Antithyroid drugs are currently the mainstay of Graves’ disease therapy with considerable limitations regarding their use and often unsatisfactory results. Although generally well tolerated there can be a high prevalence of mild side effects such as skin rash, itching and leucopenia. Agranulocytosis is rare but potentially life-threatening and other major adverse effects including aplastic anaemia, thrombocytopenia, systemic lupus erythematous and vasculitis have been reported (2). Propylthiouracil (PTU) may induce liver damage (3) and fatal hepatotoxicity (4).

As shown in our case series remission rates with ATDs can be poor (20%–30%) (5–8). High relapse rates exist in certain patient groups and the only apparent modifiable factor appears to be duration of ATD therapy with a reduced relapse rate in those treated with prolonged oral therapy (9, 10). To optimise management, patients would ideally be stratified depending on their clinical and biochemical characteristics at presentation to either ATD therapy, if prognostic factors indicate a high likelihood of remission, or definitive treatment.

Many studies investigating prognostic factors at presentation for remission following ATDs have been performed in adults (10–16); however, results are not transferable and sometimes appear to contrast those of the paediatric population. Retrospective studies in children and adolescents (6, 10, 17–26) have identified younger age, larger goitre size, lower initial body mass index (BMI), higher thyroid hormone concentrations, and higher TRAb levels at diagnosis as predictors of relapse after stopping ATD therapy. Only two studies have investigated prognostic factors in children prospectively (9, 27) (Table 2).

Following a previous review, Glaser and Styne (27) performed a prospective multicentre study with remission defined as euthyroid status 12 months after completing a 2-year course of ATDs. Of the 70 children who enrolled, 12 dropped out before completing the 2-year ATD treatment period – four due to adverse reactions (although 11 children in total developed adverse reactions) and eight due to patient preference. A univariate analysis was performed on the results from the 51 children who completed the study and showed significantly lower T4 and T3 concentrations at presentation in children who achieved remission. There was no significant difference in BMI standard deviation (SD) score but when the population was divided into groups based on ethnicity, patients who achieved remission in the Caucasian group had higher BMI SD scores whereas children of other ethnicities who achieved remission had lower BMI SD scores. Although a multivariate logistic regression analysis did not identify significant factors, binary recursive partitioning analysis (used to derive a highly sensitive clinical prediction rule) showed age and initial T3 concentration as significant variables with older age (>14–15 years) (p=0.06) and lower T3 level indicating a greater chance of remission. In this study goitre size was not a useful predictor of remission; however, it was measured manually by clinicians knowing the hyperthyroid status of the patients and consequently may have been prone to overestimating goitre size introducing bias. More precise ultrasound measurement at presentation may provide better prognostic information.

Kaguelidou et al. (9) performed a prospective observational study including 147 children to identify risk factors for relapse within 2 years of completing a 2-year course of ATD therapy. Some 32% of patients found it difficult to comply with the treatment regimen while 6% of patients experienced adverse reactions to the ATDs. Four variables at presentation were found to be independent predictors of relapse: younger age; high FT4 levels; high TRAb levels and being non-Caucasian in ethnicity. When these variables underwent internal validation, ethnicity and FT4 levels were particularly strong predictors of relapse. A prognostic scoring system was developed to aid decision making on an individual patient basis.

Hyperthyroidism can perpetuate itself by worsening the autoimmune process and autoimmunity results in the production of more TSH receptor antibodies (28). Therefore, noncompliance with ATDs can have a profound effect making remission less likely. Noncompliance or nonadherence with medication in children and adolescents is estimated to range from 25% to 60% (29, 30). Concordance is a particular problem in those with chronic conditions, especially in adolescents (31) (the age group with the highest incidence of Graves’ disease) who desire independence with less parental input (30), and in those with learning disabilities (32). Educational strategies are often implemented; however, evidence suggests that these result in only small to medium benefits (29). Adjusting dosing regimens can help and compliance has been proven to be better with once daily methimazole/carbimazole compared to thrice daily PTU (33).

Failure to achieve permanent remission with ATDs leads to the consideration of other treatment modalities such as RAI or surgery. Surgery may be the optimal definitive treatment in certain patient groups including children under 5 years of age and children with a large goitre (>80 g) at diagnosis (34–36). Complication rates can vary from 15% when performed by paediatric surgeons compared to 4% (p<0.01) for experienced thyroid surgeons (>30 thyroidectomies per year) (37). Furthermore, thyroidectomy scars, which may be hypertrophic requiring medical or surgical treatment may be considered cosmetically unacceptable and rule out this form of therapy for some adolescents.

Over 1200 cases of RAI treatment in childhood have been reported in the literature, achieving a 95% remission rate in Graves’ disease (38–41). Turner et al. (42) performed a survey of current RAI administration in patients with thyrotoxicosis over a 19-year period from 1990 to 2008 in the United Kingdom. Of 69,258 RAI treatments, 560 were in the group of patients under 21 years of age. There was a 7-fold rise in young people receiving RAI expressed as a percentage of the total from 0.2% in 1990 to 1.5% in 2008 with the youngest age receiving treatment falling from 19 to 11 years of age. This may be due to the superior efficacy of RAI treatment compared to ATDs, proven by many studies including a Cochrane Systematic review [relative risk already present (RR) 1.70, 95% confidence interval (CI) 1.29–2.24] (43).

The reluctance to use RAI in paediatric practice stems from concerns about short- and long-term complications. Nausea can occur immediately post-RAI treatment and pain over the thyroid gland, which is self-limiting and can be managed with nonsteroidal anti-inflammatory drugs (38). Thyroid storm, more common after RAI treatment in those with severe thyrotoxicosis and large goitre, can be minimised by administering ATDs before RAI ablation. Symptoms of iatrogenic hyperthyroidism after RAI treatment can be controlled with β-blockers or saturated potassium iodine (Lugol’s solution). Whereas a small proportion of adult patients have worsening eye disease following RAI, this does not appear to be a problem in children (38).

Long-term studies have shown RAI treatment to be safe. Although low-dose RAI exposure, from the environment or diagnostic procedures, is associated with an increase in thyroid carcinoma, high-dose RAI treatment which completely ablates the thyroid gland results in a much lower thyroid carcinoma incidence (44, 45). Read et al. (39) reported 26- and 36-year follow-up outcomes of 116 patients under the age of 20 who had been treated with RAI for Graves’ disease between 1953 and 1973. The age range at the time of initial treatment was 3 years 7 months–19 years 9 months. Initial doses of RAI were conservative and given to induce euthyroidism. Some patients needed repeat treatments leading to an increase in dose to ablate the thyroid. None of these patients developed leukaemia or cancer of the thyroid gland. Two patients developed single thyroid nodules which were found to be benign on biopsy. There were no other signs of genetic damage or thyroid problems, apart from iatrogenic hypothyroidism, associated with RAI treatment. Although these doses were much lower than doses used for ablation today (3000–5000 cGy versus 10,000–20,000 cGy), higher doses destroying more thyroid tissue may reduce the risk of thyroid malignancies. Other studies have also reported the safe use of RAI in children with long-term follow-up data (46). However, a large retrospective study showed a positive association of radiation dose from CT scans in children with leukaemia and brain tumours (47) and another study showed adult patients who received RAI treatment for hyperthyroidism (mean age 62 years) had an increased incidence of cancer overall (RR 1.25; 95% CI: 1.08–1.46) and specifically kidney and breast cancers (48).

The Cooperative Thyrotoxicosis Therapy Follow-up Study reported on 34,684 adult patients with hyperthyroidism treated between 1946 and 1968 with ATDs, surgery or RAI (49). At 10–20 years of post-treatment follow-up there was a 5-fold higher incidence of thyroid carcinomas in patients with Graves’ disease treated with ATDs compared to RAI and an 8-fold higher incidence than in patients treated with surgery. Rates of thyroid adenomas were 10 and 20 times higher in adults treated with ATDs than in patients treated with RAI and surgery, respectively. However, as younger children are at a greater risk of thyroid carcinoma after external exposure to RAI, there may be a small increase in the rate of thyroid cancer after RAI treatment in children, which declines progressively throughout childhood (38). In an unfortunate situation that a thyroid cancer develops, papillary carcinomas are the most common type which are slow growing and have an excellent prognosis. Long-term follow-up studies assessing nonthyroid cancer risks in children are yet to be performed. There is no evidence to support an increased risk of congenital anomalies in offspring of those treated with RAI as children (38).

There is a paucity of data on RAI as the first-line of therapy. A task force of expert clinicians commissioned by the American Thyroid Association in association with the American Association of Clinical Endocrinologists has produced thyrotoxicosis management guidelines (50). They suggest RAI can be considered at presentation if there is a low risk of remission using ATDs. The guidelines state that RAI should not be given to children under 5 years of age but is acceptable in children between 5 and 10 years of age depending on the dose. Pretreating with ATDs and β-blockers until T4 normalises before proceeding with RAI is recommended. Rivkees proposes that children approaching 10 years of age or older should be given ATDs if prognostic factors indicate a potential likelihood of remission, otherwise RAI should be considered as the first-line therapy whilst discussing the advantages and disadvantages of different treatment options with the patient and family (37).

Therefore, it would seem appropriate for children above 5 years of age, especially those of non-Caucasian ethnicity, with clinical and/or biochemical features of severe hyperthyroidism at diagnosis to be considered for RAI as the first-line therapy. There may be other patient groups where initial oral therapy is deemed inappropriate such as those with poor engagement, noncompliance or behavioural difficulties and patients with contraindicators to surgery that may proceed directly to definitive RAI treatment. Figure 1 provides a possible treatment strategy based on the clinical, biochemical and psychosocial factors of individual patients on presentation.

Figure 1:

Flow diagram stratifying initial patient therapy based on clinical and biochemical features at presentation.