Multiple longitudinal studies have shown that the appearance of islet autoantibodies precedes clinical (stage 3) T1D, however the rate of progression from seroconversion to clinical disease remains highly heterogeneous [10]. Known islet autoantibodies include insulin autoantibody (IAA), glutamic acid decarboxylase antibody (GADA), islet-antigen-2 antibody (IA-2A also known as IC-512A) and zinc transporter 8 antibody (ZnT8A). Important information about islet autoimmunity has been obtained from genetically at-risk cohorts followed over time. Studies recruiting from birth include the Diabetes Autoimmunity Study in the Young (DAISY), the Finnish Type 1 Diabetes Prediction and Prevention (DIPP) study, the BABYDIAB study, and The Environmental Determinants of Diabetes in the Young (TEDDY) study. Data from the Diabetes Prevention Trial-Type 1 (DPT-1) and the Type 1 Diabetes TrialNet Pathway to Prevention (PTP) Study also have provided further data by screening first and second-degree family members (age 1–45 years) of probands with T1D. Key features of these cohorts are highlighted in Table 1. All of these studies consistently showed that individuals with multiple autoantibodies have a high risk of progressing to clinical T1D. Additionally, age at seroconversion and specific autoantibodies present can influence the speed of disease progression, making early detection pivotal in identifying those most at risk [10, 11]. Accurate prediction of the most rapidly progressing individuals is important for risk assessment, and ultimately for identification of those more appropriate for disease modifying therapy.

Autoantibody seroconversion in genetically at-risk children typically occurs at a young age, with prominent peaks in early childhood. The TEDDY study found that the seroconversion rate was highest at age 6 months to 3 years, peaking at 9 months. This peak was notably evident for IAA as the initial autoantibody with progression to multiple autoantibodies [12]. The BABYDIAB study found that the highest incidence of autoantibody seroconversion in genetically at-risk children occurs between 9 months and 2 years of age [13]. Similarly, the DIPP study reported that children with HLA-conferred susceptibility to T1D who progressed to clinical T1D before puberty were most likely to seroconvert before age 4 years, with a peak seroconversion rate at age 2 years [14]. This early onset of seroconversion has been shown to correlate with faster progression to T1D, particularly in children who develop multiple islet autoantibodies before the age of 3 years [7].

Median ages for specific autoantibody appearances support the association of earlier seroconversion with a higher risk of progression to clinical T1D. IAA tends to appear within the first two years of life, while GADA appears between ages 3 and 5 years in children with increased genetic risk or a family history of T1D. IA-2A and ZnT8A generally appear after the initial autoantibody seroconversion. The TEDDY study found that the median age at initial IAA-only seroconversion was 1.83 years, while GADA-only seroconversion occurred later, with a median age of 4.28 years. When GADA emerged as a second autoantibody, it appeared around 3.7 months after the first autoantibody; in contrast, IAA as a second autoantibody had a median onset of 5.9 months. The development of IA-2A or ZnT8A as a second autoantibody had a median onset of 13.3 months. The overall median time between the appearance of the first and second autoantibodies was 6.8 months, with no significant differences based on whether GADA or IAA was the first to appear [12].

Risk of T1D progression strongly relates to age. TrialNet found that the most important factor associated with a more rapid rate of progression is age. Children with multiple autoantibodies progressed to clinical T1D more rapidly than adults [15].

Increased number of autoantibodies correlates with higher risk and faster progression to T1D. Pooled data from DAISY, DIPP, and BABYDIAB showed that in genetically at-risk children, progression to T1D within 10 years in those with no autoantibodies was 0.4%, one autoantibody was 14.5%, and multiple autoantibodies was 69.7%. Progression to clinical T1D after multiple autoantibody seroconversion was 43.5%, 69.7%, and 84.2% at 5, 10, and 15 years of follow-up. The lifetime risk of clinical T1D approaches 100% once two or more islet autoantibodies are detected [7].

In contrast, the TEDDY study reported on the 5-year risk of clinical T1D. Within this population, the 5-year risk was 11% with one autoantibody, 36% with two autoantibodies, and 47% with three autoantibodies [16]. In relatives of individuals with T1D, the DPT-1 study found that the 5-year risk of developing clinical T1D was 25% for two autoantibodies, 40% for three autoantibodies, and 50% for four autoantibodies [17]. Of note, these rates were noted to be lower than in the larger pooled cohort described above. The Fr1da study found significantly greater risk associated with children who had 4 autoantibodies rather than 2 autoantibodies (Hazard Ratio (HR) 1.85) [18].

The TrialNet study revealed that more than 85% of autoantibody-positive relatives with impaired glucose tolerance develop clinical T1D within 5 years [15]. A multicenter retrospective study of 3,015 first-degree relatives found that progression risk is strongly influenced by age and the number of autoantibodies. Younger relatives (age < 20 years) with multiple autoantibodies had the highest progression risk, with a 52.9% risk of developing T1D within 5 years and 82.3% risk within 10 years. In contrast, individuals with only one autoantibody or older age (age ≥ 20) exhibited lower progression rates. The 20-year diabetes risk was 91.2% if both high-risk factors (age < 20 and multiple autoantibodies) were present but decreased to 59.9% with one risk factor and 35.7% for relatives age ≥ 20 with only one autoantibody [19].

In addition, studies have shown that once multiple-autoantibody positivity has developed, individuals from the general population have comparable risk of T1D development as those with a family history of T1D, therefore insights from prior genetically selected cohorts may hold relevance to the general population [18].

The sequence and type of autoantibody appearance play a critical role in determining the risk and rate of progression to clinical T1D. IAA is often the first detected autoantibody in young children, with its incidence declining with age. The DAISY study reported that children with persistently positive IAA progressed to T1D more rapidly, with 100% developing T1D by 5.6 years, compared to 63% in those with fluctuating IAA levels over a 10-year follow-up period. Age at first autoantibody appearance and IAA levels were major predictors of T1D diagnosis, while GADA and IA-2A did not have the same predictive strength in this cohort [20]. Children with IAA as the first autoantibody were found to have a higher risk of progression (10-year risk of 71.4%, 15-year risk of 82.7%, 20-year risk of 92.1%) compared to those who are without IAA in the first positive sample (10-year risk of 47.6%, 15-year risk of 53.9%, 20-year risk of 61.6%) [11].

In contrast, GADA is typically the first autoantibody in older youth and adults. In younger individuals, GADA is associated with slower disease progression, whereas in older cohorts, its presence correlates with significantly higher risk of clinical T1D. TrialNet’s PTP study found that single GADA positivity was associated with an increased risk of progression to multiple autoantibodies and clinical T1D, with adults aged 45 years exhibiting a fourfold greater risk of progression than young children. This highlights the age-dependent impact of GADA, where its prevalence and influence on disease progression are more pronounced in adults [21].

The differences in IA in adults extend beyond this finding. While IAA often initiates the cascade towards multiple autoantibody positivity in children, GADA more frequently appears alone and remains stable if it does not progress further. IA-2A and ZnT8A are less common as the initial autoantibodies, but their presence is strongly associated with faster progression to T1D compared to when both autoantibodies are absent. In first-degree relatives of individuals with T1D, IA-2A and/or ZnT8A is associated with a 5-year progression rate to clinical T1D of 45% [22]. IA-2A as a second autoantibody was associated with a significantly greater risk of progression compared to IAA, GADA, or ZnT8A, independent of initial autoantibody type [12].

Several studies have examined the relationship between GADA titers and clinical features of T1D. Higher GADA titers have been associated with lower urine C-peptide levels, greater insulin dependence, and higher frequency of autoimmune thyroid disease [23, 24]. Elevated GADA titers at the time of T1D diagnosis have also been linked to an increased risk of long-term complications such as diabetic retinopathy [25]. In a study of GADA-positive individuals with type 2 diabetes (T2D), those with high GADA titers (> 32 arbitrary units) demonstrated more pronounced features of insulin deficiency, including higher HbA1c, lower BMI, lower total cholesterol, and a lower prevalence of metabolic syndrome compared to those with low GADA titers (≤ 32 arbitrary units). These findings suggest that high GADA titers may reflect a more severe autoimmune phenotype within increased beta cell destruction. Additionally, higher GADA levels were associated with increased IA-2A positivity, further supporting the link between titer levels and autoimmune burden [23]. Recent efforts, such as the Type 1 Diabetes Intelligence (T1DI) study, have advanced the understanding of autoantibodies by attempting to create phenotypic clusters based on their association with T1D risk [11]. In this study, IAA was typically the first autoantibody to appear (median age 1.6 years), followed by GADA (1.9 years), and IA-2A (2.1 years). The highest risk of T1D was observed in children positive for all three autoantibodies, with a 5-year risk of 69.9% and 10-year risk of 89.9%. Children with persistent IAA and GADA had the second-highest risk, with a 5-year risk of 39.1% and a 10-year risk of 73.8%. Notably, this finding suggests that while IA-2A is a strong predictor of T1D in young children, its presence is not essential for disease progression. Clusters characterized by persistent GADA and IA-2A had a 5-year risk of 30.9% and 10-year risk of 68.2%. In this group, IAA often appeared early but reverted during follow-up, with clinical T1D developing later (around age 9 years). However, findings regarding IAA reversion have varied [26]. Late persistent GADA positivity showed lower risk, with a 5-year risk of 10.5% and a 10-year risk of 24.7%. A single, often reverting, antibody was associated with significantly lower risks (5-year risk of 1.6% and 10-year risk of 4%).

Screening for islet autoantibodies has evolved considerably, with various assay techniques now in use. Clinical trials and research protocols often confirm autoantibody positivity with repeat testing to minimize false positives, which is particularly critical in populations with low disease incidence or when considering interventions. Traditional methods, such as radioimmunoassay (RIA), have been integral to autoantibody detection and are considered the “gold standard” [27]. This approach, often used in research settings, involves radiolabeled antigens to detect antibody-antigen complexes. Newer assay technologies include electrochemiluminescence (ECL), enzyme-linked immunosorbent (ELISA), luciferase immunoprecipitation (LIPS), and antibody detection by agglutination PCR (ADAP) [28].

ECL assays are increasingly recognized for their high sensitivity and specificity, and they have also demonstrated improved prediction for T1D progression over RIA. In TrialNet, individuals with autoantibodies detected by RIA underwent further analysis with ECL. Those positive for GADA and/or IAA by ECL had a 34–58% risk of progressing to clinical T1D within six years, compared to a 5% risk for ECL-negative individuals. Most ECL-negative cases were single-autoantibody positive by RIA, underscoring ECL’s enhanced discrimination capacity [29]. ECL assays have also been adapted into multiplex platforms which can detect multiple IAs (IAA, GADA, IA-2A) and tissue transglutaminase antibody (TGA) simultaneously [30]. Other approaches, such as ELISA, offer accessibility and cost-effectiveness, as seen in the Fr1da study, which used a 3-screen ELISA for initial GADA, IA-2A, and ZnT8A detection [18]. Emerging assays like ADAP and LIPS provide novel methodologies. ADAP amplifies DNA-antigen conjugates for highly sensitive detection, while LIPS uses luminescence to quantify serum antibodies. ADAP has been adapted to a 5-plex assay that combines all 4 islet autoantibodies (IAA, GADA, IA-2A, ZnT8A) and TGA [31]. It is being used by the Antibody Detection Israel Research (ADIR) general population screening program and was found to be comparable in performance to standard RIA and ELISA [32]. In addition, LIPS assays for IAA, GAD, IA-2A, and ZnT8A were found to be comparable to RIA [33,34,35]. Within an autoantigen, there are also epitopes that are disease-relevant and improve disease prediction in at-risk individuals. Details on antibody epitopes and further differences in assays has previously been reviewed [36].

Further studies are needed to directly compare these different assays, especially regarding their advantages and disadvantages as we progress towards general population screening. The first international workshop to attempt standardization of insulin autoantibodies was in 1987 [37]. Additional evaluations were done by the Diabetes Autoantibody Standardization Program (DASP) [38], which was then superseded by the Islet Autoantibody Standardization Program (IASP). The IASP works to assess assay proficiency, harmonize interlaboratory measures, evaluate novel assays, and highlights the ongoing heterogeneity and lab-specific dependence of results [27].

Affinity refers to the strength of binding of an antibody to its binding site. Predictive value of autoantibodies increases when measured by high-affinity methods. High affinity of IAA has been associated with progression to multiple autoantibodies and clinical T1D. In the BABYDIAB cohort, high affinity of IAA was associated with HLA DRB1*04, younger age of IAA appearance, and subsequent progression to multiple islet autoantibodies or T1D. IAA affinity in multiple antibody-positive children was on average 100-fold higher than in children who remained single IAA positive or became autoantibody negative [39]. Another study from Germany found that using a threshold of ≥ 109 l/mol, 22 of the 24 children who developed multiple islet autoantibodies or diabetes were correctly identified by high-affinity IAA and 18 of 22 who did not develop multiple islet autoantibodies or diabetes were correctly identified by low-affinity IAA [40].

Similar to IAA, high-affinity GADA was also associated with multiple autoantibodies and the development of clinical T1D. GADA affinity was higher in multiple autoantibody-positive children and in HLA DR3-positive children [41, 42]. Additional studies are needed on autoantibody affinity to further characterize these findings. Interestingly, despite IA-2A’s importance in predicting clinical T1D in other studies, one study of IA-2A affinity was not associated with progression to clinical T1D or HLA haplotype [43].