Development of Clinical Prediction Score for Failure of Antithyroid Drug Treatment in Graves’ disease Patients

Abstract

BACKGROUND: The current modalities for treatment of Graves’ disease include antithyroid drugs, radioactive iodine ablation, and total thyroidectomy. The patient’s decision depends on clinical characteristics and the patient’s needs. Although long-term antithyroid drug treatment is another choice for patients, nearly half of patients experience failure of antithyroid drug treatment and undergo radioiodine ablation or thyroidectomy. 

OBJECTIVES: To develop a clinical prediction score using prognostic factors for failure of antithyroid drug treatment in Graves’ disease patients.

METHODS: The study was a retrospective observational cohort design. 630 Graves’ disease patients at Surin hospital from 2018 to 2025 were reviewed. Important clinical data from multivariable logistic regression were collected to compute the coefficient, which was then transformed into a score.

RESULTS: All 630 Graves’ disease patients were divided into two groups: 306 patients with antithyroid drug failure and 324 patients in remission. The clinical prediction score was developed using five significant clinical parameters, including relapse history, thyroid gland size(≥45 g), thyroid-related orbitopathy, FT4>3.5 ng/dL, and MMI dose at 6^th month ≥ 10 mg/d. The prediction score ranged from 0-7, with a score of 0-4 denoting a low risk (PPV= 38.9%) and a score of 4.5-7 a high risk (PPV=81.3%) for failure of antithyroid drug treatment. The AuROC of the score was 0.78 (95% CI 0.74-0.82) 

CONCLUSIONS: Graves’ disease patients who have a clinical prediction score of 4.5-7 have a high probability for failure of antithyroid drug treatment. Physicians should recommend these patients be treated with radioactive iodine or thyroidectomy.