Skip to content

Type 1 diabetes — genetics and environmental triggers

TL;DR — T1D is strongly heritable but not genetically determined. HLA class II variation supplies the largest inherited component, while many non-HLA loci collectively modify immune regulation and β-cell biology; a standardized genetic risk score can distinguish T1D from type 2 diabetes and enrich newborn screening (Sharp 2019, PMID 30655379). Genetics cannot explain changing incidence or discordance between identical twins, so environmental exposures must interact with susceptibility. Enterovirus evidence is the most developed trigger line, with a meta-analysis finding molecular detection associated with islet autoimmunity/T1D, but association is not proof that one virus initiates most cases (Isaacs 2023, PMID 37390839). TEDDY has identified temporal infection and microbiome associations, while nutritional prevention has not yielded a proven general strategy (Lönnrot 2017, PMID 28770319; Vatanen 2018, PMID 30356183).

Genetic architecture

Domain Examples Main implication
HLA class II DR3-DQ2, DR4-DQ8, protective haplotypes Antigen presentation dominates inherited risk
HLA class I Selected A and B alleles CD8 T-cell recognition and age/tempo effects
Immune regulation PTPN22, CTLA4, IL2RA Activation thresholds and regulatory pathways
Insulin locus INS variable-number tandem repeat region Thymic insulin expression and tolerance
Antiviral signaling IFIH1 Host response to viral RNA
β-cell/immune shared pathways Multiple common variants Small effects aggregate; mechanism often uncertain

No common variant is diagnostic by itself. HLA risk is large relative to other loci, but even the highest-risk genotype has incomplete penetrance. Conversely, many people who develop T1D do not carry the single highest-risk HLA combination.

Genetic risk scores

The T1D genetic risk score combines HLA and non-HLA variants into a continuous measure. An improved standardized score was developed for newborn screening and incident diagnosis, demonstrating that polygenic information can enrich a population for follow-up and help distinguish autoimmune from type 2 diabetes (Sharp 2019, PMID 30655379).

Use case Potential value Limitation
Newborn enrichment Concentrates autoantibody surveillance in a smaller group Misses cases outside selected thresholds
Adult diagnostic classification Supports T1D vs type 2 distinction Accuracy depends on ancestry and comparator population
Trial recruitment Raises event rate Reduces generalizability if over-enriched
Family counseling Communicates relative susceptibility Does not provide an individual diagnosis date

Portability is a central issue. Scores trained largely in European-ancestry cohorts can lose calibration elsewhere. A 2025 Chinese study demonstrates that a T1D score can discriminate types in a Chinese population, but ancestry-specific validation remains necessary (Hu 2025, PMID 40569436).

Family history and the population paradox

Relatives of people with T1D have substantially higher risk than the general population, which made TrialNet-style family screening efficient. Yet most new cases arise in people without an affected first-degree relative because the general population is much larger. That arithmetic motivates population screening even though individual baseline risk is low (Bonifacio 2025, PMID 40134221).

Genetic enrichment and autoantibody screening answer different questions. A genetic score estimates inherited susceptibility and is stable across life; autoantibodies show that immune activation has occurred and are much closer to disease progression.

Enteroviruses

The 2023 systematic review and meta-analysis of controlled observational studies evaluated enteroviral nucleic acids and proteins. Detection was associated with islet autoimmunity or T1D, supporting biological relevance but leaving heterogeneity in specimen, assay, timing, and viral type (Isaacs 2023, PMID 37390839).

Possible causal roles include:

  1. initiating islet autoimmunity through infection and inflammatory antigen presentation;
  2. accelerating an already established autoimmune response;
  3. persistently infecting pancreatic tissue;
  4. serving as a marker of altered host antiviral immunity rather than a direct cause.

Maternal viral infection studies add a prenatal hypothesis, but observational evidence cannot exclude confounding or exposure misclassification (Allen 2018, PMID 29569297). Causal confidence would increase if virus-specific vaccination or antiviral treatment reduced seroconversion or progression.

Respiratory infections and timing

TEDDY reported that respiratory infections were temporally associated with initiation of islet autoimmunity (Lönnrot 2017, PMID 28770319). Temporal proximity is stronger than a cross-sectional association but still does not establish specificity: respiratory illnesses are frequent in early childhood, diagnoses are syndromic, and immune activation may reveal rather than initiate an existing process.

Microbiome

TEDDY mapped normal early-childhood microbiome development and separately analyzed children who developed early-onset T1D (Stewart 2018, PMID 30356187; Vatanen 2018, PMID 30356183). Associations involve maturation, diversity, metabolic pathways, geography, diet, infection, and antibiotic exposure.

The central causal problem is bidirectionality. Autoimmunity, diet, illness, and metabolic change can alter the microbiome; the microbiome can also influence immune education and barrier function. No microbiome signature is validated to diagnose, predict, or prevent T1D in routine care.

Nutrition and supplementation

Dietary hypotheses have included breastfeeding, timing and type of complementary foods, cow's-milk proteins, gluten, omega-3 fatty acids, and vitamin D. These exposures are difficult to measure, culturally patterned, and linked to infection, socioeconomic status, and microbiome composition. Observational associations should therefore not be translated into prevention advice without intervention evidence.

The literature retrieved in this build did not establish a nutritional intervention that prevents T1D in the general population. A PubMed search for vitamin D prevention returned predominantly type 2 diabetes trials rather than a definitive T1D prevention RCT; the proposed T1D preventive effect is therefore recorded as unproven rather than cited to mismatched evidence.

Geography, migration, and secular change

Wide geographic variation and incidence changes over time are inconsistent with a genetic-only model. Migration studies can separate inherited ancestry from environment but remain vulnerable to healthcare access, ascertainment, age distribution, and generational change. Global T1D modeling confirms large geographic burden differences while also exposing where primary surveillance is missing (Gregory 2022, PMID 36113507).

A causal-evidence ladder

Evidence type What it contributes Major weakness
Geographic correlation Generates hypotheses Ecological confounding
Case-control exposure study Efficient for rare disease Recall and reverse-causation bias
Prospective birth cohort Exposure precedes outcome Multiple testing; residual confounding
Molecular detection in tissue/blood Biological proximity Persistence and contamination questions
Mendelian/randomized intervention Stronger causal test Often unavailable or difficult
Prevention trial Directly tests modifiability Large, long, expensive

The environmental field is strongest where multiple layers align: genetic antiviral pathways, prospective timing, molecular detection, and a plausible pancreatic mechanism. Even there, population attributable risk and the fraction of cases following that pathway are unresolved.

Gene–environment interaction

Gene–environment interaction is often invoked but rarely estimated with sufficient power. A true interaction means the joint effect differs from what is expected from separate genetic and exposure effects; showing that both are associated is not enough. Prospective cohorts require dense exposure sampling before seroconversion and large multi-ancestry samples to evaluate this properly.

IFIH1 and other antiviral-response loci make viral interactions particularly plausible. HLA may also shape which viral or β-cell peptides are presented. These are mechanistic hypotheses awaiting intervention-level validation (Atkinson 2023, PMID 37478842).

Quantifying score transportability

More than 70 genetic regions are associated with T1D, with HLA class II accounting for over half of estimated heritability in reviewed European-ancestry data (Luckett 2023, PMID 37439792). This concentration makes HLA-frequency and linkage-disequilibrium differences consequential across populations.

Population/application Result Meaning
Pune, India European nine-SNP score T1D-vs-type-2 AUC 0.84 vs 0.87 in Europeans Useful with measurable transport loss
SEARCH multiethnic youth; n=2,045 T1D and type-2 scores independently classified etiologic groups Score distributions differed by ancestry
Indian 67-SNP study AUC 0.83 overall and 0.86 antibody-positive vs 0.92 in Europeans More variants did not erase HLA differences

The Pune study included 262 clinical T1D, 345 type 2 diabetes, and 324 controls; most discrimination came from HLA (Harrison 2020, PMID 32528078). SEARCH found particular utility in autoantibody-negative youth, where genetic probability identified those likely to progress to absolute insulin deficiency (Oram 2022, PMID 35312757). Classification performance does not by itself establish clinical utility for newborn screening.

Viral evidence: magnitude and heterogeneity

The 2011 molecular meta-analysis included 4,448 participants and reported OR 3.7 (95% CI 2.1–6.8) for islet autoimmunity and 9.8 (5.5–17.4) for T1D (Yeung 2011, PMID 21292721). The 2023 update included 60 studies/12,077 participants and estimated OR 2.1 (1.3–3.3) for autoimmunity, 8.0 (4.9–13.0) for T1D, and 16.2 (8.6–30.5) for detection within one month of diagnosis; T1D heterogeneity was I²=85% (Isaacs 2023, PMID 37390839).

Finding Causal reading Alternative reading
Stronger detection near diagnosis Infection precipitates conversion Metabolic/immune change increases detection
Consecutive positivity OR 1.55 (1.09–2.20) Persistence promotes autoimmunity Host susceptibility causes both
Enterovirus-B and T1D OR 12.7 (4.1–39.1) A specific viral group is diabetogenic Heterogeneous assays inflate subgroup effects

Metagenomic virome studies produced smaller associations: each positive stool sample had OR 1.14 (1.00–1.29), consecutive positivity OR 1.55 (1.09–2.20), and enterovirus-B positivity OR 1.20 (1.01–1.42) for islet autoimmunity (Faulkner 2021, PMID 33378601). The gradient from modest prospective to large peri-diagnostic estimates needs an intervention test.

Environmental nulls are informative

Hypothesis Design and estimate Boundary
Childhood vaccines trigger T1D Danish cohort, 681 cases/4.72 million person-years; rate ratios near 1 No support across six vaccine categories
Rotavirus vaccine prevents T1D UK cohort found no reduction over 4–6 years Safe in observed window; longer latency possible
Childhood vitamin D protects DAISY intake HR 1.13 (0.95–1.35); 25(OH)D HR 1.12 (0.88–1.43) Longitudinal null
Infant vitamin D protects Observational pooled OR 0.71 (0.60–0.84) No randomized prevention trial
Early probiotics protect TEDDY HR 0.66 (0.46–0.94); DR3/4 HR 0.40 (0.21–0.74) Observational; no recommendation

The Danish cohort found no association for Hib, DTP/polio, whole-cell pertussis, MMR, or oral polio vaccination, including among children with an affected sibling (Hviid 2004, PMID 15070789). A UK rotavirus-vaccine cohort likewise found no T1D reduction (Inns 2021, PMID 34183004).

Vitamin D illustrates genuine conflict. An observational meta-analysis reported lower T1D risk with infant supplementation but found no randomized trials available in 2008 (Zipitis 2008, PMID 18339654). DAISY’s repeated intake and plasma measurements found no association with autoimmunity or progression (Simpson 2011, PMID 21858504). A 2025 systematic review found no significant reduction in islet autoimmunity (pooled OR 0.91, 95% CI 0.67–1.25) or T1D (OR 0.55, 0.22–1.38) with vitamin supplementation; therefore, the evidence rechecked on 2026-08-30 does not establish vitamin D as prevention (Low 2025, PMID 40270966).

Endotypes rather than one exposure pathway

TEDDY separated IAA-first from GADA-first initiation. Male sex, Finnish site, sibling history, DR4, very-early probiotic exposure, and INS variation predicted IAA-first; weight at 12 months and CLEC16A/ERBB3 variants predicted GADA-first. Predictors of antibody spreading differed again from clinical progression (Krischer 2022, PMID 36150053). Different exposure–genotype combinations may therefore converge on the same clinical stage.

Non-HLA effects can also differ by disease phase: DIPP reanalysis of 976 autoantibody-positive cases and 1,910 controls found loci whose associations varied between seroconversion and progression, supporting endotype- and phase-specific rather than uniform genetic action (Laine 2022, PMID 35812428). DAISY analyses of vitamin-D-pathway variants generated association signals but did not convert supplementation into a proven preventive intervention (Frederiksen 2013, PMID 23979957). Likewise, prospective DIPP dietary analyses of 5,674 genetically susceptible children examine food exposures before autoimmunity, yet residual dietary and socioeconomic confounding remains unavoidable (Mattila 2024, PMID 38142920).

Open questions

  • Would an enterovirus vaccine or targeted antiviral reduce persistent autoimmunity or stage progression? Molecular association is established, preventability is not (Isaacs 2023, PMID 37390839).
  • Which microbiome changes precede autoimmunity independently of geography, diet, antibiotics, and infection? (Stewart 2018, PMID 30356187; Vatanen 2018, PMID 30356183)
  • Can ancestry-calibrated genetic scores support equitable newborn screening without excluding most future cases? (Sharp 2019, PMID 30655379; Hu 2025, PMID 40569436)
  • Why do some genetically high-risk children never seroconvert, while lower-risk people develop adult-onset disease? (Evans-Molina 2025, PMID 40230204)

References

  1. Sharp SA, et al. Development and Standardization of an Improved Type 1 Diabetes Genetic Risk Score. Diabetes Care. 2019;42:200-207. PMID 30655379
  2. Hu J, et al. A type 1 diabetes genetic risk score discriminates between type 1 diabetes and type 2 diabetes in a Chinese population. Diabetologia. 2025;68:1969-1982. PMID 40569436
  3. Isaacs SR, et al. Enteroviruses and risk of islet autoimmunity or type 1 diabetes. Lancet Diabetes Endocrinol. 2023;11:578-592. PMID 37390839
  4. Lönnrot M, et al. Respiratory infections are temporally associated with initiation of type 1 diabetes autoimmunity. Diabetologia. 2017;60:1931-1940. PMID 28770319
  5. Allen DW, et al. Maternal virus infections in pregnancy and type 1 diabetes in their offspring. Rev Med Virol. 2018;28:e1974. PMID 29569297
  6. Stewart CJ, et al. Temporal development of the gut microbiome in early childhood from the TEDDY study. Nature. 2018;562:583-588. PMID 30356187
  7. Vatanen T, et al. The human gut microbiome in early-onset type 1 diabetes from the TEDDY study. Nature. 2018;562:589-594. PMID 30356183
  8. Bonifacio E, et al. Type 1 diabetes risk factors, risk prediction and presymptomatic detection. Diabetes Obes Metab. 2025;27 Suppl 6:28-39. PMID 40134221
  9. Gregory GA, et al. Global incidence, prevalence, and mortality of type 1 diabetes in 2021 with projection to 2040. Lancet Diabetes Endocrinol. 2022;10:741-760. PMID 36113507
  10. Evans-Molina C, et al. Type 1 diabetes presenting in adults: Trends, diagnostic challenges and unique features. Diabetes Obes Metab. 2025;27 Suppl 6:57-68. PMID 40230204
  11. Atkinson MA, et al. The pathogenic “symphony” in type 1 diabetes. Cell Metab. 2023;35:1500-1518. PMID 37478842
  12. Luckett AM, et al. Utility of genetic risk scores in type 1 diabetes. Diabetologia. 2023;66:1589-1600. PMID 37439792
  13. Harrison JW, et al. Type 1 diabetes genetic risk score is discriminative of diabetes in non-Europeans. Sci Rep. 2020;10:9450. PMID 32528078
  14. Oram RA, et al. Utility of Diabetes Type-Specific Genetic Risk Scores for Classification Among Multiethnic Youth. Diabetes Care. 2022;45:1124-1131. PMID 35312757
  15. Yeung WC, et al. Enterovirus infection and type 1 diabetes mellitus: systematic review and meta-analysis. BMJ. 2011;342:d35. PMID 21292721
  16. Faulkner CL, et al. The virome in early life and childhood and development of islet autoimmunity and type 1 diabetes. Rev Med Virol. 2021;31:1-14. PMID 33378601
  17. Hviid A, et al. Childhood vaccination and type 1 diabetes. N Engl J Med. 2004;350:1398-1404. PMID 15070789
  18. Inns T, et al. Paediatric rotavirus vaccination, coeliac disease and type 1 diabetes in children. BMC Med. 2021;19:147. PMID 34183004
  19. Zipitis CS, Akobeng AK. Vitamin D supplementation in early childhood and risk of type 1 diabetes. Arch Dis Child. 2008;93:512-517. PMID 18339654
  20. Simpson M, et al. No association of vitamin D intake or 25-hydroxyvitamin D levels with islet autoimmunity and type 1 diabetes. Diabetologia. 2011;54:2779-2788. PMID 21858504
  21. Uusitalo U, et al. Association of Early Exposure of Probiotics and Islet Autoimmunity in the TEDDY Study. JAMA Pediatr. 2016;170:20-28. PMID 26552054
  22. Krischer JP, et al. Predictors of the Initiation of Islet Autoimmunity and Progression. Diabetes Care. 2022;45:2271-2281. PMID 36150053
  23. Laine AP, et al. Non-HLA polymorphisms in T1D: phase- and endotype-specific effects. Front Immunol. 2022;13:909020. PMID 35812428
  24. Frederiksen BN, et al. Vitamin D metabolism gene polymorphisms, islet autoimmunity, and T1D progression. J Clin Endocrinol Metab. 2013;98:E1845-E1851. PMID 23979957
  25. Mattila M, et al. Fruit, berry, and vegetable consumption and risk of islet autoimmunity and T1D. Am J Clin Nutr. 2024;119:537-545. PMID 38142920
  26. Low ZY, et al. Vitamin supplementation and its effect on incident type 1 diabetes mellitus and islet autoimmunity: a systematic review and meta-analysis. Front Immunol. 2025;16:1505324. PMID 40270966