Diabetologia

AimsPopulation datasets are increasingly used to study type 1 or 2 diabetes, and inform clinical practice. However, correctly classifying diabetes type, when insulin treated, in population datasets is challenging. Many different approaches have been proposed, ranging from simple age or BMI cut offs, to complex algorithms, and the optimal approach is unclear. We aimed to compare the performance of approaches for classifying insulin treated diabetes for research studies, evaluated against two independent biological definitions of diabetes type. MethodWe compared accuracy of thirteen reported approaches for classifying insulin treated diabetes into type 1 and type 2 diabetes in two population cohorts with diabetes: UK Biobank (UKBB) n=26,399 and DARE n=1,296. Overall accuracy and predictive values for classifying type 1 and 2 diabetes were assessed using: 1) a type 1 diabetes genetic risk score and genetic stratification method (UKBB); 2) C-peptide measured at >3 years diabetes duration (DARE). ResultsAccuracy of approaches ranged from 71%-88% in UKBB and 68%-88% in DARE. All approaches were improved by combining with requirement for early insulin treatment (<1 year from diagnosis). When classifying all participants, combining early insulin requirement with a type 1 diabetes probability model incorporating continuous clinical features (diagnosis age and BMI only) consistently achieved high accuracy, (UKBB 87%, DARE 85%). Self-reported diabetes type alone had high accuracy (UKBB 87%, DARE 88%) but was available in just 15% of UKBB participants. For identifying type 1 diabetes with minimal misclassification, using models with high thresholds or young age at diagnosis (<20 years) had the highest performance. An online tool developed from all UKBB findings allows the optimum approach of those tested to be selected based on variable availability and the research aim. ConclusionSelf-reported diagnosis and models combining continuous features with early insulin requirement are the most accurate methods of classifying insulin treated diabetes in research datasets without measured classification biomarkers.

Aims/hypothesisGenetic prediction of type 2 diabetes risk has proven difficult using current methods. Recent studies have shown that genetic variants associated with physical activity behavior are linked to type 2 diabetes incidence. This study investigated how a polygenic risk score (PRS) for type 2 diabetes relates to the incidence of type 2 diabetes and its comorbidities and whether incorporating genetic risk from physical activity-related traits and measured lifestyles improves prediction. We hypothesized that adding physical activity genotypes into prediction models would improve predictive accuracy. MethodsPRSs were calculated for 279,373 Finns in the FinnGen cohort (average age 62 years, 52% women). Cox proportional hazards models were used with follow-up from birth. In addition, we assessed whether predictive ability (concordance index) improved when PRSs for physical activity, sedentary time, cardiorespiratory fitness, muscle strength, and body mass index were included alongside the type 2 diabetes PRS. Finally, we assessed how smoking and body mass index changed the models predictive ability. ResultsEach standard deviation unit increase in the type 2 diabetes PRS was associated with an 8% higher risk of developing type 2 diabetes. Among individuals with type 2 diabetes, the PRS was linked to higher risks of comorbidities: 4% higher for nephropathy and retinopathy and 5% for severe cardiovascular disease, but not neuropathy. Physical activity-related PRSs were also independently associated with the risk of type 2 diabetes--lower risk for physical activity (7%), cardiorespiratory fitness (6%), and muscle strength (4%) and higher risk for sedentary time (14%) and body mass index (35%). However, physical activity-related PRSs did not significantly improve the models concordance index (0.644 before vs. 0.672 after adding all other PRSs). In contrast, including body mass index and smoking status increased predictive ability (c-index 0.744). Conclusions and applicabilityPRSs for type 2 diabetes and physical activity-related phenotypes independently predict the incidence of type 2 diabetes and comorbidities. However, adding physical activity-related scores to the model does not significantly improve prediction beyond the type 2 diabetes score. Notably, the PRS for body mass index was better than the PRS for type 2 diabetes in predicting type 2 diabetes incidence. These findings support the hypothesis that genetic pleiotropy may partially explain associations between type 2 diabetes and physical activity behavior. Summary boxesWhat is already known about this subject? O_LIGenetic factors contribute substantially to the risk of type 2 diabetes, but it is primarily a multifactorial condition in which modifiable lifestyle factors--including physical activity--play a critical role in onset and progression. C_LIO_LIGenetic variants related to physical activity behavior have been associated with type 2 diabetes. C_LIO_LIThe clinical utility of polygenic risk scores in predicting type 2 diabetes risk remains limited, as they explain only a small proportion of genetic variance and provide minimal improvement in risk prediction beyond established clinical risk factors. C_LI What is the key question? O_LICan the risk estimates for type 2 diabetes and its comorbidities be improved by incorporating genetic risk factors associated with physical activity and measured lifestyle behaviors? C_LI What are the new findings? O_LIPolygenic risk scores for both type 2 diabetes and physical activity-related phenotypes independently predict the incidence of type 2 diabetes and its comorbidities. C_LIO_LIIncorporating physical activity-related polygenic risk scores into the model does not significantly improve predictive accuracy beyond the type 2 diabetes risk score alone. C_LIO_LIThe findings support the hypothesis that genetic pleiotropy may partially explain associations between type 2 diabetes and physical activity behavior. C_LI How might this impact on clinical practice in the foreseeable future? O_LIAlthough polygenic risk scores for type 2 diabetes may aid in identifying high-risk individuals for targeted prevention, their integration into clinical practice requires further validation. C_LI

BackgroundGenetic risk scores (GRS) for type 1 diabetes (T1D) have been developed primarily in European populations, limiting their generalisability across ancestries. Indians differ from Europeans in clinical characteristics of T1D and overall genetic architecture, yet systematic evaluation of T1D GRS performance in multi-regional Indian cohorts is lacking. MethodsThe study included 597 T1D patients and 3347 non-diabetic controls from different regions in India. Genotyping, imputation, quality control analysis, and construction of the 67-SNPs T1D GRS were performed using standardised pipelines. Discriminative performance was assessed using Receiver Operative Curve-Area under Curve (ROC-AUC) analysis, and optimal thresholds were derived using Youdens index. HLA-DQ diplotype frequencies were compared, and association analysis was conducted using multivariable logistic regression. FindingsT1D GRS showed consistent discriminative performance across Indian cohorts [ROC-AUC=0.84 (range=0{middle dot}78-0{middle dot}87)], supporting its comprehensive use for T1D classification in India. Notably, its performance was lower in islet cell autoantibody (IA) negative compared with IA positive T1D patients (ROC-AUC, 0{middle dot}75 vs 0{middle dot}85) and in adult-onset than in childhood-onset patients (0{middle dot}74 vs 0{middle dot}84). We observed a lower frequency of protective HLA-DQ diplotypes and a strong association of HLA-DQ81 containing diplotypes in childhood-onset T1D. Application of an India-specific T1D GRS score improved the sensitivity than the European cut-off. InterpretationT1D GRS is a valuable unified diagnostic tool in Indians, but its performance varies by islet cell autoantibody status and age at onset, likely reflecting population-specific HLA architecture. European-derived T1D GRS thresholds under-classify the genetic risk, highlighting the importance of ancestry-aware optimisation in Indians. FundingCDRC grant CDRC202111026 and CSIR Intramural Grant P50. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSPrevious studies have shown that a 67-SNPs T1D genetic risk score (GRS) can distinguish T1D patients from non-diabetic controls and other forms of diabetes, but its performance varies across ancestries. Islet cell autoantibodies (IA) have important diagnostic value for classifying type 1 diabetes (T1D). However, their prevalence in India varies widely, with up to one-quarter of patients testing negative, limiting their clinical utility. Evidence supporting the use of the T1D GRS in India, combined with IA antibodies status is limited to a single cohort representing one linguistic group. The applicability of T1D GRS across multi-centric clinical settings has not been systematically evaluated. Added value of this studyThis study validates the 67-SNPs T1D GRS across multiple Indian cohorts representing major linguistic groups, supporting its use as a unified diagnostic tool. Differences in T1DGRS performance between childhood-and adult-onset T1D are linked to enrichment of protective HLA-DQ diplotypes in adult-onset disease, providing genetic insight into disease heterogeneity. The study also demonstrates that European-derived GRS thresholds systematically under-classify genetic risk in Indians and the population-specific threshold is essential. Implications of all the available evidenceThe European-derived T1D GRS can be applied across Indian clinical settings with consistent discriminative performance. However, its utility is influenced by islet cell autoantibody status and the age at onset of disease. Ancestry-aware threshold optimisation substantially improves diagnostic accuracy and is essential for equitable implementation of T1D GRS in Indians. Larger studies are needed to identify population-specific risk variants and further refine genetic tools for clinical diagnosis.

Background and hypothesisDiabetic kidney disease (DKD) is a severe diabetic complication affecting one third of individuals with type 1 diabetes. Although several genes and common variants have been associated with DKD, much of the predicted inheritance remain unexplained. Here, we performed next-generation sequencing to assess whether low-frequency variants -- single or aggregated -- contribute to the missing heritability in DKD. MethodsWe performed whole-exome sequencing (WES) of 498 individuals and whole-genome sequencing (WGS) of 599 individuals with type 1 diabetes. After quality control, we had next-generation sequencing data available for altogether 1064 individuals, of whom 546 had developed either severe albuminuria or end-stage kidney disease, and 528 had retained normal albumin excretion despite a long duration of type 1 diabetes. Single variants and gene aggregate tests were performed separately for WES and WGS data and combined with meta-analysis. Furthermore, we performed genome-wide aggregate analyses on genomic windows (sliding-window), promoters, and enhancers with the WGS data set. ResultsIn single variant meta-analysis, no variant reached genome-wide significance, but a suggestively associated THAP7 rs369250 variant (P=1.50x10-5) was replicated in the FinnGen general population GWAS data for chronic kidney disease (CKD) and DKD phenotypes. Gene-aggregate meta-analysis identified suggestive evidence (P<4.0x10-4) at four genes for DKD, of which NAT16 and LTA (TNB-{beta}) replicated in FinnGen. Of the intergenic regions suggestively associated with DKD, the enhancer on chromosome 18q12.3 (P=3.94x10-5) showed interaction with the METTL4 gene; the lead variant was replicated, and predicted to alter Mafb binding. ConclusionsOur sequencing-based meta-analysis revealed multiple genes, variants and regulatory regions suggestively associated with DKD. However, as no variant or gene reached genome-wide significance, further studies are needed to validate the findings. What was knownO_LIGenetics is an important factor in the development and progression of diabetic kidney disease (DKD) in individuals with type 1 diabetes. C_LIO_LIPreviously identified genetic associations have mostly been common variants as they originated from GWAS studies. Based on inheritance estimates, the current findings only explain a fraction of the predicted disease risk. C_LI This study addsO_LIOur study with 1097 sequenced individuals with type 1 diabetes is to date one of the largest sequencing studies on DKD in type 1 diabetes. C_LIO_LIThe study reveals several suggestive variants, genes and intergenic regulatory regions associated with DKD. Low-frequency protein-altering variants inside NAT16 and LTA (encoding for TNF-{beta}), and chromosome 18q12.3 enhancer variant linking to METTL4 were also replicated in FinnGen kidney disease phenotypes. C_LI Potential impactO_LIThe results suggest novel genes that may be important for the onset and development of serious DKD in individuals with type 1 diabetes. In addition to revealing novel biological mechanisms leading to DKD, they may reveal novel treatment targets for DKD. However, further validation and functional studies are still needed. C_LI

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.

BackgroundType 2 diabetes (T2D) is a multi-organ disease defined by hyperglycemia resulting from different disease mechanisms. Using clinical parameters measured at diagnosis (age, BMI, HbA1c, HOMA2-B, HOMA2-IR and GAD autoantibodies) adult patients with diabetes have been reproducibly clustered into five subtypes, that differed clinically with respect to disease progression and outcomes.1 In this study we use genetic information to investigate if these subtypes have distinct underlying genetic drivers. MethodsGenome-wide association (GWAS) and genetic risk score (GRS) analysis was performed in Swedish (N=12230) and Finnish (N=4631) cohorts. Family history was recorded by questionnaires. ResultsSevere insulin-deficient diabetes (SIDD) and mild obesity-related diabetes (MOD) groups had the strongest family history of T2D. A GRS including known T2D loci was strongly associated with SIDD (OR per 1 SD increment [95% CI]=1.959 [1.814-2.118]), MOD (OR 1.726 [1.607-1.855]) and mild age-related diabetes (MARD) (OR 1.771 [1.671-1.879]), whereas it was less strongly associated with severe insulin-resistant diabetes (SIRD, OR 1.244 [1.157-1.337]), which was similar to severe autoimmune diabetes (SAID, OR 1.282 [1.160-1.418]). SAID showed strong association with the GRS for T1D, whereas the non-autoimmune subtype SIDD was most strongly associated with the GRS for insulin secretion rate (P<7.43x10-9). SIRD showed no association with variants in TCF7L2 or any GRS reflecting insulin secretion. Instead, only SIRD was associated with GRS for fasting insulin (P=3.10x10-8). Finally, a T2D locus, rs10824307 near the ZNF503 gene was uniquely associated with MOD (ORmeta=1.266 (1.170-1.369), P=4.3x10-9). ConclusionsNew diabetes subtypes have partially different genetic backgrounds and subtype-specific risk loci can be identified. Especially the SIRD subtype stands out by having lower heritability and less involvement of beta-cell related pathways in its pathogenesis. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSIn March 2018 we suggested a novel subclassification of diabetes into five subtypes. This classification was based on clustering using clinical parameters commonly measured at diabetes diagnosis (age at diabetes onset, HbA1c, bodymass index, presence of GAD autoantibodies and HOMA2 indices for insulin resistance and secretion). These subtypes differed with respect to clinical characteristics, disease progression and risk of complications, but it remained unclear to what extent these subtypes have different underlying pathologies. In our original publication we analysed a small set of genetic risk variants for diabetes and found differential associations between subtypes, suggesting potential aetiological differences. Added value of this studyIn this study we have conducted a full genome analysis of the original ANDIS cohort, including genome-wide association studies and polygenic risk score analysis with replication in an independent cohort. We have also compared heritability and prevalence of having a family history of diabetes in the subtypes. Implications of all the available evidenceWe demonstrate that stratification into subtypes facilitates identification of genetic risk loci and that the aetiology of the subtypes is at least partially distinct. These results are especially important for the future study and treatment of individuals belonging to the severe insulin-resistant diabetes (SIRD) subtype, whose pathogenesis appears to differ substantially from that of traditional T2D.

ObjectiveCertain ethnicities such as South Asians and East Asians have higher rates of type 2 diabetes mellitus, in part, driven by insulin deficiency. Insulin deficiency can be due to beta-cell insufficiency, low beta-cell mass, or early cell death. Transcription factor XBP1 maintains beta-cell function and prevents early cell death by mitigating cellular endoplasmic reticulum stress. We examine the role of XBP1 expression in maintaining glucose homeostasis, glycaemic control, and response to diabetes therapeutics. Research Design and MethodsColocalisation analyses were used to determine if expression of XBP1 in pancreatic islets and type 2 diabetes shared common causal genetic variants. We identify a lead eQTL variant associated exclusively with XBP1 expression and examine its association HOMA-B and stimulated glucose in cohorts of newly diagnosed Asian Indians from Dr. Mohans Diabetes Specialities Centre, India (DMDSC) and the Telemedicine Project for Screening diabetes and complications in rural Tamil Nadu (TREND). We then examine longer term glycaemic control using HbA1c in Asian Indian cohorts, the Tayside Diabetes Study (TDS) of white European ancestry in Scoltand, and the Genes & Health (G&H) study of British South Asian Bangladeshi and Pakistani ancestry. Finally, we assess the effect of eQTL variant on drugs designed to improve insulin secretion (sulphonylureas and GLP1-RA). ResultsVariants affecting XBP1 expression in the pancreatic islets colocalised with variants associated with T2DM risk in East Asians but not in white Europeans. Lower expression of XBP1 was associated with higher risk of T2DM. rs7287124 was the lead eQTL variant and had a higher risk allele frequency in East (65%) and South Asians (50%) compared to white Europeans (25%). In 470 South Asian Indians, the variant was associated with lower beta-cell function and higher stimulated glucose ({beta}log HOMAB =-0.14, P=5x10-3). Trans-ancestry meta-analysed effect of the variant in 179,668 individuals was 4.32 mmol/mol (95%CI:2.60,6.04, P=8x10-7) per allele. In 477 individuals with young onset diabetes with non-obese BMI, the per allele effect was 6.41 mmol/mol (95%CI:3.04, 9.79, P =2x10-4). Variant carriers showed impaired response to sulphonylureas. ConclusionXBP1 expression is a novel target for T2DM with particular value for individuals of under-researched ancestries who have greater risk of young, non-obese onset diabetes. The effect of XBP1 eQTL variant was found to be comparable with or greater that the effect of novel glucose-lowering therapies. Visual abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=140 SRC="FIGDIR/small/23289501v1_ufig1.gif" ALT="Figure 1"> View larger version (35K): org.highwire.dtl.DTLVardef@1f6a99aorg.highwire.dtl.DTLVardef@f8fc04org.highwire.dtl.DTLVardef@6983baorg.highwire.dtl.DTLVardef@1473e16_HPS_FORMAT_FIGEXP M_FIG Visual abstract: ER: Endoplasmic Reticulum, UPR: Unfolded Protein Response, IRE1:Inositol-Requiring Enzyme 1, mRNA: messenger ribonucleic acid, ERAD: Endoplasmic Reticulum Associated protein Degradation, eQTL: expression Quantitative Trait Loci, HbA1c: glycated haemoglobin. Created with Biorender.com C_FIG

Background. Cardiovascular disease accounts for most mortality in type 2 diabetes (T2D), yet treatment is anchored on glucose-derived metrics and fasting insulin is rarely measured. We tested whether fasting insulin carries cardiovascular-mortality information across the dysglycaemic spectrum and approaches non-diabetic levels with longer oral-therapy T2D duration. Methods and Findings. We analysed six NHANES cycles [2007-2018] linked to National Death Index follow-up through 2019. Baseline characteristics were described in participants with complete kidney-function data: normoglycaemic (n = 2,913), pre-diabetes (n = 3,993), and oral-therapy T2D (n = 1,381). Survey-weighted Cox models used the covariate-complete mortality sample (pre-diabetes n = 4,019; oral-therapy T2D n = 1,387) and adjusted for age, sex, race/ethnicity, BMI, smoking, physical activity, education, poverty:income ratio, and insulin assay generation. Per +1 natural-log-unit fasting insulin, all-cause hazard ratios (HRs) were 1.69 (95% CI 1.14-2.50; P = 0.008) in pre-diabetes and 0.71 (0.51-0.98; P = 0.039) in oral-therapy T2D; cardiovascular HRs were 3.09 (1.69-5.63; P < 0.001) and 0.50 (0.26-0.99; P = 0.046), respectively. Cancer mortality was not associated with fasting insulin. In oral-therapy T2D, geometric-mean fasting insulin remained 1.6- to 2.3-fold the normoglycaemic referent across duration bands. After BMI adjustment, fasting insulin declined during the first five years (-4.0%/year, P = 0.018) and was flat thereafter (-0.1%/year, P = 0.66). Conclusions. Fasting insulin predicted cardiovascular mortality in pre-diabetes. In oral-therapy T2D, the inverse association was most consistent with survivor effects and accumulated renal and vascular damage. Fasting insulin remained above normoglycaemic levels throughout treated T2D.

AimsWe investigated the relationship between polygenic score (PGS) for BMI and other adiposity PGS with age at type 2 diabetes onset in white European (EUR) and south Asian (SAS) ancestries, and the mediating role of BMI. MethodsIn this retrospective study, using polygenic score (PGS) for BMI, clinically measured BMI and age at type 2 diabetes onset, we conducted mediation analysis separately for SAS (n=3,901, Genes & Health) and EUR (n=729, UK Biobank) aged 40 years or older. For SAS, we also used multivariable linear regression with backward selection to identify the best adiposity PGS (waist circumference (WC), waist-to-hip ratio (WHR), visceral adipose tissue (VAT), body fat (BF), trunk fat (TF) and gluteofemoral fat (GFAT), hand-grip strength (HGS)) predicting age at type 2 diabetes onset. ResultsA one SD increment in BMI-PGS was associated with earlier type 2 diabetes onset by -0.73 years (95%CI -1.01; -0.45) in SAS and -0.57 years (95%CI -1.05; -0.08) in EUR. BMI fully mediated the PGS effect in EUR (100%) and only partially in SAS (28%). Alongside BMI-PGS, WC-PGS and TF-PGS were good at discriminating measured BMI, WC and WHTR in SAS and were correlated with BMI-PGS. Other best predictors of early onset type 2 diabetes in SAS were WC-PGS, WHR-PGS, BF-PGS and GFAT-PGS, which differed between SAS subgroups and by sex. ConclusionsThese findings underscore the importance of incorporating adiposity-related genetics in predicting type 2 diabetes onset among SAS and demonstrate the limitations of using BMI alone to capture associated risk, particularly in diverse populations with typically lower BMI. Research in contextWhat is already known about this subject? O_LISouth Asians develop type 2 diabetes, on average, a decade earlier than white Europeans and at lower BMI levels. C_LIO_LIThe well-established association between BMI and type 2 diabetes risk in white Europeans is more complex in south Asians, who have different patterns of adipose tissue distribution that are not well reflected in BMI. C_LIO_LIThe contribution of adiposity-related polygenic scores (PGS) to type 2 diabetes onset has not been examined across ancestries. C_LI What is the key question? O_LIWhat is the association between adiposity PGS and type 2 diabetes onset across ancestries, what how much of this effect is mediated by BMI? C_LI What are the new findings? O_LIHigher BMI-PGS is associated with earlier type 2 diabetes onset in both south Asians and white Europeans; this relationship is fully mediated by BMI in white Europeans but only partially in south Asians. C_LIO_LIOther PGS for central adiposity are significant predictors of early type 2 diabetes onset and central adiposity anthropometrics in south Asians. C_LI How might this impact on clinical practice in the foreseeable future? O_LIThese findings highlight the importance of incorporating adiposity-related genetics in predicting type 2 diabetes onset in diverse populations and underscore the need to move beyond BMI when assessing metabolic risk and potentially evaluating the effectiveness of interventions. C_LI

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 ([&ge;]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 [&ge;]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

ObjectiveWith the high prevalence of pediatric obesity and overlapping features between diabetes subtypes, accurately classifying youth-onset diabetes can be challenging. We aimed to develop prediction models that, using characteristics available at diabetes diagnosis, can identify youth who will retain endogenous insulin secretion at levels consistent with type 2 diabetes (T2D). MethodsWe studied 2,966 youth with diabetes in the prospective SEARCH study (diagnosis age [&le;]19 years) to develop prediction models to identify participants with fasting c-peptide [&ge;]250 pmol/L ([&ge;]0.75ng/ml) after >3 years (median 74 months) of diabetes duration. Models included clinical measures at baseline visit, at a mean diabetes duration of 11 months (age, BMI, sex, waist circumference, HDL-C), with and without islet autoantibodies (GADA, IA-2A) and a Type 1 Diabetes Genetic Risk Score (T1DGRS). ResultsModels using routine clinical measures with or without autoantibodies and T1DGRS were highly accurate in identifying participants with c-peptide [&ge;]0.75 ng/ml (17% of participants; 2.3% and 53% of those with and without positive autoantibodies) (area under receiver operator curve [AUCROC] 0.95-0.98). In internal validation, optimism was very low, with excellent calibration (slope=0.995-0.999). Models retained high performance for predicting retained c-peptide in older youth with obesity (AUCROC 0.88-0.96), and in subgroups defined by self-reported race/ethnicity (AUCROC 0.88-0.97), autoantibody status (AUCROC 0.87-0.96), and clinically diagnosed diabetes types (AUCROC 0.81-0.92). ConclusionPrediction models combining routine clinical measures at diabetes diagnosis, with or without islet autoantibodies or T1DGRS, can accurately identify youth with diabetes who maintain endogenous insulin secretion in the range associated with type 2 diabetes.

The aim of this study was to determine the effectiveness of the Type 1 Diabetes Genetic Risk Score (T1D GRS) for the prioritisation of children with newly diagnosed hyperglycaemia for genetic testing of monogenic diabetes. MethodsA cohort of 808 children and adolescents with newly diagnosed hyperglycaemia were collected. All underwent standard clinical follow-up and genetic testing based on the knowledge and means accessible at the time. In this cohort and in 189 control subjects (165 monogenic diabetes patients and 24 healthy individuals), we assessed the T1D GRS2 and 10 SNP GRS scores. We assessed the T1D GRS2 cut-off in our cohort and investigated its utility in addition to negative autoantibody status for the prioritisation of cases for genetic testing for monogenic diabetes. Genetic testing included Sanger sequencing, panel sequencing and MLPA. ResultsApplying T1D GRS2 in addition to negative autoantibodies on the newly diagnosed hyperglycaemia cohort substantially decreased the number of unnecessarily tested cases. The pick-up rate was increased three-fold, while the sensitivity of the prioritisation decreased only slightly from 77.8% to 72.2% when compared with autoantibodies alone. The majority of monogenic diabetes cases that escaped this prioritisation for genetic testing had low levels of a single autoantibody and were most probably false positives in the autoantibody testing. The monogenic cases that would not be prioritised using GRS2 and autoantibodies were diagnosed based on clinical phenotype. On the other hand, two monogenic diabetes cases with HNF1B-MODY were not originally diagnosed and were identified only thanks to their low GRS2 value. ConclusionsUsing T1D GRS in combination with autoantibody testing is effective in decreasing the number of unnecessarily genetically tested cases. This approach, used in addition to the standard clinical evaluation, can be a valuable tool in the early selection of suitable candidates for molecular testing for monogenic diabetes. Research in ContextO_LIWhat is already known about this subject? O_LIDiagnosing monogenic diabetes is important, because gene-tailored treatment is available. In children and adolescents, monogenic diabetes has to be differentiated mostly from type 1 diabetes. C_LIO_LIIndividuals with type 1 diabetes and monogenic diabetes have a different genetic risk for type 1 diabetes. C_LIO_LIThe genetic risk score for type 1 diabetes (T1D GRS) can be computed from disease-associated polymorphisms. C_LI C_LIO_LIWhat is the key question? O_LIWhat is the utility of T1D GRS as a tool for the prioritisation of cases for genetic testing of monogenic diabetes? C_LI C_LIO_LIWhat are the new findings? O_LIMore than half of the children with diabetes with negative autoantibodies (type 2 diabetes excluded) and T1D GRS2 below a cut-off was confirmed genetically as having monogenic diabetes. C_LIO_LIThe combination of T1D GRS2 + negative autoantibodies increased the pick-up rate by genetic testing three-fold compared with negative autoantibodies alone, with only small decrease in sensitivity of the prioritisation. C_LIO_LILow T1D GRS can draw interest to cases that would otherwise not be suspected of having monogenic diabetes. C_LI C_LIO_LIHow might this affect clinical practice in the foreseeable future? O_LIChildren and adolescents with confirmed diabetes and normal BMI, negative autoantibodies and low T1D GRS can be prioritised for genetic testing soon after diabetes diagnosis. C_LI C_LI

Sequencing of large cohorts offers an unprecedented opportunity to identify rare genetic variants and to find novel contributors to human disease. We used gene-based collapsing tests to identify genes associated with glucose, HbA1c and type 2 diabetes (T2D) diagnosis in 363,977 exome-sequenced participants in the UK Biobank. We identified associations for variants in GCK, HNF1A and PDX1, which are known to be involved in Mendelian forms of diabetes. Notably, we uncovered novel associations for GIGYF1, a gene not previously implicated by human genetics, in diabetes. GIGYF1 predicted loss of function (pLOF) variants associated with increased levels of glucose (0.77 mmol/L increase, p = 4.42 x 10-12) and HbA1c (4.33 mmol/mol, p = 1.28 x 10-14) as well as T2D diagnosis (OR = 4.15, p= 6.14 x10-11). Multiple rare variants contributed to these associations, including singleton variants. GIGYF1 pLOF also associated with decreased cholesterol levels as well as an increased risk of hypothyroidism. The association of GIGYF1 pLOF with T2D diagnosis replicated in an independent cohort from the Geisinger Health System. In addition, a common variant association for glucose and T2D was identified at the GIGYF1 locus. Our results highlight the role of GIGYF1 in regulating insulin signaling and protecting from diabetes. Author SummaryGenetic studies focused on high impact variants in protein-coding regions of the genome can provide valuable insight into the biology of human disease. As these variants tend to be rare, studying them requires large cohort sizes and methods to aggregate variants that are likely to have a similar biological impact. We studied how rare genetic variants contribute to type 2 diabetes (T2D) using sequencing data from 363,977 participants in the UK Biobank, employing methods to aggregate variants at the level of individual genes. As well as identifying genes known to be involved in inherited forms of diabetes, we uncovered a novel association for GIGYF1. GIGYF1 loss of function associated with increased risk of T2D and increased levels of the diabetes biomarkers glucose and HbA1c. This association was also seen in an independent dataset. GIGYF1 encodes a protein that binds a negative regulator of the insulin receptor that has not been well-characterized in the literature. By highlighting the importance of GIGYF1 in modulating insulin signaling these results may lead to new therapeutic approaches for diabetes as well as a new appreciation for GIGYF1 loss of function as a genetic risk factor for T2D.

Immune-mediated destruction of the beta-cells in the pancreatic islets of Langerhans is the underlying cause of type 1 diabetes (T1D). Despite decades of research, the exact mechanisms involved at the beta-cell level during the development of disease remain poorly understood. This includes the mode(s) of beta-cell death and signaling events implicated in exacerbating local islet inflammation and immune cell infiltration, commonly known as insulitis. In disease models, beta-cell apoptosis seems to be the predominant cell death form which has led to the general assumption that beta cells mostly die by apoptosis in T1D. However, apoptosis is an anti-inflammatory programmed cell death mechanism, and this dogma is therefore challenged by the pathogenetic nature of T1D, as a progressive increase in islet inflammation is seen. This infers that other modes of beta-cell death that inherently increase insulitis may predominate. One such mechanism could be the newly characterized form of programmed cell death; pyroptosis (from the Greek "fire-falling"). Pyroptosis is characterized by gasdermin-mediated cell lysis with a bursting release of pro-inflammatory factors. Beta-cell death by pyroptosis in T1D may therefore offer a plausible explanation for the exacerbated paracrine islet inflammation that spreads during the progression of insulitis. Here, we briefly debate the evidence supporting beta-cell pyroptosis in T1D as a central mechanism of islet inflammation and beta-cell demise. The paper intends to challenge the current understanding of beta-cell destruction to move the field forward. Importantly, we present experimental data from human islets and EndoC-{beta}H5 cells that directly support beta-cell pyroptosis as a rational death mechanism in T1D. We suggest a model of beta-cell demise in T1D in which pyroptosis plays a prominent role in concert with other cell death mechanisms. As the role of pyroptosis in disease is still in its infancy, we hope also to inspire researchers working in other disease fields.

BackgroundHeterogeneity exists in type 1 diabetes (T1D) development and presentation. Islet autoantibodies form the foundation for T1D diagnostic and staging efforts. We hypothesized that autoantibodies can be used to identify heterogeneity in T1D before, at, and after diagnosis, and in response to disease modifying therapies. at clinically relevant timepoints throughout T1D progression. MethodsWe performed a systematic review assessing 10 years of original research studies examining relationships between autoantibodies and heterogeneity during disease progression, at the time of diagnosis, after diagnosis, and in response to disease modifying therapies in individuals at risk for T1D or within 1 year of T1D diagnosis. Results10,067 papers were screened. Out of 151 that met data extraction criteria, 90 studies characterized heterogeneity before clinical diagnosis. Autoantibody type/target was most commonly examined, followed by autoantibody number, titer, order of seroconversion, affinity, and novel islet autoantibodies/epitopes. Recurring themes included positive relationships of autoantibody number and specific types and titers with disease progression, differing clinical phenotypes based on the order of autoantibody seroconversion, and interactions with age and genetics. Overall, reporting of autoantibody assay performance was commonly included; however, only 43% (65/151) included information about autoantibody assay standardization efforts. Populations studied were almost exclusively of European ancestry. ConclusionsCurrent evidence most strongly supports the application of autoantibody features to more precisely define T1D before clinical diagnosis. Our findings support continued use of pre-clinical staging paradigms based on autoantibody number and suggest that additional autoantibody features, particularly when considered in relation to age and genetic risk, could offer more precise stratification. Increased participation in autoantibody standardization efforts is a critical step to improving future applicability of autoantibody-based precision medicine in T1D. Plain Language SummaryWe performed a systematic review to ascertain whether islet autoantibodies, biomarkers of autoimmunity against insulin-producing cells, could aid in stratifying individuals with different clinical presentations of type 1 diabetes. We found existing evidence most strongly supporting the application of these biomarkers to the period before clinical diagnosis, when certain autoantibody features (number, type) and the age when they develop, can provide important information for patients and care providers on what to expect for future type 1 diabetes progression.

ImportanceThe hemoglobin A1c and fasting plasma glucose (FPG) have known limitations for diabetes diagnosis, but models to identify individuals who would benefit from 2-hour oral glucose tolerance testing (OGTT) are limited. ObjectiveTo determine if OGTT-only diagnosed diabetes has comparable outcomes to A1c- or FPG-diagnosed diabetes and if standard clinical features could be leveraged to identify undiagnosed diabetes. DesignMultivariable prediction model development and validation. SettingUS National Health and Nutrition Examination Survey (NHANES) data with corresponding US National Center for Health Statistics (NCHS) mortality data. ParticipantsOf 105,862 NHANES subjects from 1999 to 2016, we identified 13,800 subjects with FPG, A1c, or OGTT results (11,550 with mortality data) and 92,062 other subjects (53,255 with mortality data). ExposureOGTT-diagnosed diabetes Main Outcomes and MeasuresThe primary outcomes were association of mortality with diabetes diagnostic approach and models to diagnose diabetes. We used a gradient boosted machine decision tree to predict diabetes from standard clinical features. In the test set, we compared the AUROC and the net benefit by decision curve analysis to A1c, FPG, and a combination of the two. We performed survival analysis based on method of diabetes diagnosis and diabetes model predictions. ResultsThe rate of OGTT-only diabetes was 1.34%. Subjects with OGTT-only diabetes had equivalent risk of mortality compared to subjects with FPG- or A1c-diagnosed diabetes after adjusting for age, sex, and race/ethnicity. A model using the A1c and standard clinical features (A1c+ model) outperformed the A1c to exclude diabetes (Sensitivity at Youdens Index: 0.72 vs. 0.37). Adding FPG to that model (A1c/FPG+) outperformed FPG for excluding diabetes (Sensitivity: 0.87 vs. 0.48). Subjects with A1c/FPG+-predicted diabetes but sub-diagnostic A1c and FPG had equivalent mortality (HR=8.2, p<2*10-16) to those with A1c or FPG-diagnosed diabetes (comparison p<0.17). Conclusions and RelevanceDiabetes diagnosed by OGTT alone has equivalent mortality to A1c and FPG-diagnosed diabetes. A model using standard clinical features can bolster the A1c and FPG to identify potentially undiagnosed diabetes. Model predictions associated with mortality equivalently to having a diabetes diagnosis. Implementation of a clinical decision support tool could improve diagnosis of diabetes and lead to earlier interventions. Key PointsO_ST_ABSQuestionC_ST_ABSAre individuals with diabetes only diagnosable by oral glucose tolerance test (OGTT) at similar risk of mortality as those diagnosed by hemoglobin A1c or fasting plasma glucose (FPG)? Can readily-available clinical data improve diagnosis? FindingsStratifying NHANES subjects by OGTT-only (1.34%) vs A1c or FPG diagnosis (4.13%) of diabetes found that OGTT-only diagnosis had equivalent mortality. A model combining A1c and standard clinical features (A1c+) had superior AUROC for diabetes compared to the A1c alone. The addition of FPG (A1c/FPG+) had superior AUROC compared to the FPG and A1c+. The A1c/FPG+ predictions strongly associated with mortality in subjects with sub-diagnostic A1c and FPG. MeaningIncorporation of clinical features can improve diabetes diagnosis missed by A1c and FPG. Model prediction of diabetes associates with mortality.

Hyperglycaemia is a hallmark of diabetes and a major risk factor for diabetic kidney disease (DKD). However, the molecular consequences of long-term cumulative hyperglycaemia (CH) remain unclear. As a stable epigenetic modification, DNA methylation may capture past glycaemic exposure. Here, we assessed CH-associated DNA methylation in 1,245 participants with type 1 diabetes (T1D) from Finland and the United Kingdom-Republic of Ireland cohorts. We identified 17 CH-associated CpGs, with the strongest association at cg19693031 (TXNIP). Longitudinal analyses demonstrate that these CH-associated DNA methylation levels remain stable despite short-term glycaemic fluctuations, suggesting lasting epigenetic imprints of earlier metabolic control. Integrative analyses combining genomic, epigenetic, and proteomic data characterized these CpGs and potential target proteins. Mendelian randomization suggested a causal association between cg20853880 (KLF11) and DKD, supported by chromatin accessibility and kidney KLF11 expression. Our findings suggest that epigenetic changes contribute to metabolic memory and may mediate the effects of hyperglycaemia on DKD.

Aims/hypothesisCurrent guidelines recommend use of sodium-glucose cotransporter-2 inhibitors (SGLT2 inhibitors) for kidney protection in people with type 2 diabetes and early-stage chronic kidney disease (CKD) based on a urinary albumin/creatinine ratio (uACR) of [&ge;]3 mg/mmol. However, individuals with a normal uACR or low-level albuminuria were not represented in kidney outcome trials, leaving uncertainty about absolute treatment benefit in this group. To address this gap and support treatment decisions in clinical practice, we developed and validated a model to predict individual-level kidney protection benefit through use of SGLT2 inhibitors. MethodsThis observational cohort study used electronic health record data from UK primary care (Clinical Practice Research Datalink, 2013-2020) of adults with type 2 diabetes, eGFR [&ge;]60 ml/min per 1.73 m2 and uACR <30 mg/mmol, without heart failure or atherosclerotic vascular disease, who were starting treatment with either SGLT2 inhibitors or the comparator drugs dipeptidyl peptidase-4 (DPP4) inhibitors/sulfonylureas. First, we confirmed the real-world applicability of the relative treatment effect from a previous SGLT2 inhibitor trial meta-analysis, using overlap-weighted Cox proportional hazards models. Second, we assessed calibration of the CKD-PC risk score for kidney disease progression ([&ge;]50% eGFR decline, end-stage kidney disease or kidney-related death). Third, we integrated the relative treatment effect with the risk score to predict 3-year individual-level absolute risk reductions for SGLT2 inhibitors, and validated the accuracy of predictions vs overlap-weighted estimates based on observed data. Finally, we compared the clinical utility of a model-based treatment strategy with that of the [&ge;]3 mg/mmol albuminuria threshold. ResultsIn 53,096 initiations of SGLT2 inhibitor treatment compared with 88,404 initiations of DPP4 inhibitor/sulfonylurea treatment, there was a 42% lower relative risk of kidney disease progression with SGLT2 inhibitors (HR 0.58; 95% CI 0.48, 0.69), consistent with a previous trial meta-analysis. The CKD-PC risk score did not require recalibration (calibration slope 1.05; 95% CI 0.94, 1.17). The median overall model-predicted absolute risk reduction with SGLT2 inhibitors was 0.37% at 3 years (IQR 0.26-0.55), and showed good calibration (calibration slope 1.10; 95% CI 1.09, 1.12). As an illustration of clinical utility, using the model predictions to target the same proportion of the population (n=25,303, 17.9%) as the albuminuria threshold would prevent over 10% more events over 3 years (253 vs 228) by identifying a subgroup of 6.7% of individuals with uACR <3 mg/mmol who showed significantly greater absolute risk reduction in response to SGLT2 inhibitor treatment than the remainder with uACR <3 mg/mmol (3.2% vs 1.2% in extended 5-year observational analyses, p=0.05). Conclusions/interpretationA model adapting the international CKD-PC risk score can accurately predict the individual-level kidney protection benefit from treatment with SGLT2 inhibitors in people with type 2 diabetes and no or early-stage CKD. This could guide treatment decisions in clinical practice worldwide. and could target treatment more effectively than the [&ge;]3 mg/mmol albuminuria threshold recommended by current international guidelines. Research in contextO_ST_ABSWhat is already known about this subject?C_ST_ABSO_LISodium-glucose cotransporter-2 (SGLT2) inhibitors reduce the risk of kidney failure in people with type 2 diabetes C_LIO_LICurrent guidelines recommend use of SGLT2 inhibitors for kidney protection in individuals with type 2 diabetes and urinary albumin/creatinine ratio [&ge;]3 mg/mmol, but this is an extrapolation beyond current evidence from kidney outcome trials C_LIO_LIIt is unclear which people with type 2 diabetes and preserved eGFR and a normal urinary albumin/creatinine ratio or low-level albuminuria have clinically relevant kidney protection benefit from SGLT2 inhibitors C_LI What is the key question?O_LICan a model integrating an established risk score with the relative treatment effect from a SGLT2 inhibitor trial meta-analysis accurately predict individual-level kidney protection benefit? C_LI What are the new findings?O_LIThe model accurately predicted individual-level kidney protection benefit with SGLT2 inhibitor treatment in an external validation using UK primary care data C_LIO_LICompared with the [&ge;]3 mg/mmol albuminuria threshold, the model more effectively identified individuals who are likely to benefit, and could prevent more adverse kidney events C_LI How might this impact on clinical practice in the foreseeable future?O_LIThe model enables individualised prescribing of SGLT2 inhibitors for kidney protection, which could optimise treatment allocation and improve kidney outcomes C_LI

ObjectivesGenetic Risk scores (GRS) classify diabetes types, type 1 (T1D) and type 2 (T2D) in Europeans but the power is limited in other ancestries. We explored the performance of T1DGRS and potential reasons for inferior discrimination ability in diabetes-type classification in Indians. Research Design and MethodsIn a well-characterized Indian cohort comprising 645 clinically diagnosed T1D, 1153 T2D and 327 controls, we estimated the discriminative ability of T1DGRS (comprising 67 SNPs from Europeans) using receiver operating characteristics-area under the curve (ROC-AUC). We also compared the islet autoantibody status (AA), frequency and effect size of various HLA alleles/haplotypes between Indians and Europeans. ResultsThe T1DGRS was discriminative of T1D from T2D and controls but the ability is lower in Indians than Europeans (AUC=0.83 vs 0.92 respectively, p<0.0001). The T1DGRS was higher in AA-positive patients compared to AA-negative patients [13.01 (12.79-13.23) vs 12.09 (11.64-12.56)], p<0.0001) and showed greater discrimination in the AA-positive T1D (ROC-AUC 0.85). While association of common HLA-DQA1[~]HLA-DQB1 haplotypes with T1D is replicated, important differences in the risk allele frequency, nature/direction and magnitude of association between Indians and Europeans were noted. ConclusionsA T1DGRS derived from Europeans is discriminative of T1D in Indians, highlighting similarity in heritability of T1D. Differences in allele frequency, effect size and directionality, especially in the HLA region are important contributors to inferior discrimination performance of T1DGRS in Indians. Further studies of diverse populations may improve its performance.

PurposeThe Biomarkers of heterogeneity in type 1 diabetes study cohort was set up to identify genetic, physiological and psychosocial factors explaining the observed heterogeneity in disease progression and the development of complications in people with long-standing type 1 diabetes (T1D). Data and samples are available for new studies and collaborations. ParticipantsData- and samples were collected in two subsets. 1) A prospective cohort of 611 participants aged [&ge;]16 years with [&ge;]5 years T1D duration was recruited from four Dutch Diabetes clinics between June 2016 and March 2021. At baseline and 1- and 2-year follow-up visits, physical assessments were performed, and blood and urine samples were collected. Participants completed questionnaires about diabetes-related problems, quality of life, neuropathy and impaired awareness of hypoglycaemia at baseline and at the last follow-up visit. A subgroup of participants underwent mixed-meal tolerance tests (MMTT) at baseline (n=169) and at 1-year follow-up (n=104). Genetic data and linkage to medical and administrative records were also available. 2) A second cross-sectional cohort, aiming to include 200 participants aged [&ge;]18 years with [&ge;]35 years T1D duration, was recruited from 7 centres, collecting measurements and samples plus 5-year retrospective data. Findings to dateFasting residual C-peptide secretion associated with decreased risk of impaired awareness of hypoglycaemia. Stimulated residual C-peptide was detectable in an additional 10% of individuals compared with fasting residual C-peptide secretion. MMTT measurements at 90 minutes and 120 minutes showed good concordance with the MMTT total area under the curve. An overall decrease of C-peptide at 1-year follow-up was observed. Future plansResearch groups are invited to consider the use of this data and sample collection. Future work will include additional hormones, beta-cell-directed autoimmunity, specific immune markers, microRNAs, metabolomics and gene expression data, combined with glucometrics, anthropometric/clinical data and additional markers of residual beta-cell function. Strengths and limitations of this studyO_ST_ABSStrengthsC_ST_ABS- The Biomarker cohort is a large longitudinal prospective cohort study with three time points, collecting biosamples and clinical data from participants with well-established and long-standing type 1 diabetes ([&ge;]5 years). - A subgroup with detailed clinical data underwent MMTT tests at two timepoints allowing further residual beta-cell marker studies. - The Biomarker and Long-Term type 1 Diabetes cohorts represent a "real-world" population, also including participants from non-academic/-specialised centres. Limitations- Despite the fact that data and biosamples were collected from more than 600 participants, this number may be too low for (sub) stratification of the data (e.g. insulin delivery modality, different treating centres and therapies etc.). - In the prospective group there was a relatively high dropout rate of 25% after 2 years, largely affected by the Covid-19 outbreak.