We assessed the effects of dose, regimen and duration of anti-thyroid drug therapy for Graves' thyrotoxicosis on recurrence of hyperthyroidism, course of ophthalmopathy, adverse effects, health-related quality of life and economic outcomes. We undertook a systematic review and meta-analyses of randomised controlled trials (RCTs). We identified RCTs regardless of language or publication status by searching six databases, and trial registries. Dual, blinded data abstraction and quality assessment were undertaken. Trials included provided therapy for at least 6 months with follow-up at least 1 year after drug cessation. Fixed or random effects meta-analyses were used to combine study data. Twelve trials compared a Block-Replace regimen (requiring a higher dose of anti-thyroid drug treatment) with a Titration regimen. Overall, there was no significant difference between the regimens for relapse of hyperthyroidism (relative risk (RR) = 0.93, 95% confidence interval (CI) 0.84 to 1.03). Participants were more likely to withdraw due to adverse events with a Block-Replace regimen (RR = 1.89, 95% CI 1.25 to 2.85). Prescribing replacement thyroxine, either with the anti-thyroid drug treatment, or after this was completed, had no significant effect on relapse. Limited evidence suggested 12-18 months of anti-thyroid drug treatment should be used. The titration regimen appeared as effective as the Block-Replace regimen, and was associated with fewer adverse effects. However, relapse rates over 50% and high participant drop-out rates in trials mean that the results should be interpreted with caution, and may suggest that other strategies for the management of Graves' disease, such as radioiodine, should be considered more frequently as first-line therapy. There were no data on the course of ophthalmopathy, health-related quality of life and economic outcomes.