, so that the false positive rate remained constant as the information available to the model grew. External validation was conducted in the BioMe Biobank (N = 9,818) without retraining. ResultsDiscrimination improved consistently across layers, from AUROC 0.673 (M1) to 0.797 (M5). Under the fixed screening policy, sensitivity nearly doubled from 0.27 to 0.49, with a cumulative recovery of 30.4% of cases missed by the base model. Gains were driven by distinct subgroups at each transition: laboratory features identified biologically high-risk individuals; medication features captured those with high treatment intensity reflecting advanced cardiometabolic burden; longitudinal care trajectory features rescued cases with biological instability observable only through repeated measurements; and SDOH features recovered individuals with limited clinical observability, with rescue probability highest among those with the fewest recorded monitoring domains. Sparse data in the clinical record indicated low observability, not low risk. Social and genetic features each contributed most when downstream physiologic signal was limited, supporting a contextual rather than universal role for each. In BioMe, discrimination was attenuated (M4 AUROC 0.659), but the relative ordering of information layers was fully preserved, and a systematic upward shift in predicted probability distributions underscored the need for recalibration before deployment in a new setting. ConclusionsDKD risk detection in T2D is substantially improved by integrating complementary information layers under a fixed clinical screening policy, with gains arising from distinct domains that identify at-risk individuals in different clinical contexts. The layered landmark framework introduced here reveals how risk observability -- shaped by monitoring intensity, healthcare engagement, and access -- determines what a screening model can detect, and provides a foundation for context-aware EHR-based screening that accounts for data availability at the time of risk assessment. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=140 SRC="FIGDIR/small/26351384v1_ufig1.gif" ALT="Figure 1"> View larger version (51K): org.highwire.dtl.DTLVardef@175bfc4org.highwire.dtl.DTLVardef@181170dorg.highwire.dtl.DTLVardef@108c98org.highwire.dtl.DTLVardef@7e5c86_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical abstract.C_FLOATNO Study design and layered DKD screening framework The top row defines the cohort timeline, in which predictors are derived from clinical data collected between T2D diagnosis and the 1-year landmark, and incident DKD is ascertained after the landmark. The second row depicts the nested model architecture, in which five successive models sequentially incorporate intrinsic risk, laboratory snapshot features, medication exposure, longitudinal care trajectories, and social determinants of health, while retaining all features from prior layers. The third row summarizes model development in the All of Us Research Program (N = 39,431) and external validation in the BioMe Biobank (N = 9,818), where the same trained models and risk thresholds were applied without retraining. The bottom row highlights the three evaluation domains: predictive performance, fixed-policy screening, and missed-case recovery context. DKD, diabetic kidney disease; T2D, type 2 diabetes; PRS, polygenic risk scores; AUROC, area under the receiver operating characteristic curve; AUPRC, area under the precision-recall curve; PPV, positive predictive value; SHAP, SHapley Additive exPlanations. C_FIG

Background: Acute pancreatitis (AP) is an established risk factor for diabetes, with approximately 20% of children developing either prediabetes or diabetes within one year of their first episode. Little is known about the diabetes pathophysiology or which individuals are at highest risk. We aimed to evaluate whether genetic risk scores (GRS) for type 1 (T1D) and polygenic risk scores (PRS) type 2 diabetes (T2D) are associated with progression to dysglycemia following AP. Methods: Clinical data were available for 123 children (mean age (IQR), 12 (8-15) years; mean body mass index (BMI), 21.8) with AP who were followed for >1 year. Array genotyping coupled with imputation using the TOPMed reference panel was performed. Genetic ancestry was predicted using a random forest classifier. GRS for T1D and T2D were calculated using either an ancestry-appropriate (T1D-GRS) or a multi-ancestry (T2D-PRS) weighted framework. To evaluate risk compared to the population we used predefined GRS thresholds from UK Biobank. Results: Among the 123 subjects, 24 developed dysglycemia (5 with diabetes and 19 with prediabetes). The majority (75.6%, n=93) of children were of European ancestry. Comparison of the T1D-GRS burden with the UK BioBank showed numerically higher proportions for any given threshold. At the top 5% threshold, 9.7% of our cohort were classified as high-risk compared to 5% in UK Biobank (p<0.05). The elevated T1D-GRS could be primarily attributed to non-HLA variants and was more enriched in those testing positive for [&ge;]1 islet-autoantibody. The T2D-PRS was also elevated in the dysglycemic group but only reached statistical significance in those who were obese. Conclusion: These findings highlight the potential role of both T1D-GRS and T2D-PRS in investigating diabetes susceptibility following AP.

Objective: Despite the complex and non-linear progression of diabetes, its shared pathways with atherosclerotic cardiovascular disease (ASCVD) are conventionally described using models based on single time points. We identified longitudinal diabetes clusters before diagnosis using deep learning and studied their association with ASCVD events and mortality. Methods: We analyzed 157,670 visits from 15,871 adults (25-65 years) without diabetes from four pooled U.S. cohorts (median follow-up: 22 years [IQR: 9-30]). A gated recurrent unit model with decay (GRU-D) was used to predict 1-year risk of diabetes or censoring within 10 years, by learning longitudinal embeddings across 25 clinical characteristics and biomarkers. Parallel Factor Analysis-2 (PARAFAC-2) and Gaussian mixture models (GMM) were used to group longitudinal participant representations as clusters. Landmark time Cox proportional hazards regressions, relative to last observation in the training window, were used to study covariate-adjusted associations of clusters with ASCVD and mortality. Prognostic utility of clusters beyond the PREVENT risk score was assessed using Harrell's C-index. Findings were replicated in a fifth cohort. Results: The analytic sample was aged 49 years [SD: 11], 58% female, and 68% white; 1,202 (8%) developed diabetes within the first 10 years. We identified five clusters (Cluster A to E) that differed in their clinical characteristics over time. Cluster E (46%) had the highest cumulative incidence of diabetes in the study period, followed by Cluster C (40%) and Cluster A (38%). Cluster C, which was defined by older age, high blood pressure, and suboptimal renal function at the first visit, had higher rates of ASCVD (HR: 1.09, 95%CI: 0.98-1.21) and mortality (HR: 1.08, 95%CI: 1.00-1.16), relative to Cluster A despite being similar in age and BMI at the first visit. Relative to Cluster A, all other clusters had similar or lower rates of ASCVD and mortality. We observed substantial cluster effects for three clusters (Clusters C to E), which were based on only two cohorts. The two clusters (Clusters A and B) that included participants from all four cohorts were reproduced in the fifth cohort and showed similar rates of outcomes. Clusters did not improve ASCVD prognosis, relative to a model that included only the PREVENT risk score. Conclusions: Longitudinal clusters reveal substantial heterogeneity in the period before diabetes diagnosis, and their risk for ASCVD and mortality. However, clusters discovered may, in part, be explained by cohort effects from variations in recruitment and visit patterns after recruitment.

Background: Genetic variants impacting red blood cell biology disrupt the relationship between glycaemia and glycated haemoglobin (HbA1c), with implications for diagnosis and management of type 2 diabetes (T2D). Thalassaemia trait is estimated to affect 350 million people globally, but its impact on T2D and related outcomes is not clear. Methods: We explored associations between thalassaemia trait, HbA1c, and T2D diagnosis and complications in 43,088 British Bangladeshi and Pakistani participants in the Genes & Health study with linked multisource England National Health Service (NHS) electronic health record data and whole exome sequencing. Findings: 2,490 participants (5.8%) were heterozygous carriers of ClinVar pathogenic / likely pathogenic thalassaemia variants, however 3 in 4 of these were not diagnosed with thalassaemia in their NHS health records. rs33950507, a common variant causal for HbE thalassaemia, was associated with increased HbA1c (beta=0.13, 95%CI:0.08-0.18, p=7.8x10-8), but not glucose levels (beta=0.01, 95%CI:-0.04-0.06, P=0.72). rs33950507 was associated with increased hazards of prediabetes (HR=1.38, 95%CI:1.26-1.52, p=2.2x10-6) and T2D (HR=1.11, 95%CI:1.01-1.22, p=0.03), and reduced hazards of diabetic eye disease (HR=0.74, 95%CI:0.56-0.96, p=0.02) and cerebrovascular disease (HR=0.44, 95%CI:0.20-0.94, p=0.03). Sensitivity analyses suggested mediation by overdiagnosis and overtreatment of T2D. Interpretation: Alternatives to HbA1c, and/or precision medicine approaches to defining and managing hyperglycaemia, are needed, particularly on a global scale. This may be particularly relevant to individuals from ancestral groups among whom erythrocytic traits are more common but often undiagnosed. Funding: Wellcome Trust, MRC, NIHR, Barts Charity, Genes & Health Industry Consortium

BackgroundThe glucagon-like peptide-1 receptor (GLP1R) is a key regulator of glucose metabolism and appetite and a major therapeutic target for type 2 diabetes (T2D) and obesity. Genetic studies have implicated the GLP1R locus in both body mass index (BMI) and T2D, but it remains unclear whether their underlying genetic associations are the same. MethodsWe analyzed 431,107 participants of genetically inferred European ancestry from the Million Veteran Program. Within {+/-} 500 kb of GLP1R, we performed locus-wide linear regression models for BMI and logistic regression models for T2D, adjusted for age, sex, and 10 principal components. We identified primary and secondary BMI sentinel variants using conditional analyses and evaluated their associations with T2D. Bayesian fine-mapping was used to construct credible sets of GLP1R locus for BMI and T2D. ResultsConditioning on the primary sentinel variant rs12213929 (upstream of GLP1R, {beta} = 0.11; 95% CI 0.09-0.14; p = 1.94x10-17), we identified a secondary variant (rs13216992, intron of GLP1R) independently associated with BMI ({beta} = 0.10; 95% CI 0.07-0.13; p = 7.88x10-14). The two sentinel variants showed low linkage disequilibrium (r2 = 0.03). A two-variant allelic burden score (0-4; sum of the rs12213929 G-allele count and rs13216992 C-allele count) showed that participants with 4 risk alleles had 0.47 kg/m2 higher BMI than those with 0 risk alleles (95% CI 0.39-0.55; p < 2x10-16). Both variants were associated with higher T2D risk, but with distinct patterns after BMI adjustment: the rs12213929-T2D association persisted after adjustment for BMI (OR = 1.02; 95% CI 1.01-1.03; p = 0.0004), whereas the rs13216992-T2D association was fully attenuated (OR = 1.00; 95% CI 0.99-1.01; p = 0.68). Fine-mapping identified a compact 95% BMI credible set of 17 variants and a broader 95% T2D credible set of 42 variants, with all BMI credible variants contained within the T2D set. ConclusionsThe GLP1R locus harbors at least two independent BMI-associated variants that exhibit heterogeneous relationships with T2D: rs12213929 influences T2D risk partly through BMI-independent pathways, whereas rs13216992 appears to act predominantly via adiposity. These findings refine the genetic architecture at this key therapeutic target gene and provide a foundation for functional and pharmacogenomic studies to determine whether GLP1R variation can inform precision prevention and treatment of obesity and T2D.

BackgroundCardiometabolic diseases arise from metabolic dysfunction that develops decades before clinical onset. Conventional genetic risk models are typically derived in middle-aged or older populations, where genetic effects are confounded by cumulative environmental exposures, chronic comorbidities, and clinical interventions. Whether the life stage at which genetic liability is modelled influences the biological signal captured by polygenic scores remains unclear, particularly in underrepresented populations. We therefore tested whether genetic liability modelled in early adulthood, a period of relative physiological stability, is associated with cardiometabolic risk across the life course in Asian populations. MethodsWe developed a polygenic score for metabolic syndrome, GenMetS, using data from 1,368 Singaporean women aged 18-45 years. The model integrates 15 established polygenic scores for metabolic traits and applies elastic-net penalized regression to optimize variant weights. GenMetS was evaluated in five cohorts comprising 670,952 individuals aged 0-94 years across population-based and disease-enriched settings, including Asian and European ancestry groups. Associations with metabolic traits, cardiometabolic diseases, multimorbidity, and early-life growth patterns were assessed. ResultsIn Asian populations, GenMetS explained 5.0-12.4% of the variance in metabolic syndrome in adults and 10.3% in children, with negligible performance in European populations (R{superscript 2} < 0.001). Higher GenMetS was associated with increased odds of cardiometabolic diseases, including type 2 diabetes, heart failure, and stroke (odds ratios 1.32-1.52 per standard deviation). In UK Biobank participants of Asian ancestry, GenMetS improved discrimination of cardiometabolic multimorbidity beyond age alone. Associations were consistent across sexes. In children, higher GenMetS was associated with obesogenic growth trajectories and increased abdominal adiposity. ConclusionsGenetic liability to metabolic dysfunction modelled in early adulthood captures a stable biological signal associated with metabolic traits, disease risk, and multimorbidity from childhood to adulthood in Asian populations. These findings indicate that the life stage of model derivation shapes the biological signal captured by polygenic scores and support the development of life-stage- and ancestry-informed approaches for cardiometabolic risk assessment and prevention.

BackgroundThe first 1000 days from conception to age 2 years represent a critical window for kidney development, during which nutritional exposures may have lifelong programming effects. Whether early-life sugar restriction reduces long-term kidney disease risk remains unknown. MethodsUsing the UK sugar rationing policy (1942-1953) as a natural experiment, we compared risks of chronic kidney disease (CKD) and acute kidney injury (AKI) among 64,942 UK Biobank participants born around the rationing period. Duration of early-life exposure was categorised. Cox proportional hazards models estimated hazard ratios (HRs). Negative control analyses included non-UK-born UK Biobank participants and the Chinese CHARLS cohort. Mediation analyses integrated clinical phenotypes and metabolomic profiles. FindingsCompared with never-exposed individuals, those exposed to sugar rationing throughout the first 1000 days (in utero to age 2 years) had lower risks of CKD (adjusted HR 0.78, 95% CI 0.66-0.93) and AKI (0.79, 0.69-0.90). Negative control analyses showed null associations. Mediation analyses indicated that metabolic efficiency (basal metabolic rate), body composition (fat-free mass), and lipid metabolism mediated 4-9% of the protective association. A distinct metabolomic signature characterised by higher polyunsaturated fatty acids and lower VLDL subfractions was identified. InterpretationSugar restriction during the first 1000 days is associated with lower risks of CKD and AKI in adulthood, partially mediated by favorable metabolic efficiency, body composition, and lipid profiles. These findings identify early-life sugar exposure as a modifiable developmental programming factor for lifelong kidney health and support public health strategies to reduce added sugar intake during pregnancy and early childhood. FundingNational Natural Science Foundation of China, and others

Metformin has been linked to mortality benefits in type 2 diabetes that may extend beyond glycemic control, but population-level evidence connecting these benefits to inflammation-related pathways remains limited. Using NHANES 2013-2018 data with mortality follow-up through 2019, we examined associations between metformin use and four inflammatory markers, including neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), serum albumin, and high-sensitivity C-reactive protein (hs-CRP), and evaluated their relevance to all-cause and cardiovascular mortality. Among 2,122 adults with self-reported diabetes (60% metformin users; 2,116 with valid mortality follow-up), survey-weighted linear regression adjusted for demographic, socioeconomic, and metabolic covariates showed metformin use was associated with lower NLR ({beta} = - 0.35; 95% CI -0.57, -0.14), lower MLR ({beta} = -0.04; 95% CI -0.06, -0.02), and higher serum albumin ({beta} = +0.11 g/dL; 95% CI 0.06, 0.16); the hs-CRP association was directionally consistent but not significant. Associations for NLR and MLR were essentially unchanged after BMI and HbA1c adjustment, remained robust in an active comparator analysis against sulfonylurea monotherapy, and were consistent across propensity score and overlap weighting sensitivity analyses. Survey-weighted Cox regression linked metformin to lower all-cause (HR 0.64; 95% CI 0.48, 0.86) and cardiovascular mortality (HR 0.49; 95% CI 0.26, 0.94). NLR was independently associated with all-cause mortality, with the highest tertile carrying nearly twice the hazard of the lowest, and inclusion of NLR or MLR modestly attenuated the metformin-mortality association. Metformin use is associated with a distinct cellular immune-inflammation profile in adults with type 2 diabetes, supporting further investigation of non-glycemic pathways relevant to its long-recognized clinical benefits.

Background: Renal effects of statins in type 2 diabetes mellitus (T2DM) remain uncertain. We evaluated whether statin exposure is associated with time to dialysis initiation. Methods: We conducted a retrospective cohort study of adults with T2DM, indexing follow-up at diagnosis during first hospital admission (day 0) between january 2017 and march 2025. Statin use was modeled as time-varying from statin days; (classified in 3 categories: baseline users, new users, and never users). The primary outcome was dialysis. Analysis estimated cause-specific hazards, censoring deaths; proportional hazards were checked with prespecified windows of statin exposure (0?1, 1?3, > 3 years). Competing-risk analyses (Fine?Gray) assessed the sub-distribution hazard of dialysis with death as a competing event in two models: (i) prevalent users at baseline and (ii) new-users with post-initiation intervals of 30 and 90 days. An Observational Medical Outcomes Partnership Common Data Model standardized dataset of a Brazilian quaternary hospital, and the Real-World Data tool MD Clone were used in the study. Results: Of 36,246 adults identified, 32,125 entered the time-varying cohort (39,943 risk intervals; 656 dialysis events); median follow-up among censored patients was 753 days. At baseline, 70.3% never used statins, 5.5% were users (? 0 days), and 24.2% initiated after diagnosis. Crude dialysis incidence was 4.51 vs. 12.31 per 1,000 patient-years during unexposed vs. exposed time. In the adjusted time-varying Cox model, current statin exposure was associated with a modestly higher hazard of dialysis (HR = 1.043, 95% CI 1.011?1.077). In the new-users analysis, HRs were 0.83 (95% CI 0.66?1.05), and 0.73 (95% CI 0.57?0.92) with a 30-day and 90-day intervals, respectively. Conclusions: In this retrospective cohort of hospitalized diabetic patients at baseline, statin initiation at least 90-days in advance is associated with reduced indication of renal replacement therapy.

Aims CGM devices report glucose only within fixed limits (typically 40-400 mg/dL; 2.2-22.2 mmol/L), truncating extreme values to a boundary ("capping"). We characterised prevalence, duration, and consequences of capping in type 1 diabetes trial data. Materials and Methods We analysed 46,990,617 CGM readings from 948 participants across four publicly available clinical trial datasets (Dexcom G4 Platinum or G6 sensors). Capping prevalence, run duration, and associations with age, HbA1c and sex were characterised across all datasets. In the 77 participants of the Replace-BG trial CGM-plus-blood glucose monitor (BGM) arm, CGM-derived metrics were compared with contemporaneous BGM measurements across 1,162 non-overlapping 14-day windows. Results Between 93.5% and 100% of participants had at least one capped reading, and capped values comprised 0.47-0.98% of all readings. In the three datasets for which duration could be calculated, over 70% of upper-cap runs exceeded 15 minutes and over one third exceeded 60 minutes. Upper-limit capping was inversely associated with age (Spearman {rho} -0.20 to -0.47, p[&le;]0.002) in three of the datasets, and positively associated with baseline HbA1c ({rho} 0.39-0.62, p<0.001) in all four datasets. A within-participant analysis showed that capping burden did not predict CGM-BGM divergence in any summary metric (all p>0.2), and a systematic CGM-BGM offset in mean glucose and time in range (TIR) reflected the physiological lag between blood and interstitial fluid rather than capping artefact. Conclusions Sensor limit capping is near-universal in type 1 diabetes, produces sustained periods of right-censored glucose data disproportionately affecting younger patients, and does not substantially distort standard summary metrics at the population level. Clinicians and trialists should be aware that CGM data can confirm extreme glucose events but cannot quantify their severity.

Aims Type 2 diabetes mellitus (T2DM) increases risks of stroke and dementia, yet these risks vary across individuals. We hypothesized that clinically derived diabetes subtypes contribute to this heterogeneity. We aimed to identify data-driven subtypes using routine clinical features and examine their associations with dementia, stroke, mortality, and brain structure. Methods K-means clustering was applied to 14,353 UK Biobank participants with prevalent T2DM using age at diagnosis, body mass index, glycated hemoglobin, insulin resistance (triglyceride/HDL ratio), systolic blood pressure, and C-reactive protein. Cox models assessed associations with incident dementia (all-cause, Alzheimers disease [AD], vascular dementia [VaD]), stroke (all-cause, ischemic [IS], intracerebral hemorrhage [ICH]), and mortality. Brain MRI outcomes were analyzed in 779 participants using inverse probability-weighted linear regression. Results Three subtypes were identified: severe obesity-related inflammatory diabetes (SOID), mild metabolic diabetes (MMD, reference), and mild age-related hypertension-predominant diabetes (MARD-H). Compared with MMD, SOID showed higher risks of dementia (HR 1.24), VaD (HR 1.42), stroke (HR 1.38), IS (HR 1.48), all-cause mortality (HR 1.59), and cardiovascular death (HR 1.88). MRI showed lower gray matter volume and greater white matter hyperintensity burden in SOID. Conclusions Data-driven subtyping revealed heterogeneity in neurological risk in T2DM, with the obesity-inflammation subtype showing elevated vascular and neuroimaging risk.

Secondary hyperparathyroidism persists in the majority of kidney transplant recipients and is associated with adverse graft and cardiovascular outcomes. The immunosuppressive drug class used post-transplant may modulate parathyroid hormone (PTH) levels through distinct mechanisms: calcineurin inhibitors (CNI) stabilize PTH mRNA, while mTOR inhibitors (mTORi) suppress parathyroid cell proliferation in experimental models. We report supporting evidence from two independent analyses. In a multinational real-world database analysis (TriNetX Global Collaborative Network), kidney transplant recipients with documented mTORi use and eGFR in the target range had lower PTH than those on CNI across eGFR strata examined (15-30, 30-45, 45-60, 60-75, >75 mL/min/1.73 m2), with risk ratios for PTH >130 pg/mL ranging from 0.47 to 0.67 in propensity-matched analyses (all p < 0.05). The known confounders - calcium (higher in CNI) and phosphate (higher in mTORi) - both act to oppose this pattern, strengthening the possibility of a drug effect. In a longitudinal single-center cohort (n = 118; 796 PTH measurements), a linear mixed-effects model with time-varying mTORi exposure confirmed a 42% lower PTH during on-mTORi periods after adjustment for eGFR, transplant vintage, diabetes, age, and sex (fold-change 0.58 [95% CI 0.50-0.68]; p < 0.0001). These findings suggest a direct PTH-lowering effect of mTORi. Immunosuppression choice may be considered in the management of post-transplant hyperparathyroidism in selected patients.

Importance: Despite the benefits of statin therapy in individuals with diabetes, fewer than 70% of adults with diabetes meet contemporary guidelines for statin therapy and reducing low-density lipoprotein cholesterol (LDL) to <100 mg/dL. Evidence describing delays in statin initiation after diabetes diagnosis and associated clinical outcomes may motivate process of care interventions to improve guideline recommended care in individuals newly diagnosed with type 2 diabetes mellitus (T2D). Objective: To examine the timing of statin initiation and achievement of LDL <100 mg/dL after diabetes diagnosis, and to determine the association of early LDL reduction among statin initiators with incident atherosclerotic cardiovascular disease (ASCVD). Design: Retrospective observational cohort study using data from 2005-2021 Setting: Veterans Affairs Health Care System (VA) Participants: Individuals with newly diagnosed T2D Exposure: Primary exposure was ASCVD risk based on ACC/AHA Pooled Cohort Equations; secondary exposure was LDL <100 mg/dL in the first year after T2D diagnosis among statin initiators Main Outcomes and Measures: Co-primary outcomes were initiation of statin therapy and achievement of LDL <100 mg/dL within 5 years of diabetes diagnosis; incident 5-year ASCVD was a secondary outcome. Results: Among 100,406 individuals with newly diagnosed T2D, 59,615 were prescribed statin therapy within five years (59.4%), and 44,783 (57.5%) of those with LDL above goal achieved LDL <100 mg/dL within 5 years. Relative to those at low (<7.5%) 10-year ASCVD risk, individuals at intermediate (7.5-20%) and high (>20%) risk were more likely to be initiated on a statin (intermediate: Hazard Ratio [HR] 1.14 [95% CI 1.11, 1.17]; high: HR 1.16 [95% CI 1.13, 1.19]) and to achieve LDL <100 mg/dL (intermediate: HR 1.23 [95% CI 1.19, 1.26]; high: HR 1.34 [95% CI 1.30, 1.38]). Among those prescribed a statin within one year of diabetes diagnosis, achieving LDL <100 mg/dL in the first year after diabetes diagnosis was associated with lower risk of 5-year incident ASCVD (HR 0.84 [95% CI 0.77, 0.92]). Conclusions and Relevance: Gaps in guideline-directed primary prevention of ASCVD arise early following initial diabetes diagnosis. Guideline recommended early LDL lowering among statin initiators was associated with improved clinical outcomes.

Background Lifestyle interventions can increase the probability of remission of prediabetes to normal glucose tolerance, but their economic value remains unclear. We assessed the within-trial and lifetime-horizon modeled cost-effectiveness of intensive and conventional lifestyle interventions in risk-stratified participants with prediabetes. Methods A health economic evaluation was conducted alongside the 12-month multicenter PLIS trial (n=1,105). High-risk participants were randomized to intensive (HR-INT) or conventional (HR-CONV); low-risk participants to conventional lifestyle intervention (LR-CONV) or control (only short single consultation; LR-CTRL) with risk stratification based on insulin secretion, insulin sensitivity, and liver fat content. Within-trial analyses estimated incremental costs per additional remission to normoglycemia and per quality-adjusted life year (QALY). Lifetime cost-effectiveness was modelled using a four-state Markov Model. Findings At 12 months, HR-INT and LR-CONV increased remission compared with their respective comparators. The incremental cost per additional remission was {euro}7,081 (95% CI: dominated-47,277) for HR-INT and {euro}4,278 (1,312-11,793) for LR-CONV from a health insurance perspective. A willingness-to-pay of {euro}22,000 (HR-INT) and {euro}7,500 (LR-CONV) per additional remission corresponded to 90% probability of cost-effectiveness. Neither intervention was cost-effective in terms of QALYs gained within the 12-months period. Lifetime modelling suggested that both HR-INT and LR-CONV are not only cost-effective, but also cost-saving, relative to HR-CONV and LR-CTRL, respectively. Also in the probabilistic sensitivity analysis, most simulations indicated dominance (71.7% for HR and 88% for LR). Interpretation Based on short-term economic evaluation, the interventions assessed were cost-effective regarding additional participants with remission, not for incremental QALYs gained. Lifetime modelling suggests cost savings for both risk groups. Targeting populations with lifestyle interventions to achieve prediabetes remission seems to generate good value for money in the long term.

Objective: Patients and physicians frequently focus on HbA1c and weight alone. We hypothesized that individuals with similar HbA1c and BMI may present markedly distinct metabolic backgrounds. We investigated whether the adipo-B index- composite of adipose insulin resistance (adipo-IR) and beta-cell function (HOMA-B)-can uncover hidden heterogeneity in this clinically homogeneous population. Methods: A total of 399 newly diagnosed, drug-naive Japanese subjects with T2DM were analyzed. Histograms of HbA1c and BMI demonstrated peak distributions within HbA1c 8-10% and BMI 24-26. Based on these distributions, a clinically homogeneous subgroup was defined to minimize confounding by glycemic severity and adiposity. Metabolic parameters including FBG, insulin, FFA, HOMA-R, HOMA-B, adipo-IR, adipo-B, T-C, TG, HDL-C and non-HDL-C were analyzed. Simple regression, multivariable linear regression, and subgroup stratification analyses were performed. Results: Despite comparable HbA1c and BMI by design, adipo-B stratification revealed significant differences in HOMA-B, FFA, non-HDL-C, and TG, whereas HOMA-R stratification identified only higher insulin and adipo-IR without differences in lipids or HOMA-B. Thus, adipo-B-but not HOMA-R-identified a lipotoxic, beta-cell-stressed phenotype invisible to conventional markers. Simple regression showed significant positive correlations between adipo-B and HbA1c, FBG, FFA, T-C, TG, and non-HDL-C, and negative correlations with insulin and HOMA-B. Multivariable linear regression confirmed that adipo-B was independently associated with non-HDL cholesterol, TG, and FFA after adjustment for HbA1c and BMI. Conclusion: Even among patients with identical HbA1c and BMI, the adipo-B index uncovers clinically relevant metabolic heterogeneity, supporting its role as a functional marker of the adipose-pancreas axis and a potential tool for precision phenotyping in early T2DM.

Exercise is a cornerstone therapy for diabetes because working skeletal muscles take up glucose at dramatically greater rates than postprandial insulin-stimulated glucose uptake and, notably, do so without a requirement for insulin. This remarkable ability of working muscles is preserved in diabetes, when muscles become resistant to insulin. However, the mechanism of insulin-independent glucose uptake by working muscles is not fully understood. Here we describe a previously unrecognized glucose uptake pathway in muscle, which we refer to as "mSGLT" based on shared properties with the Sodium Glucose Linked Transporter family. In contrast to the abundant GLUT4 transporter, mSGLT is not regulated by insulin, requires Na,K-ATPase-2 activity, and transports the hexose -methyl-D-glucoside (MDG), a glucose derivative that is handled by SGLTs but not GLUT4. The mSGLT pathway and GLUT transport pathways are independent and additive. In addition to exercise, mSGLT imports glucose under other conditions of adrenergic stimulation, which inhibits pancreatic insulin release and reduces the insulin sensitivity of muscle. SGLT2-specific antibodies recognize a protein in muscle of similar size to the kidney SGLT2; this protein localizes to the muscle t-tubules, together with Na,K-ATPase-2 and MAP17, the regulatory subunit of SGLT2. However, skeletal muscles do not express a full-length transcript of Slc5a2 (SGLT2), and SGLT2-specific inhibitors do not inhibit mSGLT with high affinity. The novel transporter may be a muscle variant of Slc5a2 that results from post-transcriptional or post-translational mechanisms. mSGLT and its regulation offer potential muscle-specific therapeutic targets for treating hyperglycemia and other conditions when insulin-stimulated glucose disposal into muscle is impaired.

Introduction: APOL1 high-risk variants markedly increase susceptibility to kidney disease among individuals of African ancestry; however, only a subset of carriers develops clinically significant CKD or ESKD. This discrepancy highlights a gap between genetic risk and clinical trajectory. Current prognostic tools rely primarily on eGFR and albuminuria, which incompletely reflect the underlying biological processes driving APOL1-associated kidney injury. We hypothesized that plasma biomarkers reflecting inflammatory and tubular injury pathways could identify biologically active disease states within this genetically high-risk population and improve prognostic stratification. Methods: Participants from the Mount Sinai BioMe Biobank carrying two APOL1 high-risk alleles (G1, G1; G1, G2; or G2 G2) were followed for a median of 6 years. Baseline plasma biomarkers of inflammation and tubular injury (TNFR1, TNFR2, KIM-1, MCP-1, YKL-40, IL-18, suPAR) were measured. The composite outcome was sustained 40% decline in eGFR or ESKD. Multivariable Cox models assessed associations between biomarkers and outcomes. A weighted biomarker risk score was derived from tertile-based hazard ratios and categorized into low-, moderate-, and high-risk groups. Results: Among 498 participants (median eGFR 83 ml/min/1.73 m2), 80 (16.1%) reached the composite outcome. Higher concentrations of TNFR1, TNFR2, suPAR, KIM-1, and IL-18 were independently associated with kidney events after multivariable adjustment. Event rates were 7% in the low-risk group, 16% in the moderate-risk group, and 36% in the high-risk group. Conclusions: Plasma biomarkers reflecting inflammatory and tubular injury pathways reveal marked heterogeneity in kidney outcomes among individuals with high-risk APOL1 genotypes. Integration of these signals into a biology-weighted score identifies distinct prognostic phenotypes beyond genotype and traditional clinical measures, supporting multidomain biomarker frameworks for risk stratification and potential trial enrichment in APOL1-associated kidney disease.