Graves' Disease in Children: Long-term Outcomes of Medical Therapy

Shona Rabon; Amy M. Burton; Perrin C. White


Clin Endocrinol. 2016;85(4):632-635. 

In This Article


There were 291 children meeting diagnostic criteria for Graves' disease. Sixty-two were male (21%). One hundred seventeen (40%) were Hispanic, 90 (31%) were Caucasian, 59 (20%) were African American, and 25 (8%) were Other, primarily south Asian. Mean age (±standard deviation) at diagnosis was 12·3 ± 3·8 (range 3–18·5 years). Two hundred sixty-eight patients were initially started on an ATD (251 on methimazole, 17 on PTU), whereas 23 underwent thyroid ablation or surgery at diagnosis (at the behest of the patient and family). Fifty-seven of the 268 patients started on medical therapy (21%) achieved provisional remission (Fig. 1a). Remission rates remained steady for approximately 5·5 years and then decreased, but one patient remitted after 11 years of medical therapy. Sixteen (28%) of 57 patients relapsed after achieving provisional remission, most within 16 months (Fig. 1b). Of the 268 patients started on an ATD, 152 (57%) eventually received radioactive iodine or underwent thyroidectomy. Ninety-eight patients remained on medication, and 41 were in remission at the time of last contact. Reasons for definitive therapy included noncompliance, convenience, side effects of medication, or poor control on medical therapy. Based on Kaplan–Meier survival analyses, neither remission nor relapse rates were affected by gender or ethnicity. Note that this type of analysis takes right-censored data into account (i.e. patients who had not experienced remission or relapse, respectively, at their last follow-up, or who had undergone definitive therapy without ever going into remission) and therefore remains valid with variable durations of follow-up. We also did not observe any correlation between the probability of remission and initial free or total T4, or presence of antithyroid peroxidase or antithyroglobulin antibodies (not shown). Thyroid-stimulating immunoglobulin or antithyrotrophin receptor antibodies were rarely available on patients diagnosed before 2004 (14/85 in 2003 or before, 147/206 in 2004 or after). There was no difference in rate of remission based on presence of thyroid-stimulating immunoglobulin or antithyrotrophin receptor antibodies in patients diagnosed in 2004 or thereafter (not shown).

Figure 1.

Kaplan–Meier survival analyses. (a) Remission. Grey dots denote patients who experienced an event (remission), and the time in years from diagnosis until the event occurred. Black dots denote patients who did not experience remission, with the dot located at the time of definitive treatment or at the last follow-up (right censoring). There are many dots denoting other censored patients that are not visible because they are obscured by dots denoting patients who went into remission. (b) Subsequent relapse. Grey dots represent patients who relapsed from an initial remission, and the time in years from remission to relapse. Black dots represent censored patients who had not relapsed at the time of their last follow-up. Many dots denoting censored patients are not visible.

Of the 251 patients treated with methimazole, 53 (21%) had at least one adverse event, most commonly rash (29 patients), arthralgias (12), elevated transaminases (10), and neutropenia (nine). Neutropenia was mild in all cases, and none of the patients had an absolute neutrophil count <1000 cells/μl. Less common complaints were pruritus (four patients), myalgias (three), shortness of breath (two), and muscle cramps with elevated creatine kinase (one). Median (range) time to adverse event was 22 (5 days to 2·5 years) days. No patient was admitted to the hospital for a drug reaction, although several were evaluated in the emergency department. Methimazole was discontinued in 37 patients after adverse events; they underwent definitive treatment, the majority opting for radioactive iodine ablation.