Recently, we demonstrated that higher levels of autoantibodies to the human TSH receptor (TBII) predict relapse of hyperthyroidism in Graves' disease (GD). The aim of this study was to extend this outcome prediction by dividing TBII into stimulating (TSAb) and blocking (TBAb) TSH receptor autoantibodies. Altogether, ninety patients (81 female, 9 male) were retrospectively analyzed; sixty-four patients (71 %) did not go into remission or relapsed, whereas twenty-six patients (29 %) went into remission (median follow-up: 17.5 months). TSAb and TBAb measurement was performed in a CHO cell bioassay with cAMP readout at the time of their first visit in our outpatient clinic (single point measurement in median 6.5 months after initial diagnosis). In the remission group, eighteen of twenty-six patients (69 %) were TSAb-positive, whereas fifty-three of sixty-four patients (83 %) were TSAb-positive in the relapse group (p = ns). The mean stimulation indices (SI) were 4.1 in the remission group and 12.9 in the relapse group, respectively (p = 0.015). By using a threshold of 10 SI, the specificity for relapse was 96.0 %, as only one in twenty patients with an SI above 10 went into remission during follow-up (PPV 95 %). Most TSAb-positive patients also had high levels of TBII. Neither group showed any difference with respect to blocking type autoantibodies, which were mostly negative in both groups. In summary, high TSAb levels are similar but not superior to TBII for predicting relapse in GD patients. In contrast, TBAb measurement does not add any valuable information in this context. In the clinical routine, TSAb/TBAb measurement may not play an important role for diagnosis or outcome prediction of GD, since sensitive 2 (nd) generation TBII assays are easier to perform and offer similar information to the clinician. Bioassays should be reserved for special clinical questions such as Graves' disease in pregnancy.