POLSKI
Introduction
Type 1 diabetes (T1D) is a chronic autoimmune condition with around half of the cases diagnosed in childhood and adolescence [1]. While genetic predisposition plays a central role in disease susceptibility, environmental triggers are believed to initiate or accelerate the autoimmune process. These triggers, including viral infections such as enteroviruses and respiratory pathogens, have long been implicated in T1D pathogenesis in genetically susceptible individuals [2–4].
The emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) introduced a novel environmental context in which biological (infection and immune activation) and societal stressors (lockdowns and delayed care), might have affected the onset and presentation of T1D. In parallel, the rapid global implementation of coronavirus disease (COVID-19) vaccination programs has raised concerns about possible immunologic effects, including the theoretical risk of triggering autoimmune diabetes in susceptible children [5–7]. As with earlier vaccine-related concerns, such as those raised around MMR and type 1 diabetes, these hypotheses must be carefully examined in light of the overall benefits of immunization [8].
Several observational studies have reported increases in new-onset T1D following SARS-CoV-2 infection, with some reporting hazard ratios above 1.8 for new diagnoses within six months of confirmed infection [7, 9]. Other studies reported an increased severity of the T1D onset during the timing of the COVID-19 [10]. Others, however, found no significant change or even a decline in incidence, highlighting the need for comprehensive evaluation. Case reports and pharmacovigilance data have suggested a temporal relationship between COVID-19 vaccination and T1D onset, but these remain anecdotal, with no demonstrated causality [11, 12]. Furthermore, other studies have shown that increased incidence during the pandemic may not be directly related to infection or vaccination, and that severity at presentation did not worsen [13, 14].
While recent literature has explored the relationship between COVID-19 and T1D, no comprehensive synthesis has specifically addressed the dual role of both SARS-CoV-2 infection and COVID-19 vaccination as potential environmental triggers for new-onset T1D in children and adolescents [13, 15–17]. Although some authors provided a narrative overview of increased T1D incidence during the pandemic, they lacked systematic methodology and did not analyze immune or vaccine-related mechanisms. Others conducted a meta-analysis focused exclusively on glycemic fluctuations after COVID-19 vaccination in individuals with established T1D, without assessing disease onset or pediatric subgroups. Reviews that addressed autoimmune diabetes post-infection or post-vaccination in broader terms often combined adult and pediatric populations and did not apply structured synthesis methods or risk-bias appraisal. In contrast, our review is registered in PROSPERO and follows PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines to systematically evaluate both infection and vaccination as potential environmental triggers for T1D, with a dedicated focus on children and adolescents [18]. We aim to extract not only incidence trends but also timing of onset, immunologic markers, and severity at diagnosis. Furthermore, we plan to assess methodological quality, extract data on confounding variables (stress, comorbidities), and explore differences across regions, exposure types, and population characteristics.
Therefore, this systematic review with meta-analysis aims to synthesize and map the available evidence on the relationship between COVID-19 infection and/or vaccination and the incidence of new-onset T1D in children and adolescents.
Material and methods
Review design
This systematic review and meta-analysis will follow the PRISMA checklist. The protocol is registered in PROSPERO (ID: CRD420251051150). If data availability does not permit a quantitative synthesis (meta-analysis), we will conduct a structured narrative synthesis. Certainty of the evidence will be assessed using the GRADE approach.
Search strategy
We will search the following databases for literature published from January 2020 onward: PubMed/MEDLINE, Embase, Scopus, Web of Science, and the Cochrane Library. Search terms will include combinations of type 1 diabetes, T1D, COVID-19, SARS-CoV-2, vaccination, infection, children, and adolescents. Both MeSH/Emtree terms and free-text keywords will be used. The complete strategy is available in PROSPERO. The detailed search strategy is provided as supplementary material. No language restrictions will be applied. Grey literature (conference abstracts, preprints) will be excluded unless they provide sufficient methodological detail and outcome data. Additionally, we will manually screen the reference lists of included studies and relevant reviews.
Eligibility criteria
The review will include observational studies (cohort, case-control, cross-sectional) evaluating new-onset T1D in individuals aged 0–19 years following confirmed SARS-CoV-2 infection or COVID-19 vaccination that meet predefined criteria based on the Population–Exposure–Comparator–Outcome–Study design (PECOS) framework. The aim is to comprehensively identify and evaluate the evidence linking COVID-19 infection and/or vaccination with the onset of T1D in pediatric populations.
Studies must report original data, describe diagnostic criteria, and provide extractable outcome data. Reviews and meta-analyses will be excluded but screened for eligible references.
Population
We will include studies involving children and adolescents aged 0 to 19 years who have been diagnosed with new-onset T1D, based on clinical and/or biochemical criteria. Studies must clearly indicate that the diagnosis corresponds to T1D, either through clinical presentation (hyperglycemia, ketoacidosis), requirement of insulin therapy, or laboratory confirmation (low or absent C-peptide, presence of diabetes-related autoantibodies). Studies involving adults will be excluded, unless pediatric-specific data are reported separately or can be extracted independently.
Exposure
Eligible studies must assess one or both of the following exposures: 1) Confirmed SARS-CoV-2 infection, as diagnosed by reverse transcription polymerase chain reaction (RT-PCR), antigen testing, or serological evidence (positive SARS-CoV-2 IgG antibodies), in the context of recent infection or post-infectious follow-up; 2) COVID-19 vaccination, regardless of manufacturer, vaccine type (mRNA, viral vector), dose, or timing relative to T1D diagnosis. Case reports of new-onset T1D temporally associated with vaccination may be included to allow for narrative synthesis of early signals. Studies evaluating general incidence trends during the pandemic will also be considered, provided they present data that can be interpreted in relation to population-level COVID-19 exposure (infection rates or vaccination coverage).
Comparators
Comparators may include individuals without documented SARS-CoV-2 infection or vaccination, historical or pre-pandemic incidence rates of T1D within the same population or region, and individuals exposed to other respiratory viruses (influenza or others) when used as contextual comparators. The absence of a comparator group will be a reason for exclusion to ensure the evidence is suitable for quantitative synthesis. Accordingly, stand-alone case reports and descriptive case series without a comparator will be excluded.
Outcomes
The primary outcome of interest is the incidence of new-onset T1D following exposure to COVID-19 infection or vaccination. This may be expressed as absolute case counts, incidence rates, relative risk measures, other effect sizes (hazard ratios or odds ratios) compared to a defined comparator group, or qualitative increases over baseline levels. Secondary outcomes include: clinical severity at diagnosis, such as the presence and severity of DKA; timing of onset in relation to the exposure event (within 1, 3, 6, or 12 months post-infection or vaccination); biochemical and immunologic profiles, including the presence of autoantibodies – GADA (glutamic acid decarboxylase antibodies, IA-2A (insulinoma-associated autoantigen), ZnT8A (zinc transporter family member 8), IAA (insulin autoantibodies), C-peptide levels, and basal HbA1c; genetic or immunologic susceptibility markers, including HLA genotypes, T-cell markers, or immune cell subsets if available. In addition, for each study and the overall synthesis, we will classify the direction of effect as (↑) increase, (↓) decrease, mixed, or no difference (=) No difference, to summarize whether the exposure was associated with a higher, lower, variable, or unchanged risk of new-onset T1D.
Study designs
We will include observational analytical studies, such as cohort studies (retrospective or prospective), registry-based surveillance reports, provided they include extractable data on new-onset T1D in the pediatric age group, case-control studies, and cross-sectional analyses. Systematic reviews and meta-analyses will not be included in the synthesis but will be screened for potentially eligible primary studies through reference checking.
Study selection and screening
Search results will be imported into Rayyan for collaborative blinded screening [19]. Two reviewers will independently screen titles and abstracts for eligibility. Disagreements will be resolved by consensus or adjudicated by a third reviewer. Full texts of potentially eligible studies will be retrieved and screened similarly.
Data extraction
To ensure a systematic and transparent synthesis, we will extract standardized information for each included study based on a structured evidence framework (Table I). This table aligns with the PECOs approach, capturing key study characteristics (author, year, country, design, sample size, age group), exposure type (COVID-19 infection, vaccination, or both), and detailed vaccine information when applicable (type, timing, and temporal relation to T1D onset). The extraction will include clearly defined comparator groups, primary outcomes (incidence of new-onset T1D, DKA, immune profile), direction of effect (↑ increase, ↓ decrease, mixed, or = no difference), effect sizes and metrics (hazard ratios, odds ratios, risk ratios), and the explicit timing of exposure relative to diagnosis. In addition, immune findings (such as autoantibodies), genetic predisposition (HLA status), and any relevant confounders will be recorded. For each study, we will also summarize authors’ conclusions regarding potential causality, assign an overall conclusion strength (strong, suggestive, inconclusive, or null), assess the risk of bias, and include reviewer comments where needed. This comprehensive framework will support consistency in data extraction and provide a robust basis for narrative and, where feasible, quantitative synthesis.
Table I
Variables that will be systematically extracted for each included study. This framework aligns with the PECOs approach to ensure transparent synthesis
Risk of bias and quality assessment
We will use the Newcastle-Ottawa Scale for observational studies [20]. Assessment will consider study design, exposure measurement, outcome reporting, and control of confounders.
Data synthesis
A descriptive synthesis will summarize findings across studies. Where data allow, findings will be grouped by exposure type (infection vs. vaccination), age category, region, or time period. If three or more studies report comparable incidence data, a random-effects meta-analysis may be performed using pooled incidence rate ratios or odds ratios. Subgroup analyses may be conducted based on: age (< 6; 6–10; ≥ 10 years), exposure type (infection vs. vaccination), geographic region, COVID-19 wave/timing.
Expected results
Based on previous data, we expect to find evidence of an increased incidence of new-onset T1D following SARS-CoV-2 infection in pediatric populations, particularly during the first waves of the pandemic, when hazard ratios and risk ratios were often reported above 1.8 [11]. Large population-based cohorts and registry studies comparing pre- and post-pandemic incidence are likely to provide the strongest comparative estimates against pre-pandemic or uninfected reference groups. These associations may be reinforced by regional observational cohorts reporting increased referrals and higher rates of DKA at presentation [10].
In contrast, the relationship between COVID-19 vaccination and T1D onset is anticipated to be much weaker, with most available data limited to descriptive signals or smaller comparative studies. Some authors suggest no association or even a lower risk of severe metabolic decompensation among vaccinated children compared to unvaccinated or infected peers [21, 22].
However, not all studies support a direct causal relationship. Some large registries like the German DPV and the SWEET, and national cohorts (Denmark, Spain) have reported no direct causal relationship, highlighting that increases in T1D incidence during the pandemic may be partly attributable to indirect factors such as diagnostic delays, stress, or shifts in healthcare-seeking behavior rather than direct viral effects [13, 23–25]. We expect to observe mixed directions of effect across studies, as well as limited data on immunologic findings, autoantibody trends, or genetic markers.
Despite emerging patterns, the evidence may be limited by considerable heterogeneity in study design, population sampling, diagnostic criteria, exposure definitions (PCR vs. antibody-confirmed infection), and outcome measurement. Most studies may lack adjustment for critical confounders such as healthcare-seeking behavior, diagnostic delays, stress, comorbidities, or viral coinfections. Furthermore, the majority of reports fail to systematically investigate autoimmunity markers or genetic susceptibility.
We anticipate this systematic review and meta-analysis will clarify the differential impact of infection versus vaccination on new-onset T1D, map the direction of effect for each exposure-outcome pair, and highlight knowledge gaps related to confounding, immune profiling, and standardization of definitions, while also providing essential context for interpreting new real-world data, helping determine whether patterns seen during the pandemic persist or evolve in the post-pandemic era.
