{
  "abstract": "Introduction This study analyzed the effects of the COVID-19 pandemic and body weight on islet and endocrine autoimmunity in children with type 1 diabetes (T1D).Research design and methods Data from 11 973 children and adolescents aged 0.5 to <18 years with new-onset T1D (2015–2023) from the Diabetes Prospective Follow-up Registry were evaluated. Rates of autoantibodies against beta cells (islet antigen 2 (IA2), zinc transporter 8 (ZnT8), glutamic acid decarboxylase (GAD), insulin), thyroid, transglutaminase (TGA), and adrenals were assessed. Logistic regression models adjusted for age and sex examined associations with COVID-19 and body mass index (BMI).Results 6136 (51%) children were diagnosed with T1D before, and 5837 (49%) after the beginning of the COVID-19 pandemic. Beta-cell autoantibodies were present in 94.3%, thyroid autoantibodies in 7.7%, TGA autoantibodies in 8.3%, and adrenal autoantibodies in 5.6%. During versus before COVID-19, IA2 and GAD autoantibody positivity significantly increased (63.3% vs 60.5%, p=0.002, and 65.9% vs 64.0%, p=0.04, respectively), ZnT8 autoantibodies declined (68.0% vs 71.9%, p=0.002), while insulin autoantibodies remained unchanged (p=0.06). Prevalence of IA2, ZnT8, and insulin, but not GAD autoantibodies, showed positive associations with BMI. Thyroid and TGA autoantibodies were not related, while adrenal autoantibodies were negatively related to the pandemic.Conclusions The COVID-19 pandemic and body weight influenced autoimmunity in children with T1D. The rise in IA2 autoantibody positivity may suggest a faster progression from pre-existing autoimmunity to clinical disease. The pandemic did not appear to trigger associated endocrine autoimmunity.",
  "authors": [
    {
      "affiliations": [
        "Paediatric Endocrinology and Diabetology, Bern University, Berne, Switzerland"
      ],
      "name": "Claudia Boettcher"
    },
    {
      "affiliations": [
        "Institute of Epidemiology and Medical Biometry, ZIBMT, University of Ulm, Ulm, Germany"
      ],
      "name": "Reinhard Holl"
    },
    {
      "affiliations": [
        "University Clinic for Pediatrics and Adolescent Medicine, Medical University Vienna, Vienna, Austria"
      ],
      "name": "Katrin Nagl"
    },
    {
      "affiliations": [
        "Division of Endocrinology and Diabetes, RWTH Aachen University, Aachen, Germany"
      ],
      "name": "Beate Karges"
    },
    {
      "affiliations": [
        "Universitätsklinikum Schleswig-Holstein, Campus Lübeck, Lübeck, Germany"
      ],
      "name": "Simone Von Sengbusch"
    },
    {
      "affiliations": [
        "Heinrich-Heine-Universitat Düsseldorf, Düsseldorf, Germany"
      ],
      "name": "Alena Welters"
    },
    {
      "affiliations": [
        "Technical University of Munich School of Medicine, Munich, Germany"
      ],
      "name": "Katharina Warncke"
    },
    {
      "affiliations": [
        "Universitätsklinikum Carl Gustav Carus, Dresden, Germany"
      ],
      "name": "Monika Flury"
    },
    {
      "affiliations": [
        "Viktoriastift Clinic, Bad Kreuznach, Germany"
      ],
      "name": "Diyah Nahdiyati"
    },
    {
      "affiliations": [
        "Auf der Bult Children’s Hospital, Hanover, Germany"
      ],
      "name": "Thekla von dem Berge"
    },
    {
      "affiliations": [
        "Department of Pediatrics, Adolescent Medicine and Neonatology, University of Freiburg Faculty of Medicine, Freiburg, Germany"
      ],
      "name": "Clemens Kamrath"
    }
  ],
  "full_text": "WHAT IS ALREADY KNOWN ON THIS TOPIC The incidence of type 1 diabetes in children increased during the pandemic. However, the underlying mechanisms are unclear.WHAT THIS STUDY ADDS We could show that both the COVID-19 pandemic and body weight influenced islet autoimmunity in children with newly diagnosed type 1 diabetes. Our results suggest faster progression from isolated islet autoimmunity to overt type 1 diabetes.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings shed new light on the intricate relationship between environmental changes, body weight and autoimmune responses in children. This knowledge could help with future risk stratification for autoimmunity and suggest ways to intervene.Background Type 1 diabetes (T1D) is characterized by the autoimmune destruction of pancreatic beta cells. 1 Almost 20% of all young people with T1D develop another autoimmune disease, such as autoimmune thyroid disease, celiac disease, and, less often, adrenal insufficiency.2The multifactorial pathogenesis of T1D involves genetic predisposition and environmental triggers. Over the past decade, several studies have highlighted the role of viral infections in modulating autoimmune responses.1Notably, the incidence of T1D in children increased significantly during the COVID-19 pandemic.3 4 However, the underlying mechanisms of this increase are still unknown, and there are conflicting findings regarding the influence of COVID-19 and the role of SARS-CoV-2 on the development of childhood T1D.5 6 National registry or cohort studies have not found an increased risk of T1D or islet autoimmunity after SARS-CoV-2 infections or only marginally, for example, in Sweden.7–13 Furthermore, obesity is a risk factor for islet autoimmunity and the development of T1D, and it is associated with earlier onset of T1D in predisposed persons.14 Recent studies suggest that higher body weight, sex, and age may influence islet autoimmunity.15 16 During the COVID-19 pandemic, substantial weight gain across all weight and age groups in children and adolescents has been reported.17–19On the basis of these findings, we hypothesized that both COVID-19 pandemic and body weight may modulate islet and endocrine autoimmunity in children and adolescents with T1D, potentially leading to novel insights and strategies for disease prevention.Methods This analysis used data from the German/Austrian/Swiss/Luxembourgian Diabetes Prospective Follow-up Registry (DPV) of children and adolescents aged 0.5 to <18 years with new-onset T1D between 2015 and 2023. In order to analyze autoimmunity, we only included patients with at least documented glutamic acid decarboxylase (GAD) and islet antigen 2 (IA2) autoantibody test results. To verify the extent to which our analysis is biased by only including patients with documented diabetes-specific autoantibodies, we performed a subanalysis. For this, we included all patients who met the above inclusion criteria, regardless of whether diabetes-specific autoantibodies were documented. This cohort was divided into two groups: one with documented islet autoantibodies (regardless of number) and one with no documented islet autoantibodies.Children and adolescents with a migrant background, defined as the patient or a parent born outside Germany, Austria, Switzerland or Luxembourg, were excluded from the analysis. Informed consent for participation in the DPV registry was obtained from T1D persons or their parents by verbal or written procedure, as approved by each center’s responsible data protection officers.Participants were divided into two groups based on the diagnosis of T1D relative to the onset of the COVID-19 pandemic: the before-pandemic group (2015–2019) and the during-pandemic group (2020–2023). The analysis of beta-cell autoimmunity included autoantibodies to IA2, zinc transporter 8 (ZnT8), GAD, and insulin autoantibodies (IAA), the latter only included when measured within 4 weeks after diagnosis of T1D. Furthermore, thyroid autoantibodies against thyroid peroxidase, thyroglobulin, and the thyroid-stimulating hormone receptor, as well as immunoglobulin A antibodies against tissue transglutaminase (TGA) and adrenal autoantibodies, were evaluated. Body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared) was assessed 3 months after T1D diagnosis. BMI categories were defined as <30, 30–70, and >70 percentiles based on German reference values (German Health Interview and Examination Survey for Children and Adolescents).18 Furthermore, BMI values were transformed to SD scores (BMI-SDS) by applying the Box-Cox transformation method.20Unadjusted outcomes were presented as median with IQR or as percentage (%). Data were compared between the before-pandemic and during-pandemic groups via Wilcoxon’s rank-sum test for continuous outcomes or χ2 test for dichotomous outcomes. The main independent variables were the BMI tertiles (<33, 33–66, and >66 percentiles) and the period of T1D manifestation (categorized as before pandemic vs during pandemic). Logistic regression models were used to assess the impact of the COVID-19 pandemic and BMI on autoimmunity. Age and sex were incorporated into the model as confounding variables. Differences in adjusted least squares means were computed to quantify the magnitude of differences in the prevalence of autoantibody positivity across groups. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline for cohort studies. Median and IQR are provided for the description of continuous variables, frequencies and percentages for the description of categorical variables. A two-sided p value ≤0.05 was considered statistically significant. All analyses were performed using SAS V.9.4 (build TS1M7) on a Windows Server mainframe (SAS Institute).Results We analyzed data from 11 973 children and adolescents (6640 males) from 355 diabetes centers in Germany, Austria, Switzerland and Luxembourg documented in the DPV registry. The study population had a median age at T1D diagnosis of 9.8 years (IQR 6.1; 13.0), a median disease duration of 0.30 years (IQR 0.21; 0.40), and a median BMI-SDS of 0.10 (IQR −0.55; 0.76). 94.3% of our patients had at least one positive beta-cell autoantibody test result ( table 1).Table 1Characteristics of the study populationVariablesAllBefore COVID-19After COVID-19P valuenMedian (IQR) or %nMedian (IQR) or %nMedian (IQR) or %Before versus during pandemicAge (years)11 9739.8 (6.1–13.0)613610.1 (6.3–13.3)58379.5 (5.9–12.8)<0.001Diabetes duration (years)11 9730.30 (0.21–0.40)61360.31 (0.22–0.41)58370.30 (0.20–0.39)<0.001BMI-SDS11 9640.10 (−0.55 to 0.76)61350.10 (−0.55 to 0.76)58290.11 (−0.54 to 0.78)>0.99Males11 97355.5613655.8583755.1>0.99All beta-cell-ab+11 97394.3613693.7583794.80.096IA2-ab+11 97361.7613660.5583763.00.058GAD-ab+11 97364.5613663.9583765.2>0.99Insulin-ab+*901556.4446756.9454855.8>0.99ZnT8-ab+586769.3223371.8363467.80.021Thyroid-ab+10 6527.754088.052447.3>0.99TGA-ab+98828.349538.549298.5>0.99Adrenal-ab+12855.66977.76652.2<0.001*Only during the first 4 weeks after diabetes diagnosis.ab+, autoantibody positivity; BMI, body mass index; GAD, glutamic acid decarboxylase; IA2, islet antigen 2; SDS, SD scores; TGA, transglutaminase.Thyroid autoantibodies were positive in 7.7%, TGA autoantibodies in 8.3%, and adrenal autoantibodies in 5.6% of patients. 6136 (51%) individuals had a T1D diagnosis before and 5837 (49%) after the emergence of COVID-19. During versus before the COVID-19 pandemic, patients were younger at manifestation of T1D (median age 9.5 years vs 10.1 years, p<0.001), but BMI did not differ between the two groups (p>0.99; table 1). Table 1 gives an overview of the characteristics of the study population, and online supplemental table S1 gives additional characteristics of age groups and sex.SP110.1136/bmjdrc-2025-005349.supp1Supplementary dataWithout the inclusion criterion of documented results of islet cell autoantibody measurements, we could evaluate the data of 21 654 children and adolescents in the DPV registry. Of these patients, approximately one-third (n=7202) had no documented islet cell autoantibody measurement results, while two-thirds (n=14 452) had documented results. Online supplemental table S2 shows patients without any documentation of islet cell autoantibody findings versus patients with documentation of such findings, once in the overall group (A), before the COVID-19 pandemic (B), and thereafter (C). Patients without measurements had slightly higher BMI and longer diabetes duration, while age and sex distributions were comparable. Findings of autoantibodies other than those associated with T1D (eg, thyroid, adrenal gland and celiac disease) were also documented significantly less frequently in the group without documented islet cell autoantibody findings. However, the positive rate was comparable to that of the group with documented islet cell autoantibody findings. Due to the low documentation frequency, the significance of this finding is limited.Associations of islet cell autoimmunity The onset of T1D following the beginning of the COVID-19 pandemic was associated with a significantly higher rate of positive IA2 (p=0.002) and GAD autoantibody (p=0.04) test results ( figure 1A, online supplemental table S3). Moreover, a positive association between the positivity rates of IA2 and GAD autoantibodies and the COVID-19 pandemic was observed specifically among males, leading to higher prevalence rates in this group during the pandemic (figure 1B, online supplemental table S3). In contrast, the prevalence of elevated ZnT8 autoantibodies experienced a significant decline during the pandemic (p=0.002; figure 1A, online supplemental table S3), with a more pronounced effect observed in girls, resulting in a notably lower prevalence in this group during the pandemic (p<0.0001; figure 1B, online supplemental table S3).Figure 1Prevalence of beta-cell autoantibodies before versus during the COVID-19 pandemic. Autoantibody positivity is shown for different beta-cell autoantibodies (ab) before versus during the COVID pandemic. (A) Beta-cell autoantibody positivity is shown for the whole group of participants for autoantibodies against IA2, GAD, insulin, and ZnT8. (B) Autoantibody positivity is shown for autoantibodies against IA2, GAD, and ZnT8 selectively for males and females. Autoantibody positivity is shown as percentage±SE. BMI, body mass index; GAD, glutamic acid decarboxylase; IA2, islet antigen 2; ZnT8, zinc transporter 8. *P<0.05; **p<0.01; ***p<0.001.Notably, the prevalence of IA2, insulin, ZnT8, but not GAD autoantibodies, showed significant positive associations with increased weight (figure 2A, online supplemental table S3). For insulin and ZnT8 autoantibodies, this association was more pronounced in boys than in girls (figure 2B,C, online supplemental table S3). The prevalence of all positive islet cell autoantibodies was age dependent (figure 2D, online supplemental table S3). GAD autoantibody test results increased with age (figure 2D, online supplemental table S3), and the prevalence of IA2 and ZnT8 autoantibodies was positively associated with ages above 6 years, while the prevalence of IAA was significantly associated with younger age (p<0.0001) (figure 2A–D, online supplemental table S3).Figure 2Prevalence of beta-cell autoantibodies for categories of BMI and age. Beta-cell autoantibody positivity is shown for BMI tertiles (<33, 33–66 and >66 percentiles) for the whole group of participants (A), only females (B) and only males (C). (D) shows beta-cell autoantibody positivity for age categories (<6, 6 to <12, and 12 to <18 years). Autoantibody positivity is shown as percentage±SE. BMI, body mass index; GAD, glutamic acid decarboxylase; IA2, islet antigen 2; ZnT8, zinc transporter 8. *P<0.05; **p<0.01; ****p<0.0001.Table 2 provides an overview of the association of weight, age, gender, and the COVID-19 pandemic with the prevalence of beta-cell autoimmunity and individual islet autoantibodies in children and adolescents with new-onset T1D.Table 2Overview of the impact of weight, age, gender, and the COVID-19 pandemic on the prevalence of beta-cell autoimmunity and of the individual islet autoantibodies in children and adolescents with new-onset type 1 diabetesPrevalence of autoantibodiesEffect of body weightEffect of COVID-19 pandemicAssociated factors: sex and ageIA2 autoantibodiesPositively associated with higher BMIPositively associated with the COVID pandemic only in malesMales, age group 6 to <12 yearsZnT8 autoantibodiesPositively associated with higher BMI in malesNegatively associated with the COVID pandemic only in femalesMales, aged >6 yearsIAAPositively associated with higher BMI in malesNot significantly influenced by the COVID pandemicYounger ageGAD autoantibodiesNo effect of BMIPositively associated with the COVID pandemic only in malesFemales, older ageBMI, body mass index; GAD, glutamic acid decarboxylase; IA2, islet antigen 2 ; IAA, insulin autoantibodies; ZnT8, zinc transporter 8.Associations of thyroid, celiac and adrenal autoimmunity TGA and thyroid autoantibodies were found to be more prevalent in girls ( figure 3A). The prevalence of TGA autoantibodies was further associated with younger age and lower BMI (figure 3B,C). In contrast, thyroid autoantibodies were more common in older T1D persons (figure 3B). Notably, BMI was not related to the occurrence of thyroid or adrenal autoimmunity (figure 3C). While the prevalence of thyroid and TGA autoantibodies remained unchanged during the pandemic, there was a significant decline in the positivity rate of autoantibodies against the adrenal following the emergence of COVID-19 (p<0.001; figure 1D).Figure 3Prevalence of non-beta-cell autoantibodies for BMI, sex, COVID-19 and age. The percentages of positivity of autoantibodies (ab) against TGA, thyroid and adrenals are shown for (A) sex, (B) age categories (<6, 6 to <12, and 12 to <18 years), (C) BMI tertiles (<33, 33–66 and >66 percentiles) and (D) before versus after the beginning of the COVID-19 pandemic. Autoantibody positivity is shown as percentage±SE. BMI, body mass index; TGA, transglutaminase. *P<0.05; **p<0.01; ****p<0.0001.Discussion This study identified a complex interaction between the COVID-19 pandemic and body weight in relation to autoimmunity among children and adolescents with new-onset T1D. More specifically, autoantibody positivity against IA2, ZnT8, and insulin was directly associated with body weight, whereas IA2 and, to a lesser degree, GAD autoantibody positivity were associated with the COVID-19 pandemic in boys, and ZnT8 autoantibody positivity in girls. Specifically, and in line with other reports, 21 22 older children exhibited higher prevalence rates of IA2, ZnT8, and GAD autoantibodies, whereas younger T1D persons showed a higher prevalence of IAA.These age and sex-related differences are consistent with previous studies suggesting heterogeneous immune responses driven by distinct ‘endotypes’ of T1D.23 Our data also suggest that the COVID-19 pandemic may have served as a unique stimulus for islet autoimmune progression across various endotypes. This could explain the temporary shift of the onset of T1D to a younger age during the pandemic.24 Changes in overall environmental exposures among young children during the pandemic may have contributed to the increase in T1D by shifting the balance of environmental factors from protection to promotion of islet autoimmunity.25 A study based on data from the Swedish Better Diabetes Diagnosis study found that genotypes with a lower risk of T1D were associated with a higher proportion of overweight or obese children with newly diagnosed T1D, suggesting that increased BMI in carriers of these genotypes may be a risk factor for T1D development. The results support the hypothesis that environmental factors increasingly contribute to the incidence of T1D, especially in children with genetic risk profiles previously considered low risk.26 Furthermore, it is established that among single-autoantibody positive participants, the human leucocyte antigen (HLA) genotype, GAD autoantibody positivity and elevated BMI were found to increase the risk of disease progression significantly.21 22 27T1D onset after the beginning of the COVID-19 pandemic was associated with a higher rate of positive results of tests against IA2 and, to a lesser degree, also of GAD autoantibodies in boys, while there was a negative effect of the pandemic on ZnT8 autoantibody positivity in girls, suggesting altered sex-specific beta-cell autoimmunity during the pandemic. Therefore, our data indicate that the immune process that generates IA2 and ZnT8 autoantibodies may be more sensitive to external influences, such as pandemic-related changes. Importantly, IA2 autoantibodies are rarely the first islet autoantibodies to appear in children developing T1D.28 While seroconversion to insulin and GAD autoantibodies is associated with the onset of the autoimmune process in T1D, the appearance of IA2 antibodies is associated with progression to clinically overt diabetes.29–31 In line with our findings, Boboc et al observed that children with positive SARS-CoV-2 serology had a higher percentage of detectable IA2 autoantibodies and were more likely to be positive for multiple islet autoantibodies compared with those without SARS-CoV-2 serology.32Our results of an increase in the prevalence of IA2 but not IAA after the beginning of the COVID-19 pandemic suggest that the rise in childhood T1D incidence after the emergence of COVID-19 may have been more likely due to faster progression from pre-existing islet autoimmunity (stage 1 or 2) to clinically overt T1D (stage 3), rather than new-onset islet autoimmunity. An accelerated progression to clinical disease in stage 1 children after a COVID-19 infection was demonstrated by the Fr1da study group.33 Lahn et al suggested a faster rate of autoimmune beta-cell destruction during the pandemic based on the finding that remission in children and adolescents with newly diagnosed T1D occurred less frequently during the first 5 months after diabetes onset during the COVID-19 pandemic than in the pre-pandemic period.34 In addition, during the pandemic, an increased insulin requirement was reported among patients with newly diagnosed T1D at follow-up, suggesting a more rapid loss of beta cells.35 Furthermore, a more rapid autoimmune process would be in line with reported childhood T1D incidence from registry studies showing a wave-like pattern, with an increase during the first 2 years and a decline to pre-pandemic levels in 2022, and even below in 2023.36 37We also found that the prevalence of IA2, ZnT8 and IAA was associated with increasing weight. This aligns with observations indicating that a higher BMI may contribute to islet cell autoimmunity in children.14–16 Furthermore, the link between the COVID-19 pandemic and increased body weight in the prevalence of islet autoantibodies may explain the increase in the incidence of T1D.37 This indicates that a higher BMI may accelerate the disease process in genetically susceptible individuals by exacerbating the autoimmune process towards the development of clinical T1D.3 4 This relationship highlights BMI as a potentially modifiable risk factor in the preclinical stages of T1D. However, in contrast to other reports,17–19 we could not demonstrate a higher BMI after T1D manifestation during the COVID-19 pandemic.Our study found that approximately 94% of children and adolescents diagnosed with T1D tested positive for beta-cell autoantibodies, which is significantly higher than the 82% reported in an earlier analysis.38 This difference is most likely due to the inclusion of patients with both GAD and IA2 autoantibodies being measured, and to the recent availability of ZnT8-ab measurement.In addition to islet cell autoimmunity, we also analyzed the prevalence of endocrine autoimmunity, which is commonly associated with T1D. The higher frequency of TGA and thyroid autoantibodies in girls is in line with the literature.38–40 Interestingly, BMI did not appear to influence thyroid or adrenal autoimmunity. The negative association between weight and TGA positivity is rather due to the celiac disease phenotype. Regarding thyroid autoantibodies, our study found a prevalence of just under 8%, whereas other studies reported higher prevalences of around 20%.38 39 This difference is due to the young age in our patients, as thyroid antibodies are associated with older age. Furthermore, we examined thyroid antibodies at the onset of T1D, not over the course of the disease.38 39While the lack of an association between the COVID-19 pandemic and the prevalence of TGA or thyroid autoantibodies is consistent with other reports,40 41 the reduced prevalence of adrenal autoantibodies during the COVID-19 pandemic has not been reported elsewhere. The high prevalence of adrenal cortex autoantibodies in our analysis (5.6%) is striking, compared with around 3.5% in an older survey.38 Notably, the proportion of positive adrenal autoantibody findings prior to the pandemic was higher at 8.5%. However, the finding should be approached cautiously, as the number of individuals tested for adrenal autoantibodies was much lower than for other autoimmune disorders. The documented frequency of adrenal autoantibodies at the onset of T1D was only around 11%, and this frequency has remained stable over the past 15 years.38The strength of our study lies in the large number of children and adolescents with newly diagnosed T1D, sourced from an international Diabetes Registry spanning four countries. This broad representation helps minimize local biases. However, a limitation of the registry is that no causal inference is possible. It was also shown that the quality of documentation of autoantibody findings improved during the COVID-19 pandemic and more autoantibody test results were documented in the DPV registry. This is unlikely to have affected the proportion of positive autoantibody findings analyzed in this study. However, increased documentation of individual autoantibody findings may increase the likelihood of a positive result, which may partly account for the observed increased positivity rate.One further limitation of this study is that the methods used to measure autoantibodies are not standardized, so that they may vary across participating sites. This methodological heterogeneity may introduce site-specific variation in reported autoantibody frequencies, which restricts the ability to draw firm conclusions about comparatively small differences.In addition, islet autoantibody testing was not available for all patients. Patients without measurements had slightly higher BMI before the COVID-19 pandemic, while the age and sex distributions were comparable.Conclusion Our results suggest that the COVID-19 pandemic and body weight are relevant modulators of autoimmunity in children with T1D. The differing association patterns observed with age and sex further highlight the complexity of these mechanisms. The COVID-19 pandemic appears to have influenced specific aspects of beta-cell autoimmunity in children, especially the emergence of IA2 and ZnT8 autoantibodies. The results of our study could suggest that the pandemic may have selectively influenced T1D in at-risk children and adolescents and that an acceleration of the autoimmune process may have occurred in a subgroup of individuals with preclinical T1D. In addition, BMI, age, and sex are significantly associated with autoimmune responses in this population. The association of BMI with specific islet autoantibodies—especially in boys—suggests that BMI may influence islet autoantibody development and underscores its potential role in modulating autoimmunity. These findings provide new insights into the complex interactions between environmental changes, body weight, and autoimmune responses in children. This knowledge may aid in future risk stratification for autoimmunity and offer potential avenues for intervention.",
  "title": "Associations of body weight and COVID-19 with autoimmunity in pediatric new-onset type 1 diabetes: results from the prospective DPV registry",
  "uid": "5f15cef0-49d6-5d3c-9150-be2611e18f13"
}