Since no causative treatment of Graves' disease is available, symptomatic therapy has been derived from clinical studies (evidence class I and II). To date, conservative treatment still has a relapse rate of 30-50% that is most likely due to the heterogeneity of patients. There is no clear data about the duration of therapy necessary for the individual patient. Previous prospective randomised studies have evaluated prognostic parameters including the use of levothyroxine for relapse prevention. For an improved characterisation of the natural course of Graves' disease, the period after drug withdrawal was studied. As patients treated with high doses of antithyroid drugs (< 10 mg thiamazole) and patients with large goitres (< 50 ml) often develop hyperthyroidism, they make likely candidates for radioiodine therapy or surgery. Smoking is a very strong independent risk factor for relapse. A prospective multi-centre trial of the Thyroid Chapter of the German Endocrine Society (DGE-SSD) demonstrated that the basal TSH is the strongest prognostic parameter to predict relapse of hyperthyroidism with an accuracy of 75%. In the present study, levothyroxine supplementation did not prevent recurrent hyperthyroidism. In order to avoid hypothyroidism levothyroxine should only be used in combination with antithyroid drugs. In the face of the amelioration of nutritional iodine uptake in Germany, the application of low doses of antithyroid drugs will very likely no longer be sufficient. Future therapeutic recommendations will be based on selection criteria such as iodine supplementation, volume of goitre, smoking habits and basal TSH. used in combination with antithyroid drugs. In the face of the amelioration of nutritional iodine uptake in Germany, the application of low doses of antithyroid drugs will very likely no longer be sufficient. Future therapeutic recommendations will be based on selection criteria such as iodine supplementation, volume of goitre, smoking habits and basal TSH.