University of Tasmania
140409 - can islets autoantibody profile.pdf (2.28 MB)

Can islets autoantibody profile at diagnosis of Type 1 Diabetes in children and adolescents predict the future development of Hashimoto’s thyroiditis and coeliac disease?

Download (2.28 MB)
journal contribution
posted on 2023-05-21, 19:30 authored by Shee, A, Sarah PriorSarah Prior, Nicole ReevesNicole Reeves

Aim: This study aims to explore if any correlation exists between islets autoantibody profile at diagnosis in children and adolescents with type 1 diabetes and subsequent development of Hashimoto’s thyroiditis and coeliac disease.

Materials and Methods: In this multicentre retrospective cohort study conducted in three hospitals, for children below 18 years of age diagnosed with Type 1 Diabetes (T1D) over a ten-year period between 1 January 2009 and 31 December 2018, electronic Medical Records (DMR) were looked at and analysed by using SPSS V24.0, for several data variables such as antibody profile (type and number of three antibodies-GAD, IAA and ZnTr8 present and their titres), demographical characteristics (BMI, gender, age of onset), severity of disease at presentation (DKA vs non-DKA and HbA1c value) and subsequent development of Hashimoto’s Thyroiditis (HT) and Coeliac Disease (CD).

Results: Out of 218 children (male = 112, female = 103) tested during this study period, IAA was the most prevalent antibody found which was positive in 90 (41%) cases followed by GAD in 83(38%) and ZnTr8 in 34 (16%) children. Median age of onset of T1D was 7.709 years (range 1-17 years) and the mean HbA1c at presentation was 8.8% (range 5.5% - 14.6%). HT and CD were identified in 31 (12.8%) and 14 (5.8%) cases respectively in this cohort. Independent samples t-tests could not identify any significant difference in the GAD, IAA and ZnTr8 antibody levels for the children with or without HT{(M = 148.16, SD = 507.9) and (M = 117.6, SD = 402.2); t (103) = 0.275, p = 0.784 in GAD}, {(M = 514.46, SD = 1360.7) and (M = 316.86, SD = 972.02); t (99) = 0.697, p = 0.487) in IAA}, {(M = 309.63, SD = 397.76) and (M = 383.26, SD = 462.70) ; t (32) = 0.385, p = 0.703) in ZnTr8}. Likewise, there has not been any statistically significant difference for those with or without CD {(M = 57.9, SD = 107.5) and (M = 91.21, SD = 326.4); t (93) = 0.335, p = 0.739 in GAD}, {(M = 22.41, SD = 18.0) and (M = 420.87, SD = 1140.2); t (90) = 1.10, p = 0.274 in IAA}. One-way ANOVA {F (2) = .970, p = 0.382} could not identify any correlation between the number of antibodies present at diagnosis and subsequent development of HT. There was, however, a statistically significant difference recognised, as determined by a one-way ANOVA {F (2,112) = 3.305, p = < .05} for the number of antibodies present and development of CD. Post hoc comparisons using the Tukey test showed children presenting with all 3 antibodies were 3.87 times more likely to develop CD(p = 0.036).

Conclusion: IAA was the most prevalent islets autoantibody found in this study, followed by GAD and ZnTr8. All three antibodies were positive in 13% cases. Although, GAD, IAA and ZnTr8 antibody titres at diagnosis are not predictive of subsequent development of HT and CD, the number of antibodies present influences the future risk of CD but not for HT. Presence of all the three antibodies nearly quadruples the possibility of subsequent development of CD. As HLA typing is not a predictor of development of CD in children with T1D, this finding can influence the frequency of serology testing.


Publication title

Journal of Pediatric Endocrinology






Tasmanian School of Medicine


Austin Publishing Group

Place of publication

United States

Rights statement

Shee et al. © All rights are reserved. All site content is licensed under a Creative Commons Attribution ( CC BY) license.

Repository Status

  • Open

Socio-economic Objectives

Health status (incl. wellbeing); Neonatal and child health