Kendra Vehik

Prospective evidence linking early-life microbiome development and host genetics with type 1 diabetes (T1D) risk is limited. Here, we describe how gut microbiome maturation and host genetics influence T1D risk in the TEDDY study. We analysed 12,151 longitudinal metagenomes and host genetic data from 887 children at high genetic risk for T1D followed for up to 6 years. We identify three microbiome maturational patterns: Early Matured, Late Matured and Early Plateaued, driven primarily by non-linear changes in species from the Bifidobacterium and Ruminococcus genera. The Early Matured pattern is enriched in galactose metabolism and exhibits higher production of aromatic amino acids and B-group vitamins at early follow-ups, whereas the Early Plateaued pattern has increased microbial production of branched-chain amino acids. Notably, the Early Plateaued pattern is associated with a threefold elevated risk of T1D, whereas other patterns are not associated with T1D risk. Furthermore, we find that host genetic variants related to antimicrobial and antiviral immune responses modify the association between the Late Matured pattern and T1D risk. These findings highlight the role of early microbial exposures and host genetics in T1D susceptibility.

We aimed to determine if thyroid autoimmunity is associated with a child’s risk for subsequent development of islet or celiac autoantibodies. Children at high genetic risk of type 1 diabetes were followed for thyroid autoimmunity (thyroid peroxidase antibodies [TPOAb] and thyroglobulin antibodies [TGAb]), islet autoimmunity (IA), and celiac disease autoimmunity (CDA) in The Environmental Determinants of Diabetes in the Young study. Out of 5482 children tested for thyroid autoimmunity, IA, and CDA, 39% developed at least one autoantibody. At age 14 years, thyroid autoimmunity co-occurred with IA in 59 children (15 more cases than expected by chance alone, P = 0.02) and with CDA in 125 children (26 cases above expected, P = 0.01). The risk of developing IA or CDA after thyroid autoimmunity varied by which thyroid autoantibody appeared first: TPOAb-first was associated with both IA (HR 1.92, 95% CI, 1.09-3.40) and CDA (1.69, 95% CI, 1.03-2.76), whereas TGAb-first was not associated with the risk of either. Islet autoimmunity and CDA are frequently found in connection to thyroid autoimmunity in children and young adolescents. The relationships of thyroid autoimmunity with IA and CDA depend on which thyroid autoantibody appears first.

Summary: This paper presents a joint model of multivariate longitudinal data and multistate data with application to modeling and predicting autoantibody development in The Environmental Determinants of Diabetes in the Young (TEDDY) study. The model quantifies the risks of state transitions based on observed time-varying and non-time-varying risk factors. Based on the estimated model, a dynamic prediction approach is suggested to predict future state occupation probabilities using historical data. The proposed method can handle uncertainties in the observed data, due to measurement errors in the observed longitudinal data and interval censoring or missing information in the observed multistate data. For evaluating the predictions by the proposed approach, some performance metrics and their estimation are discussed. The proposed method is evaluated by some simulation studies. It is discussed in detail how this method can be used in analyzing the TEDDY data by properly handling the missing information and predicting future disease status using the proposed dynamic prediction algorithm.

Model personalization has attracted widespread attention in recent years. In an ideal situation, if individuals’ data are sufficient, model personalization can be realized by building models separately for different individuals using their own data. But, in reality, individuals often have data sets of varying sizes and qualities. To overcome this disparity, collaborative learning has emerged as a generic strategy for model personalization, but there is no mechanism to ensure fairness in this framework. In this paper, we develop fair collaborative learning (FairCL) that could potentially integrate a variety of fairness concepts. We further focus on two specific fairness metrics, the bounded individual loss and individual fairness, and develop a self-adaptive algorithm for FairCL and conduct both simulated and real-world case studies. Our study reveals that model fairness and accuracy could be improved simultaneously in the context of model personalization. History: Bianca Maria Colosimo served as the senior editor for this article. Funding: This work was supported by the Breakthrough T1D Award [Grant 2-SRA-2022-1259-S-B]. Data Ethics & Reproducibility Note: The code capsule is available on Code Ocean at https://codeocean.com/capsule/1331847/tree/v1 and in the e-Companion to this article (available at https://doi.org/10.1287/ijds.2024.0029 ). The real-world data, including the transportation demand management and surgical site infection data sets, are proprietary and not publicly available. Other results are available at https://github.com/ryanlif/FairCL .

Context: Autoantibodies to thyroid peroxidase (TPOAb) and thyroglobulin (TgAb) define preclinical autoimmune thyroid disease (AITD), which can progress to either clinical hypothyroidism or hyperthyroidism. Objective: We determined the age at seroconversion in children genetically at risk for type 1 diabetes. Methods: TPOAb and TgAb seropositivity were determined in 5066 healthy children with human leukocyte antigen (HLA) DR3- or DR4-containing haplogenotypes from The Environmental Determinants of Diabetes in the Young (TEDDY) study. Children seropositive on the cross-sectional initial screen at age 8 to 13 years had longitudinally collected samples (from age 3.5 months) screened retrospectively and prospectively for thyroid autoantibodies to identify age at seroconversion. The first-appearing autoantibody was related to sex, HLA genotype, family history of AITD, and subsequent thyroid dysfunction and disease. Results: The youngest appearance of TPOAb and TgAb was age 10 and 15 months, respectively. Girls had higher incidence rates of both autoantibodies. Family history of AITD was associated with a higher risk of TPOAb hazard ratio (HR) 1.90; 95% CI, 1.17-3.08; and TgAb HR 2.55; 95% CI, 1.91-3.41. The risk of progressing to hypothyroidism or hyperthyroidism was not different between TgAb and TPOAb, but children with both autoantibodies appearing at the same visit had a higher risk compared to TPOAb appearing first (HR 6.34; 95% CI, 2.72-14.76). Conclusion: Thyroid autoantibodies may appear during the first years of life, especially in girls, and in children with a family history of AITD. Simultaneous appearance of both autoantibodies increases the risk for hypothyroidism or hyperthyroidism.

Aims/hypothesis: Delivery by Caesarean section continues to rise globally and has been associated with the risk of developing type 1 diabetes and the rate of progression from pre-symptomatic stage 1 or 2 type 1 diabetes to symptomatic stage 3 disease. The aim of this study was to examine the association between Caesarean delivery and progression to stage 3 type 1 diabetes in children with pre-symptomatic early-stage type 1 diabetes. Methods: Caesarean section was examined in 8135 children from the TEDDY study who had an increased genetic risk for type 1 diabetes and were followed from birth for the development of islet autoantibodies and type 1 diabetes. Results: The likelihood of delivery by Caesarean section was higher in children born to mothers with type 1 diabetes (adjusted OR 4.61, 95% CI 3.60, 5.90, pppp28–33 years: adjusted OR 1.19, 95% CI 1.04, 1.35, p=0.01; age group >33 years: adjusted OR 1.80, 95% CI 1.58, 2.06, pp=0.07). Delivery by Caesarean section was associated with a modestly increased rate of progression to stage 3 type 1 diabetes in children who had developed multiple islet autoantibody-positive pre-symptomatic early-stage type 1 diabetes (adjusted HR 1.36, 95% CI 1.03, 1.79, p=0.02). No interaction was observed between Caesarean section and non-HLA SNPs conferring susceptibility for type 1 diabetes. Conclusions/interpretation: Caesarean section increased the rate of progression to stage 3 type 1 diabetes in children with pre-symptomatic early-stage type 1 diabetes. Data availability: Data from the TEDDY study (https://doi.org/10.58020/y3jk-x087) reported here will be made available for request at the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Central Repository (NIDDK-CR) Resources for Research (R4R) (https://repository.niddk.nih.gov/). Graphical

Context The 2 peaks of type 1 diabetes incidence occur during early childhood and puberty. Objective We sought to better understand the relationship between puberty, islet autoimmunity, and type 1 diabetes. Methods The relationships between puberty, islet autoimmunity, and progression to type 1 diabetes were investigated prospectively in children followed in The Environmental Determinants of Diabetes in the Young (TEDDY) study. Onset of puberty was determined by subject self-assessment of Tanner stages. Associations between speed of pubertal progression, pubertal growth, weight gain, homeostasis model assessment of insulin resistance (HOMA-IR), islet autoimmunity, and progression to type 1 diabetes were assessed. The influence of individual factors was analyzed using Cox proportional hazard ratios. Results Out of 5677 children who were still in the study at age 8 years, 95% reported at least 1 Tanner Stage score and were included in the study. Children at puberty (Tanner Stage ≥2) had a lower risk (HR 0.65, 95% CI 0.45-0.93; P = .019) for incident autoimmunity than prepubertal children (Tanner Stage 1). An increase of body mass index Z-score was associated with a higher risk (HR 2.88, 95% CI 1.61-5.15; P < .001) of incident insulin autoantibodies. In children with multiple autoantibodies, neither HOMA-IR nor rate of progression to Tanner Stage 4 were associated with progression to type 1 diabetes. Conclusion Rapid weight gain during puberty is associated with development of islet autoimmunity. Puberty itself had no significant influence on the appearance of autoantibodies or type 1 diabetes. Further studies are needed to better understand the underlying mechanisms.

Background/Objective. Growth and obesity have been associated with increased risk of islet autoimmunity (IA) and progression to type 1 diabetes. We aimed to estimate the effect of energy-yielding macronutrient intake on the development of IA through BMI. Research Design and Methods. Genetically at-risk children (n = 5,084) in Finland, Germany, Sweden, and the USA, who were autoantibody negative at 2 years of age, were followed to the age of 8 years, with anthropometric measurements and 3-day food records collected biannually. Of these, 495 (9.7%) children developed IA. Mediation analysis for time-varying covariates (BMI z-score) and exposure (energy intake) was conducted. Cox proportional hazard method was used in sensitivity analysis. Results. We found an indirect effect of total energy intake (estimates: indirect effect 0.13 [0.05, 0.21]) and energy from protein (estimates: indirect effect 0.06 [0.02, 0.11]), fat (estimates: indirect effect 0.03 [0.01, 0.05]), and carbohydrates (estimates: indirect effect 0.02 [0.00, 0.04]) (kcal/day) on the development of IA. A direct effect was found for protein, expressed both as kcal/day (estimates: direct effect 1.09 [0.35, 1.56]) and energy percentage (estimates: direct effect 72.8 [3.0, 98.0]) and the development of GAD autoantibodies (GADA). In the sensitivity analysis, energy from protein (kcal/day) was associated with increased risk for GADA, hazard ratio 1.24 (95% CI: 1.09, 1.53), . Conclusions. This study confirms that higher total energy intake is associated with higher BMI, which leads to higher risk of the development of IA. A diet with larger proportion of energy from protein has a direct effect on the development of GADA.