To evaluate clinical and biochemical features of 115 children (98 female, mean age 11.3 ± 3.5 years) with Graves disease to identify possible determinants of remission. We defined as positive outcome the improvement of clinical features and restoration of euthyroidism or induction of hypothyroidism after antithyroid drug (ATD) therapy and as negative outcome hyperthyroidism persistent over 2 years of ATD therapy or relapsed after ATD withdrawal. Thirty-eight children (33%) had remission after 2 years of ATD therapy. The absence of goiter at diagnosis was correlated with a better outcome. Median thyroid-stimulating hormone receptor antibody (TRAb) values at diagnosis were significantly lower in patients with a positive outcome (P = .031). We found a significant relationship between the time required for TRAb normalization and the patient outcome; TRAb normalization within 1 year from time of Graves disease diagnosis was significantly more common among patients with a positive outcome (P < .0001), and the mean time for TRAb normalization was significantly shorter in patients with a positive outcome (1.3 ± 0.8 years) compared with that observed in patients with a negative outcome (2.5 ± 2.7 years, P = .026). Although no clinical variable investigated is constantly associated with a definite outcome, the absence of goiter at the diagnosis may be associated with a better outcome. The most relevant predictor of Graves disease outcome was serum level; TRAb at time of Graves disease diagnosis less than 2.5 times the upper reference limit, TRAb normalization during ATD, and TRAb normalization timing each may predict positive outcomes. These results may have a role in the empiric clinical management of pediatric patients with Graves disease.