Objective: Long-term side effect surveillance after immunotherapy clinical trial participation in individuals with, or at risk for, type 1 diabetes is needed. Research Design and Methods: Participants from 14 randomized controlled trials (47%) joined Long-term Investigative Follow-Up in Type 1 Diabetes TrialNet or Immune Tolerance Network Type 1 Diabetes Extension Study and 98% answered at least one health-related outcomes question. Results: Participants were followed for a median of 1.78 years (IQR 0.89-4.04) after original trial completion and, in some, up to 18 years. There were no differences in the frequency of self-reported health outcomes including moderate-to-severe hypoglycemia events, hospitalizations, serious infections (including COVID-19 infection), new allergic reactions, autoimmune disease, and cancer incidence between active and placebo therapy (p>0.05 for all, n=209-473 responses per question). Conclusions: We observed no difference in self-reported health outcomes between those who received active immunotherapy versus placebo. Continued long-term follow-up of immunotherapy trial participants is essential.

An inflection point (IP) marking accelerated β-cell decline occurs ∼1-2 years before type 1 diabetes diagnosis. Precisely determining this timing could optimize clinical intervention. We developed machine learning models to predict proximity to the inferred metabolic IP using oral glucose tolerance test (OGTT) data from islet autoantibody-positive individuals in the TrialNet Pathway to Prevention study. OGTTs were retrospectively labeled by estimating time to the IP using thresholds from 1.2 to 1.6 years. Models were trained on engineered glucose and C-peptide dynamics, slopes, and composite indices, with feature selection by recursive feature elimination (RFE). Classifiers included support vector machines (SVM), random forests, and gradient boosting; a Cox proportional hazards model was also applied to estimate visit-level time to diagnosis and derive time-to-IP predictions. External validation was performed in the independent Diabetes Prevention Trial-Type 1 cohort. The best-performing model-SVM with RFE at the 1.4-year threshold-achieved an area under the curve of 0.77 (95% CI 0.72-0.82). The Cox model provided complementary numeric estimates of time to diagnosis and time to IP, sensitive to the cohort-level lead time offset (Δ). These findings show that machine learning and survival analysis can support early metabolic shift detection, enabling timely risk stratification and personalized monitoring in at-risk individuals. ARTICLE HIGHLIGHTS:We undertook this study to improve early identification of the metabolic inflection point (IP) preceding clinical type 1 diabetes in autoantibody-positive individuals. We aimed to develop and validate machine learning models using oral glucose tolerance test-derived dynamic features to detect proximity to the IP. A support vector machine trained on TrialNet Pathway to Prevention and tested on Diabetes Prevention Trial-Type 1 achieved an area under the curve of 0.77 at 1.4 years prior to diagnosis, with strong calibration and interpretability. Additionally, a Cox proportional hazards model provided numeric estimates of time to IP, offering complementary predictions. These results can support earlier intervention and timely monitoring through personalized oral glucose tolerance test-based risk stratification.

OBJECTIVE Islet autoantibody positivity (AB+) has not been evaluated across different race, ethnicity, and socioeconomic deprivation categories in individuals at risk for type 1 diabetes. We examined data from TrialNet of persons screened for ABs and evaluated patterns by race and ethnicity and socioeconomic deprivation. RESEARCH DESIGN AND METHODS This analysis included 139,963 relatives of people with type 1 diabetes screened by TrialNet between 9 January 2012 and 31 December 2022. Race and ethnicity were categorized as non-Hispanic White, Hispanic, non-Hispanic Black, and non-Hispanic other. Home addresses were used to assign deprivation, ranging from least deprived (1) to most deprived (100), analyzed in quintiles. Descriptive and multivariate analyses assessed associations among race and ethnicity, deprivation, and AB+, adjusting for age and protocol period. RESULTS Single AB+ (SAB+) and multiple AB+ (MAB+) varied by age, race and ethnicity, and deprivation. Younger individuals had increased odds of MAB+ and decreased odds of SAB+. Non-Hispanic Black participants had significantly higher odds of AB+ (adjusted odds ratio 1.44 [95% CI 1.29, 1.62]), SAB+ (1.59 [1.38, 1.85]), and MAB+ (1.26 [1.06, 1.51]), while Hispanic individuals had higher odds of SAB+ (1.26 [1.15, 1.39]) and lower odds of MAB+ (0.81 [0.72, 0.91]) compared with non-Hispanic White individuals. Lowest deprivation was associated with higher odds of MAB+ (1.21 [1.06, 1.37]) but not SAB+. CONCLUSIONS AB+ patterns vary by race, ethnicity, and deprivation. Screening and prevention strategies should be designed to identify individuals at risk based on these factors to enable equitable risk stratification and access to diabetes preventive therapies.

Type 1 diabetes arises from the interplay of genetic susceptibility and environmental exposures, leading to autoimmune β-cell destruction. Although disease-modifying therapies (DMTs) can delay progression to clinical (stage 3) type 1 diabetes, treatment responses remain inconsistent and transient. The marked heterogeneity of type 1 diabetes, shaped by age, sex, race or ethnicity, and genetic background, underscores the need to elucidate distinct mechanistic pathways. Among environmental contributors, obesity stands out as a compelling modifiable target. Data from The Environmental Determinants of Diabetes in the Young (TEDDY), Type 1 Diabetes TrialNet, and other longitudinal cohorts link BMI and adiposity to the onset of islet autoimmunity, progression through preclinical stages, and development of stage 3 type 1 diabetes. These associations are not uniform; heightened susceptibility to adiposity-related risk is seen among younger children, Hispanic populations, and individuals with specific HLA genotypes. Despite robust epidemiologic evidence, the biological pathways connecting elevated BMI to autoimmune β-cell destruction remain incompletely defined. Emerging data implicate a network of immunologic and metabolic disturbances, including insulin resistance, β-cell stress, chronic adipose tissue inflammation, altered adipokine signaling, and gut microbiome changes, that collectively heighten β-cell vulnerability, amplify autoreactive immune responses, and drive metabolic decompensation toward clinical disease. Elucidating these mechanisms and identifying related biomarkers are critical to advancing precision prevention. In future studies, investigators should evaluate whether modifying elevated BMI or targeting obesity-associated immunologic and metabolic pathways can alter the preclinical trajectory of type 1 diabetes. Such mechanistic understanding may help curb type 1 diabetes incidence and improve outcomes for populations most vulnerable to obesity-related risk.

OBJECTIVE Elevated BMI is associated with increased risk of progression to stage 3 type 1 diabetes (T1D). We tested whether normalization of BMI is associated with lower stage 3 risk among autoantibody-positive individuals with overweight or obesity. RESEARCH DESIGN AND METHODS We studied 833 autoantibody-positive participants in the TrialNet Pathway to Prevention study with overweight or obesity at baseline. Cumulative incidence of stage 3 T1D was evaluated using Cox proportional hazards models. Reported P values were adjusted for baseline age, BMI z score, and T1D stage, unless otherwise noted. RESULTS During follow-up, BMI normalized in 26.8% of participants and was associated with a lower risk of progression to stage 3 T1D (hazard ratio 0.516, P < 0.001) over a median follow-up time of 4.1 years. This association was primarily observed in youth (n = 421, hazard ratio 0.497, P = 0.001), including boys <12 years old (n = 140, P < 0.001) and girls ≥12 years old (n = 73, P = 0.043), but not in adults (n = 412, P = 0.130). BMI normalization was also associated with a lower risk of transition from stage 1 to stage 2 and from stage 2 to stage 3, but not clearly from stage 0 to stage 1 (P = 0.010, P = 0.003, and P = 0.051, respectively, adjusted for baseline age and BMI z score). In exploratory models, adjustment for insulin sensitivity/resistance indices or oral disposition index weakened the association between BMI normalization and lower progression risk, whereas adjustment for insulin secretion, glycemia-adjusted C-peptide, or progression-risk indices generally strengthened it. CONCLUSIONS Among autoantibody-positive individuals with overweight or obesity, BMI normalization was associated with a 48.4% lower risk of clinical T1D, primarily in youth.

We analyzed baseline serum samples from 41 individuals newly diagnosed with type 1 diabetes (T1D) enrolled in the AbATE trial (NCT00129259) to identify metabolic predictors of Teplizumab response. Responders to Teplizumab were defined as individuals who exhibited < 45% decline in baseline C-peptide levels at 2 years after start of treatment. We used a semi-targeted metabolomics approach via liquid chromatography–high-resolution tandem mass spectrometry. We identified fifteen significant (p<0.05) metabolites, including amino acids and their derivatives, tricarboxylic acid (TCA) cycle intermediates, and microbially derived metabolites. Responders exhibited higher levels of TCA cycle metabolites, amino acid derivatives, and microbial metabolites, whereas nonresponders showed elevated glutamate and acylcarnitines. These metabolites were used to train a supervised Random Forest (RF) model to predict treatment response. Model performance was evaluated using a 70/30 training/testing split, 5-fold cross-validation, bootstrap resampling (1,000 iterations), and permutation testing (1,000 permutations). The RF classifier achieved an accuracy of 0.769 and an area under the receiver operating characteristic curve of 0.881 in the test dataset. These findings suggest baseline serum metabolomic signatures have the potential to predict responders to Teplizumab with accuracy. This could potentially be applicable to other immunotherapies in T1D preventative efforts. Further validation of our findings is needed.

Introduction and Objective: Quantifying β-cell function in preclinical T1D is essential for understanding progression and intervention design. The Oral Minimal Model (OMM) applied to OGTTs provides a mechanistic estimate of β-cell responsiveness (ɸtotal). We derive stage-specific ɸtotal trajectories, compared to AUC C-peptide, aligned to 4 years before Stage 1 (S1) and Stage 2 (S2) transition, to better characterise β-cell decline. Methods: ɸtotal and AUC C-peptide were derived from 2597 OGTTs in 867 TrialNet Pathway to Prevention/TEDDY participants with a transition event. Generalized Additive Mixed Models adjusted for sex, age, BMI, and participant level random effects modelled trajectories. 2 year progression probability was evaluated using all S1 observations (7152 OGTTs; N=2583). Results: ɸtotal was higher in S1 than S2 (-32.2%, p < 0.001), with clearer stage contrast than AUC C-peptide (-2.9% p=0.03). Over the last 4 years of S1, ɸtotal had steady exponential decline (~4.8%/yr, p<0.001), undetectable until the final year with AUC C-peptide (Fig 1). In S2, decline was more variable with both measures until the final year of accelerated deterioration. In S1, ɸtotal ≥ 3.8 had 89% negative predictive value for non-progression within 2 years. Conclusion: ɸtotal captures early functional deterioration in S1 T1D compared to AUC C-peptide and provides robust rule-out information for near-term progression, supporting greater sensitivity and statistical power for monitoring and trial design in early-stage T1D. Disclosure A. Carr: None. S. Perazzolo: None. J.A. Hedrick: None. P. Senior: Consultant; Ended; Abbott. Consultant; Current; Dexcom, Inc. Consultant; Ended; GlaxoSmithKline plc. Research Support; Current; Eli Lilly and Company. Consultant; Current; Novo Nordisk, Sana Biotechnology Inc., Sanofi, Vertex Pharmaceuticals Incorporated. Consultant; Ended; Ypsomed AG. H.M. Ismail: None. T. Moran: Other - Data Safety Monitoring Board Member; Ended; Novo Nordisk. Advisory Panel; Ended; Abbott Diabetes. Research Support; Current; Abbott Diabetes. C.M. Dayan: Consultant; Current; Sanofi, SAB Biotherapeutics, Inc., Immunocore, Ltd. Consultant; Ended; Vertex Pharmaceuticals Incorporated. Consultant; Current; Quell. Advisory Panel; Current; Amarna, ArgenX, Amgen Inc. A. Galderisi: Advisory Panel; Current; Sanofi. Advisory Panel; Ended; Novo Nordisk. Advisory Panel; Current; vTv Therapeutics. Funding Breakthrough T1D (3-SRA-2022-1186-S-B, 3-SRA-2023-1422-S-B, 3-SRA-2023-1422-S-B). The Type 1 Diabetes TrialNet Study Group is a clinical trials network currently funded by the National Institutes of Health (NIH) through the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), the National Institute of Allergy and Infectious Diseases (NIAID), and the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), through the cooperative agreements U01 DK060782, U01 DK060916, U01 DK060987, U01 DK061010, U01 DK061016, U01 DK061029, U01 DK061030, U01 DK061034, U01 DK061035, U01 DK061036, U01 DK061037, U01 DK061040, U01 DK061041, U01 DK061042, U01 DK061055, U01 DK061058, U01 DK084565, U01 DK085453, U01 DK085461, U01 DK085463, U01 DK085465, UC4 DK085466, U01 DK085476, U01 DK085499, U01 DK085504, U01 DK085505, U01 DK085509, U01 DK097835, U01 DK103153, U01 DK103180, U01 DK103266, U01 DK103282, U01 DK106984, U01 DK106993, U01 DK106994, U01 DK107013, U01 DK107014, and a contract HHSN267200800019C; the National Center for Research Resources, through Clinical Translational Science Awards UL1 RR024131, UL1 RR024139, UL1 RR024153, UL1 RR024975, UL1 RR024982, UL1 RR025744, UL1 RR025761, UL1 RR025780, UL1 RR029890, UL1 RR031986, UL1 TR001872, and General Clinical Research Center Award M01 RR00400.

The regulatory approval and growing pipeline of disease-modifying therapies for presymptomatic type 1 diabetes have sparked an increase in screening for the presence of islet autoantibodies in individuals who may not have yet developed clinical symptoms of disease. The staging system that is currently employed to define type 1 diabetes diagnosis in presymptomatic and symptomatic phases is based on the presence of islet autoantibodies and glycemic status. Here, we make the case to consider using C-peptide values to provide context for glycemia. Inclusion of C-peptide could result in better prediction of progression to clinical disease and more specificity for features typically associated with type 1 diabetes (and, thus, be less susceptible to confounding influences on glycemia like age and insulin resistance). It also could more rapidly identify responses to disease-modifying therapies. The implementation and translation of these findings will be key to long-term success in type 1 diabetes prediction and disease modification strategies.

Introduction and Objective: Although age and BMI have been included as risk factors for T1D in Ab+ individuals, a definitive association between glucose with either age or BMI has not been evident. We hypothesized that a recently identified fraction of glucose, independent of insulin secretion (INDEP), obscures associations of glucose with age and BMI. Methods: 5475 TrialNet Pathway to Prevention Ab+ participants (age: 2-50 years) were studied. Index60, a composite measure of glucose and C-peptide, was a surrogate for insulin secretion. To obtain the INDEP fraction of AUC glucose, we first determined the AUC glucose fraction dependent on insulin secretion (DEP) from the intercept and slope (DEP=130.74+11.08* Index60) of a simple linear regression model. INDEP was then calculated by subtracting DEP from total AUC glucose. Results: Unfractionated AUC glucose had a small correlation with age (r=-0.06). However, after separation into fractions, DEP was inversely correlated with age (r=-0.46), whereas INDEP was positively correlated (r=0.15). INDEP was higher in those >40.0 years (highest decade; n=524) than in those <10.0 years (lowest decade; n=2179): 7.1±24.4 vs. -3.9±19.8 mg/dL. Unfractionated AUC glucose was not corelated with BMI (r=0.00). Like age, DEP and BMI were inversely correlated (r=-0.45), whereas INDEP and BMI were positively correlated (r=0.20). INDEP was higher in those with BMI>25.0 (overweight surrogate; n=1157) than in those <15.0 (underweight surrogate; n=581): 9.0±23.2 mg/dL vs. -4.3±19.6 mg/dL. Conclusion: Positive correlations of INDEP (opposite to inverse correlations of DEP) with age and BMI contribute to the lack of association of unfractionated AUC glucose with those risk factors in Ab+ individuals. Since glucose fractions characterize heterogeneity of the relation between glucose and insulin secretion, the partitioning of glucose should be considered in studies of Ab+ individuals. B.M. Nathan: None. D.D. Cuthbertson: None. H.M. Ismail: Consultant; Rise Therapeutics. L.M. Jacobsen: Advisory Panel; Sanofi. M.J. Redondo: None. E.K. Sims: Consultant; Sanofi. Speaker's Bureau; Med Learning Group. Other Relationship; American Diabetes Association. J. Sosenko: None. National Institutes of Health

Objective: We assessed whether there is an impactful glucose fraction independent of insulin secretion in autoantibody-positive individuals. Research Design and Methods: Baseline 2-h oral glucose tolerance test data from the TrialNet Pathway to Prevention (TNPTP; n = 6190) and Diabetes Prevention Trial-Type 1 (DPT-1; n = 705) studies were used. Linear regression of area under the curve (AUC) glucose versus Index60 was performed to identify two fractions: dependent (dAUCGLU) or independent (iAUCGLU) of insulin secretion. Results: The lack of correlation (r = 0.06) of iAUCGLU and the inverse correlation of dAUCGLU (r = -0.59) with the first-phase insulin response from DPT-1 were consistent with the independent and dependent designations of the glucose fractions. Correlations of AUC C-peptide were inverse with dAUCGLU and positive with iAUCGLU (TNPTP: r = -0.72, r = 0.57; DPT-1: r = -0.56, r = 0.60). The explained variance of AUC C-peptide increased markedly after separating AUC glucose into its fractions (from 4% to 85% in TNPTP; from 1% to 67% in DPT-1). The independent fraction contributed more to the increased glycemia of impaired glucose tolerance (IGT) than did the dependent fraction. Both dAUCGLU and iAUCGLU predicted IGT and type 1 diabetes (T1D) (P < 0.0001 for all). However, whereas dAUCGLU was more predictive of T1D (chi-square: 849 vs. 249), iAUCGLU was more predictive of IGT (chi-square: 451 vs. 176). Conclusions: A glucose fraction independent of insulin secretion was identified that was appreciable in autoantibody-positive individuals. It provides insight into the relation between glucose and C-peptide, contributes substantially to the glycemia of IGT, and predicts both T1D and IGT, particularly the latter.

Introduction and Objective: OGTT-derived AUC glucose (AUCglu) can be separated into fractions, dependent (DEP) or independent (INDEP) from insulin secretion. We hypothesized that removal of INDEP from AUCglu adds specificity for insulin secretion thereby improving the detection of treatment effects in T1D prevention trials. Methods: Index60, a composite measure of glucose and C-peptide, was a surrogate for insulin secretion to separate AUCglu into DEP and INDEP fractions. DEP was defined by the intercept and slope of a simple linear regression equation for AUC glucose vs. Index60 (DEP = 130.74+11.08* Index60). INDEP was calculated by subtracting DEP from AUCglu. Analyses were performed in participants from the positive teplizumab and the negative abatacept Trialnet prevention trials. Results: Changes in AUCglu and glucose fractions from baseline to 1 year in treatment and placebo groups are shown (Table). In the teplizumab trial, the difference between groups for DEP as the endpoint was more significant than for AUCglu. In the abatacept trial, the difference was significant for DEP, but not for AUCglu as the endpoint. The difference for INDEP was of borderline significance in the teplizumab analysis. Conclusion: Removing INDEP from AUCglu can improve specificity for insulin secretion and thus increase the likelihood of detecting a treatment effect using glucose as a metabolic endpoint in T1D prevention trials. B.M. Nathan: None. D.D. Cuthbertson: None. H.M. Ismail: Consultant; Rise Therapeutics. L.M. Jacobsen: Advisory Panel; Sanofi. M.J. Redondo: None. E.K. Sims: Consultant; Sanofi. Speaker's Bureau; Med Learning Group. Other Relationship; American Diabetes Association. W.E. Russell: None. K.C. Herold: Consultant; Sanofi, Dompé, Vertex Pharmaceuticals Incorporated, Sonoma, NexImmune. J. Sosenko: None. National Institutes of Health

Since little is known about the disposition index (DI) in autoantibody-positive individuals, we have assessed whether DI has a similar association between insulin secretion and resistance to the association observed in other populations. In TrialNet Pathway to Prevention (TNPTP; n=6620) and Diabetes Prevention Trial-Type 1 (DPT-1; n=704) study participants, two secretion-sensitivity pairs each representing a DI were analyzed cross-sectionally at baseline: AUC C-peptide/AUC glucose (AUC Ratio) and Matsuda Index (MI) from TNPTP OGTTs (oral DI), first-phase insulin response (FPIR) and 1/fasting insulin (1/FI) from DPT-1 from IVGTTs (DI). Participants were followed for progression to type 1 diabetes. Within the normal and diabetes glucose ranges, associations of AUC ratio with MI in TNPTP, and FPIR with 1/FI in DPT-1, had inverse curvilinear patterns with convexities to the origin. After logarithmic transformations to linearize the secretion and sensitivity measures, the inverse slope was steeper for the diabetes range (p<0.0001). In a Cox regression model including the AUC Ratio and MI as variables and another model including FPIR and 1/FI, the interaction terms of secretion x sensitivity (i.e., the DI/ODI), predicted stage 3 type 1 diabetes in both (p<0.0001). The DI remained significantly predictive (p<0.0001) when the DPT-1 risk score was added as a covariate in regression models. In autoantibody-positive populations, insulin secretion is inversely related to sensitivity in a quasi-hyperbolic relationship in normal and diabetes ranges of glucose. The DI can be represented by a statistical and physiologic interaction between secretion and sensitivity that is predictive of stage 3 type 1 diabetes.

Introduction and Objective: IGT (2hGLU 140-199 mg/dL) has been used to indicate higher risk of progression to T1D. However, prior findings suggest there are fluctuations between IGT and normoglycemia. We have thus systematically examined the progression to IGT and its reversion to normoglycemia. Methods: We identified 335 Ab+ TrialNet Pathway to Prevention participants (mean±SD age 19.1±13.9 yrs, 54% female) with normal 2-hr OGTTs at baseline and 6 months before incident IGT. Results: The table shows changes in mean±SD 2hGLU values between the timepoints along with concurrent changes in Index60 (a composite measure of glucose and C-peptide). There was little change in 2hGLU from baseline to 6 months before incident IGT, with a marked increase from 6 months to IGT. Changes in Index60 followed a similar pattern. Within 1 year after incident IGT, a reversion to normoglycemia was considerably higher for non-progressors to T1D (84%) than for progressors (18%). Conclusion: There is a gradual increase in 2hGLU until 6 months before incident IGT after which there is a marked increase. This is largely attributable to a marked decline in insulin secretion, suggested by a parallel increase in Index60. Reversion to normoglycemia is much less frequent with an approaching diagnosis. These findings may provide insight into β-cell function during progression. H.M. Ismail: Consultant; Rise Therapeutics. D.D. Cuthbertson: None. B.M. Nathan: None. E.K. Sims: Consultant; Sanofi. Speaker's Bureau; Med Learning Group. Other Relationship; American Diabetes Association. L.M. Jacobsen: Advisory Panel; Sanofi. M.J. Redondo: None. J. Sosenko: None. NIDDK (K23DK129799)

Introduction and Objective: We asked if a binary endpoint for change (∆) from baseline to a fixed timepoint of 1 year could be useful in future trials. Methods: We used 2hr-OGTT data from the negative abatacept prevention trial and the positive teplizumab prevention trial, and from participants in the observational TrialNet Pathway to Prevention Study (PTP) with similar characteristics. Glucose and C-peptide response curves were plotted and vectors for curve movement from baseline to 1 year were used to categorize simultaneous glucose ∆ and C-peptide ∆ as metabolic treatment failure vs. success. Results: PTP participants with ∆glucose>0 and ∆C-peptide<0 from baseline to 1 year were at substantially higher risk for stage 3 T1D than those with ∆glucose<0 and ∆C-peptide>0 (p<0.0001). Based on this, we compared placebo vs. treatment groups in both trials for failure (∆glucose>0 with ∆C-peptide<0) vs. success (∆glucose<0 with ∆C-peptide>0) after 1 year. In the table, the failure vs. success endpoint at 1 year revealed more treatment efficacy than the original endpoints used for each trial, which required 8.0 years to implement for abatacept and 10.5 years to implement for teplizumab. Conclusion: An analytic approach using a binary metabolic endpoint of failure vs. success at a fixed time interval appears to detect treatment effects at least as well as standard primary endpoints with shorter follow-up. E.K. Sims: Consultant; Sanofi. Speaker's Bureau; Med Learning Group. Other Relationship; American Diabetes Association. W.E. Russell: None. K.C. Herold: Consultant; Sanofi, Dompé, Vertex Pharmaceuticals Incorporated, Sonoma, NexImmune. D.D. Cuthbertson: None. H.M. Ismail: Consultant; Rise Therapeutics. L.M. Jacobsen: Advisory Panel; Sanofi. B.M. Nathan: None. M.J. Redondo: None. J. Sosenko: None.

Background: Staging preclinical type 1 diabetes (T1D) and monitoring the response to disease-modifying treatments rely on the oral glucose tolerance test (OGTT). However, it is unknown whether OGTT-derived measures of beta cell function can detect subtle changes in metabolic phenotype, thus limiting their usability as endpoints in prevention trials. Objective : To describe the metabolic phenotype of people with Stage 1 and Stage 2 T1D using metabolic modelling of beta cell function. Methods: We characterized the metabolic phenotype of individuals with islet autoimmunity in the absence (Stage 1) or presence (Stage 2) of dysglycemia. Participants were screened at a TrialNet site and underwent a 5-point, 2-hour OGTT. Standard measures of insulin secretion (area under the curve, C-peptide, Homeostatic Model Assessment [HOMA] 2-B) and sensitivity (HOMA Insulin Resistance, HOMA2-S, Matsuda Index) and oral minimal model-derived insulin secretion (phi total), sensitivity (sensitivity index), and clearance were adopted to characterize the cohort. Results: Thirty participants with Stage 1 and 27 with Stage 2T1D were selected. Standard metrics of insulin secretion and sensitivity did not differ between Stage 1 and Stage 2 T1D, while the oral minimal model revealed lower insulin secretion (P < .001) and sensitivity (P = .034) in those with Stage 2 T1D, as well as increased insulin clearance (P = .006). A higher baseline phi total was associated with reduced odds of disease progression, independent of stage (OR 0.92 [0.86, 0.98], P = .016). Conclusion: The oral minimal model describes the differential metabolic phenotype of Stage 1 and Stage 2 T1D and identifies the phi total as a progression predictor. This supports its use as a sensitive tool and endpoint for T1D prevention trials.

Aims/hypothesis:Immunotherapies such as Teplizumab can preserve residual beta cell function in individuals with newly diagnosed type 1 diabetes (T1D), but treatment response is variable. Currently, no biomarker exists to identify individuals most likely to benefit from immunotherapy. We believe that baseline serum metabolomic profiles can distinguish individuals who respond to treatment from nonresponders and predict therapeutic response. Methods:Baseline serum samples from 41 individuals newly diagnosed with T1D enrolled in the AbATE trial (NCT00129259) were analyzed to identify metabolic predictors of response to Teplizumab therapy in the AbATE trial. Responders to Teplizumab, as per study protocol, were defined as individuals who exhibited less than a 40% decline in baseline C-peptide levels at 2 years after start of treatment. We analyzed baseline serum samples using a semi-targeted metabolomics approach via liquid chromatography-high-resolution tandem mass spectrometry. Metabolites that were significantly different between responders and nonresponders were identified (P < 0.05), and the significant metabolites were used to train a supervised Random Forest model to predict treatment response. Model performance was evaluated using a 70/30 training/testing split, 5-fold cross-validation, bootstrap resampling (1,000 iterations), and permutation testing (1,000 permutations). Results:We identified 15 significantly different metabolites at baseline between responders and nonresponders (P < 0.05). These metabolites included amino acids and their derivatives, tricarboxylic acid (TCA) cycle intermediates, and microbially derived metabolites. At baseline, responders exhibited higher levels of TCA cycle metabolites, amino acid derivatives, and microbial metabolites, whereas nonresponders showed elevated levels of glutamate and acylcarnitines. The Random Forest classifier achieved an accuracy of 0.769 and an area under the receiver operating characteristic curve (AUC) of 0.881 in the test dataset. Cross-validation yielded a mean AUC of 0.856 (SD 0.156; 95% CI 0.719-0.992). Bootstrap analysis produced a test AUC 95% CI of 0.619-1.000, and permutation testing confirmed significance (p = 0.012). Conclusions/interpretation:Baseline serum metabolomic signatures can predict responders to Teplizumab with high accuracy. This could potentially be applicable when considering other immunotherapies in preventative efforts in T1D.Trial registration: ClinicalTrials.gov NCT00129259.

Objective:Identify microbial and microbiota-associated metabolites in monozygotic (MZ) and dizygotic (DZ) twins discordant for type 1 diabetes (T1D) to gain insight into potential environmental factors that may influence T1D. Research Design and Methods:Serum samples from 39 twins discordant for T1D were analyzed using a semi-targeted metabolomics approach via liquid chromatography-high-resolution tandem mass spectrometry (LC-HRMS/MS). Statistical analyses identified significant metabolites (p < 0.1) within three groups: All twins (combined group), MZ twins, and DZ twins. Results:Thirteen metabolites were identified as significant. 3-indoxyl sulfate and 5-hydroxyindole were significantly reduced in T1D individuals across all groups. Carnitine was reduced, and threonine, muramic acid, and 2-oxobutyric acid were significantly elevated in both All and MZ groups. Allantoin was significantly reduced and 3-methylhistidine was significantly elevated in All and DZ groups. Conclusions:Metabolite dysregulation associated with gut dysbiosis was observed. However, further validation of our findings in a larger cohort is needed. Article Highlights:Why did we undertake this study? We believed this cohort of twins discordant for type 1 diabetes (T1D) would allow for control over genetic variability to examine environmental factors.What is the specific question(s) we wanted to answer? We aimed to identify differences in microbial and microbiota-associated metabolites in twins discordant for T1D to examine the effect of the gut microbiome on T1D.What did we find? Thirteen metabolites were identified as significantly different.What are the implications of our findings? Our results show the dysregulation of several microbial metabolites in twin pairs, suggesting that the gut microbiome plays a role in the pathogenesis of T1D.

Introduction and Objective: In TrialNet’s Pathway to Prevention (PTP) study, SAB positive (SAB+) individuals develop stage 3 T1D at a rate of ~3%/year. Limited data on SAB reversion to negative (SAB-) in a non-birth cohort exists. Here, we aimed to identify the features and risk of progression to clinical T1D of SAB who revert. Methods: We grouped PTP participants with confirmed SAB+ (i.e., those who tested positive for the same antibody [Ab] on 2 consecutive visits) into reverters (tested negative on 2 subsequent, consecutive visits) or maintainers (had a subsequent SAB+ for the same SAB+). Glucose tolerance data were available for 84% of SAB+ and 81% of SAB-. Comparisons used Chi-square and T-tests. Results: Compared to SAB+ maintainers, SAB- reverters were younger (16.9±13.8 vs. 21.1±13.9, p<0.001); more likely to be Hispanic/Latino (16.7 vs 11.2%, p=0.012); less likely to be GADA+ (55.7 vs 72.9%, p<0.001), and more likely to be IA2A+ (7.6 vs 2.5%, p<0.001) or MIAA+ (36.7 vs 24.3%, p<0.001), Table 1. On follow-up, SAB- reverters did not progress to stage 3 T1D while 3.7% of SAB+ maintainers did (p=0.003); reverters were also less likely to progress to multiple Ab positivity. Conclusion: SAB- reverters have reduced risk of progression to stage 3 T1D compared to SAB+. Future work should aim to identify immunologic and metabolic protective mechanisms driving reversion. H.M. Ismail: Consultant; Rise Therapeutics. D.D. Cuthbertson: None. B.M. Nathan: None. J. Sosenko: None. K.C. Herold: Consultant; Sanofi, Dompé, Vertex Pharmaceuticals Incorporated, Sonoma, NexImmune. E.K. Sims: Consultant; Sanofi. Speaker's Bureau; Med Learning Group. Other Relationship; American Diabetes Association. I. Libman: None. M.J. Redondo: None. C. Speake: Research Support; COUR Pharmaceuticals. Consultant; GentiBio. Advisory Panel; Sanofi. NIDDK K23DK129799