Early diagnosis and relapse prediction in Graves' disease influences treatment. We assessed the abilities of four TSH-receptor antibody tests [TRAb] and one cyclic adenosine monophosphate bioassay to predict relapse of Graves' disease.Observational study investigating patients presenting with Graves' disease at a Swiss hospital endocrine referral center or an endocrine outpatient clinic. Main outcomes were diagnosis and relapse of Graves' disease after stop of anti-thyroid drugs. We used Cox regression to study associations of TRAb levels with relapse risk and calculated c-statistics [AUC] to assess discrimination. Blood draws took place as close as possible to treatment initiation.AUCs ranged from 0.90 (TSAb Biossay by RSR) to 0.97 (IMMULITE TSI by Siemens). Highest sensitivity (94.0%) was observed for IMMULITE TSI and RSR TRAb Fast, while the greatest specificity (97.9%) was found with the EliA anti-TSH-R (by Thermo Fisher). In Cox regression analysis comparing the highest versus the lower quartiles, the highest hazard ratio [HR] for relapse was found for BRAHMS TRAK (by Thermo Fisher) (2.98, 95% CI 1.13-7.84), IMMULITE TSI (2.40, 95% CI 0.91-6.35), EliA anti-TSH-R (2.05, 95% CI 0.82-5.10), RSR Fast TRAb (1.80, 95% CI 0.73-4.43), followed by RSR STIMULATION (1.18, 95% CI 0.46-2.99). Discrimination analyses showed respective AUCs of 0.68, 0.65, 0.64, 0.64, and 0.59.The assays tested had good diagnostic power and relapse risk prediction with few differences among the new assays. Due to the small sample size and retrospective design with possible selection bias, our data need prospective validation.