Thyroid and adrenal autoimmunity in children with newly diagnosed type 1 diabetes mellitus

INTRODUCTION

It is well recognized that type 1 diabetes mellitus (T1DM) is associated with the development of other autoimmune diseases, with 11.2% of individuals with T1DM reported to have associated autoimmune thyroid disease, 0.2% developing Addison disease, and 4.5% developing associated celiac disease.^[1] Having additional autoimmune pathologies in conjunction with T1DM can have consequences when it comes to diabetes management, and therefore, this warrants due consideration.^[2]

The reported prevalence of thyroid autoantibodies in children with T1DM varies from 3 to 50%.^[3,4] In particular, thyroid peroxidase (TPO) antibodies are frequently associated with T1DM with a reported range (95% confidence interval, CI) of between 13 (0.3–71) and 326 (194–510) per 10,000 patient-years.^[5] Thyroid hormone has various effects on glucose homeostasis including increased expression of glucose transporter-2 (GLUT2) in the liver increasing gluconeogenesis and glycogenolysis, stimulation of insulin secretion by pancreatic β-cells; increasing glucagon release by pancreatic α-cells; and increasing glucose transporter type 4 (GLUT4) gene expression, thereby enhancing glucose uptake in skeletal muscle.^[6] Therefore, the current guidelines from the International Society for Pediatric and Adolescent Diabetes (ISPAD) suggest screening for TPO antibodies and thyroid function at diagnosis, after initial blood glucose management, and periodic monitoring of thyroid function every 2 years or more frequently if thyroid antibodies are detected at diagnosis or symptoms become apparent.^[7]

Detectable adrenal autoantibodies are reported in up to 2% of patients with T1DM.^[7] It has been demonstrated that up to 90% of children with adrenal cortex antibodies will progress to clinical Addison disease.^[8] Glucocorticoids play a crucial role in blood glucose regulation and act on tissues including the liver, pancreas, and skeletal muscle; therefore, individuals with adrenal insufficiency are more susceptible to hypoglycemia.^[9] Despite being one of the rarer autoimmune diseases associated with T1DM, Addison disease in combination with diabetes increases the difficulty of diabetes management. In addition, patients with both T1DM and Addison disease are four times more likely to suffer premature death compared to those with T1DM alone.^[10]

It is debated whether children with T1DM require routine screening for Addison disease.^[11] ISPAD’s current recommendation highlights the importance of being alert to symptoms of Addison disease, but no routine screening is suggested. Symptoms to be aware of include hypoglycemia, an unexplained decrease in insulin requirements, hyperpigmentation of the skin, lethargy, weight loss, hyponatremia, and hyperkalemia.^[7]

While screening guidelines are clear, there is less guidance on the management of children with T1DM who have normal thyroid and adrenal function despite positive antibody tests. This study aimed to identify the local prevalence of thyroid and adrenal autoimmunity and dysfunction in children with T1DM.

MATERIAL AND METHODS

Patients and methods

A retrospective case note review was undertaken of all children who presented to the Royal Hospital for Children, Glasgow, with newly diagnosed T1DM between 1 September 2017 and 1 September 2022. This tertiary pediatric hospital takes new patient referrals from birth to the age of 16.0 years from the Greater Glasgow and Clyde area, with a median (range) of 70 (61–84) new referrals per year during the study time period. Data on TPO status, adrenal autoantibody status, thyroid and adrenal dysfunction were obtained from diagnosis to 24 May 2024 via review of hospital databases. Figure 1 demonstrates a flow diagram of patient selection. Historically, local clinical practice involved assessment of adrenal antibodies at diagnosis; however, these were not performed annually. Patients had a minimum follow-up period of 18 months and a maximum follow-up period of 6 years and 8 months. Only 46 (12%) children included in the study were diagnosed with diabetes in 2022, thus having a shorter period of follow-up. Thyroid function was assessed at diagnosis and regularly thereafter for all children, regardless of TPO antibody status. Children with prior diagnoses of hypothyroidism or Addison disease were excluded from the study. Data were also obtained regarding glutamic acid decarboxylase (GAD), islet antigen-2 (IA2), and zinc transporter-8 (ZnT8) autoantibody status. The children’s electronic medical records were accessed to determine whether treatment for thyroid or adrenal dysfunction was started and how often screening was undertaken.

Close

Figure 1:

Flow diagram of patient selection. T1DM: type 1 diabetes mellitus, TPO: Thyroid peroxidase

Export to PPT

RESULTS

In total, 504 patients were initially identified as presenting with a new diagnosis of T1DM during the study time period. However, 120 were excluded: the number lost to follow-up was 116; patients had either moved out-of-area (21), transitioned to adult services (12), or had data unavailable (83). Of those who had data unavailable, 74 patients had only one recorded set of TFTs during the study time period, three patients had no traceable data, and 6 patients had no TPO status recorded. Patients were further excluded due to a prior diagnosis of hypothyroidism (2), a prior diagnosis of Addison disease (1), or death during the study time period (1). Overall, data were available for 384 patients (53% female) with a median age of 9.6 years (range 0.9–16.1 year) at T1DM diagnosis. The median duration of diabetes was 4.1 years (range 1.7–6.7 year). Out of the 384 patients included in the study, 243 (63%) were tested for GAD antibodies, 136 (35%) were tested for IA2 antibodies, and 132 (34%) were tested for ZnT8 antibodies. Of these, 187/243 (77%) had positive GAD, 75/136 (55%) had positive IA2, and 74/132 (56%) had positive ZnT8 antibodies.

DISCUSSION

In this study, we retrospectively followed up 384 pediatric patients who presented with a new diagnosis of T1DM between 1 September 2017 and 1 September 2022. These patients were followed up until 24 May 2024, providing a minimum of 18-month follow-up per patient.

Positive TPO antibodies were found in 18% of our patient cohort (95% CI of 14–22%) which is consistent with other literature that suggests that 17–27% of type 1 diabetics express TPO antibodies compared with 13% found in the general population.^[12] In addition, there was a female predominance in the group with positive TPO antibodies which is again, concurrent with other reported findings in the literature.^[13] We found no association between the presence of TPO antibodies and age at diagnosis, disease duration, or positivity of GAD, IA2, or ZnT8 antibodies.

Overall, 3% of our cohort developed thyroid dysfunction which conflicts with reports that up to 28% of individuals with T1DM have associated autoimmune hypothyroidism.^[12] This may be due to the shorter follow-up period in our study as other prospective studies have reported a mean of 13 years between the diagnosis of diabetes and the development of hypothyroidism.^[14] Our period of follow-up ranged from 1.7 to 6.7 years, and therefore, some patients who may go on to develop hypothyroidism will not have been observed within our study. Around 80% of the patients who developed hypothyroidism were female, which is consistent with other studies.^[12,13] We observed no significant differences between the hypothyroid group and age at diagnosis, disease duration, or positivity of islet cell antibodies, and to our knowledge, there are no particular population-specific factors that may be responsible for the reduced rates of thyroid dysfunction in our cohort.

Within our cohort of 384, 380 (99%) were tested for adrenal antibodies at diagnosis of which 3 (0.8%) children were found to be positive. This is akin to larger studies reporting a prevalence of 0.9–1.5%.^[8,14] We found no relationship between age at diagnosis, disease duration, or the presence of islet cell antibodies in the patients with adrenal antibodies. None of the children with positive antibodies has developed adrenal insufficiency to date. However, a large study of pediatric patients with organ-specific autoimmune disease found that 90% of children with positive adrenal antibodies would go on to develop adrenal insufficiency within 10 years. Therefore, our shorter follow-up period may be the reason why we did not observe any children who developed adrenal insufficiency, and our observations should be interpreted with caution.^[8] However, due to the rarity of the condition, our findings are still representative of a typical cohort of children with diabetes.

Our results demonstrate that adrenal autoimmunity and adrenal insufficiency are rare within this patient cohort. Therefore, screening for adrenal autoimmunity can be much less stringent in comparison to screening for autoimmune thyroid dysfunction. However, individuals with T1DM are still at an increased risk of adrenal autoimmune pathology in comparison with the general population. Clinicians should be aware of the presenting symptoms of adrenal insufficiency in those with T1DM and the consequences that adrenal insufficiency can have on the complexity of diabetes management.

Routine screening is costly and can cause unnecessary distress to individuals with T1DM, who are concerned about the development of additional autoimmune conditions, which may also impair their diabetes control. While screening guidelines are clear, guidance on how to approach children who have normal adrenal and thyroid function in the context of a positive antibody status is less. As such, further to our study, guidelines have been created locally for a pragmatic approach to screening of children and young people in Greater Glasgow and Clyde with T1DM for thyroid dysfunction [Figure 2] and adrenal insufficiency [Figure 3]. Antibodies should be screened at the time of diagnosis for both conditions. In screening for thyroid dysfunction, TFTs should be performed at diagnosis and actioned according to TSH and FT4 levels as per Figure 2. The requirement for a short Synacthen test in screening for adrenal dysfunction is guided by the presence of adrenal antibodies, as demonstrated in Figure 3. Clinicians are also recommended to discuss symptoms of both thyroid dysfunction and adrenal insufficiency at regular diabetes review clinics.

Close

Figure 2:

Pathway for thyroid autoimmunity and dysfunction in children with T1DM, from Greater Glasgow and Clyde. TSH: Thyroid-stimulating hormone, FT4: Free thyroxine, TPO: Thyroid peroxidase, TFT: Thyroid function test.

Export to PPT

Close

Figure 3:

Pathway for management of patients with T1DM with adrenal autoimmunity and/or symptoms in Greater Glasgow and Clyde. AB: Antibodies, AD: Addison disease.

Export to PPT

A weakness of our study was missing data regarding islet cell antibodies. Inconsistency in testing meant that <65% were tested for GAD antibodies and <36% were tested for IA2 and ZnT8 antibodies. This could have prevented us from concluding whether the presence of islet cell antibodies impacts thyroid or adrenal autoimmunity and the development of their dysfunction. In addition, the retrospective nature of our study introduces limitations due to potential missing data, as 116 patients were lost to follow-up. However, our study reflects real-world data and is representative of the true environment of the study population within our center. Therefore, it remains representative of a typical cohort of children with diabetes. The study was undertaken in a single center, which may limit the ability to apply the findings to other populations. Despite this, it provides important contemporary data which was used to inform important guidelines used within our center.

CONCLUSION

In this retrospective cohort study, 18% of our patient cohort had raised TPO antibodies and 15% of these developed hypothyroidism. Females were more likely to have raised TPO antibodies. Adrenal autoantibodies were less prevalent in our patient cohort (1%), with none developing adrenal insufficiency following a diagnosis of T1DM within the follow-up period. Our study helps to confirm the need for screening for thyroid and adrenal antibodies in this patient group. Our results reinforce the rarity of adrenal insufficiency in children with diabetes, but given the non-specific, vague symptoms of hypoadrenalism that can coexist with variable glycemic control, adrenal antibody screening is a useful addition to current clinical care, especially considering the life-threatening nature of adrenal insufficiency. We have developed a guideline for ongoing management for those with positive thyroid/adrenal antibodies, offering a pragmatic approach without overburdening patients.