Looking back at the TEDDY study: lessons and future directions

Abstract

The goal of the TEDDY (The Environmental Determinants of Diabetes in the Young) study is to elucidate factors leading to the initiation of islet autoimmunity (first primary outcome) and those related to progression to type 1 diabetes mellitus (T1DM; second primary outcome). This Review outlines the key findings so far, particularly related to the first primary outcome. The background, history and organization of the study are discussed. Recruitment and follow-up (from age 4 months to 15 years) of 8,667 children showed high retention and compliance. End points of the presence of autoantibodies against insulin, GAD65, IA-2 and ZnT8 revealed the HLA-associated early appearance of insulin autoantibodies (1-3 years of age) and the later appearance of GAD65 autoantibodies. Competing autoantibodies against tissue transglutaminase (marking coeliac disease autoimmunity) also appeared early (2-4 years). Genetic and environmental factors, including enterovirus infection and gastroenteritis, support mechanistic differences underlying one phenotype of autoimmunity against insulin and another against GAD65. Infant growth and both probiotics and high protein intake affect the two phenotypes differently, as do serious life events during pregnancy. As the end of the TEDDY sampling phase is approaching, major omics approaches are in progress to further dissect the mechanisms that might explain the two possible endotypes of T1DM.

Figure 1.

Organizational chart of the TEDDY study Data Coordinating Centre (DCC) and clinical sites and flow of samples and data. The illustration is based on the TEDDY study design and the subsequent design of methods and quality control. Key steps are the contributions by the TEDDY sites in four different countries to 1) first ship all biosamples directly to the NIDDK repository in the US and 2) upload all information and questionnaire data directly to the DCC database. DCC coordinate and add quality control samples to the shipment of re-coded samples from the NIDDK repository to different TEDDY assay laboratories in the US, Finland and UK. The respective laboratories transmit analyzed biosamples to the DCC which curates and disseminate datasets for analysis.

Figure 2.

Incidence rate of insulin (IAA), GAD65 (GADA) or multiple islet autoantibodies as detailed in a recent review. A) IAA first to multiple (gray solid line); GADA first to multiple (blue solid line) and multiple first (red solid line). The incidence of IAA first to multiple is significantly higher in the first year of life as compared to GADA first to multiple (Wald chi-square=23.28, p-value<0.0001). B) incidence of IAA (gray solid line) or GADA (blue solid line) that persist as single autoantibodies.

Figure 3.

Incidence rates /1000 person years comparing a) IAA first with celiac disease autoimmunity (CDA) and b) GADA-first with CDA. The cumulative incidence at 14 years of age in the TEDDY study reached 5.1% for IAA first, 7.1% for GADA-first and 18.9% for CDA. Incidence rates /1000 person years in relation to the HLA-DR-DQ haplogenotypes DR4-DQ8/DR4.X -DQ8, DR3-DQ2/DR4-DQ8 and DR3-DQ2/DR3-DQ2 as recently reviewed . The cumulative incidence by 14 years of age for IAA-first was 5.6%, 6.0% and 2.7%, respectively, and for GADA-first 5.3%, 9.1% and 7.2%, respectively compared to 11.3%, 22.2% and 41.9%, respectively for CDA.

Figure 4.

Summary of significant TEDDY observations suggesting that the phenomena associated with either risk or protection from IAA-first are different from that of GADA-first. IAA, formerly known as MIAA, indicates that the microinsulin autoantibody analysis was used throughout the TEDDY study. Progression from islet autoimmunity (stage 1 and stage 2 type 1 diabetes) to stage 3 type 1 diabetes (T1D) is also indicated. Celiac Disease Autoimmunity (CDA) risk factors are summarized along with risk factors for Celiac Disease (CD) once CDA has developed. Relative risks (HR, RR or OR) are shown with 95^th confidence intervals. The data is categorized by the TEDDY outcomes: IAA only as the first-appearing autoantibody (IAA-first)^,,,,,, GADA only as the first-appearing autoantibody (GADA-first)^,,,,, multiple islet autoantibodies (MIAA)^,, progression from islet autoimmunity (IA) to type 1 diabetes (T1D)^,,–, Celiac Disease autoimmunity (CDA)^,,,, and Celiac Disease (CD)^,. The definition of the MIAA endpoint varies across TEDDY publications as either the appearance of a second autoantibody or 2 or more autoantibodies at first appearance.