The study was performed under the auspices of the International Atomic Energy Commission, Vienna, Austria, with the aim of determining the optimal minimum therapeutic dose of iodine-131 for Graves' disease. The study was designed as a single-blinded randomised prospective outcome trial. Fifty-eight patients were enrolled, consisting of 50 females and 8 males aged from 17 to 75 years. Each patient was investigated by clinical assessment, biochemical and immunological assessment, thyroid ultrasound, technetium-99m thyroid scintigraphy and 24-h thyroid 131I uptake. Patients were then randomised into two treatment groups, one receiving 60 Gy and the other receiving 90 Gy thyroid tissue absorbed dose of radioiodine. The end-point markers were clinical and biochemical response to treatment. The median follow-up period was 37.5 months (range, 24-48 months). Among the 57 patients who completed final follow-up, a euthyroid state was achieved in 26 patients (46%), 27 patients (47%) were rendered hypothyroid and four patients (7%) remained hyperthyroid. Thirty-four patients (60%) remained hyperthyroid at 6 months after the initial radioiodine dose (median dose 126 MBq), and a total of 21 patients required additional radioiodine therapy (median total dose 640 MBq; range 370-1,485 MBq). At 6-month follow-up, of the 29 patients who received a thyroid tissue dose of 90 Gy, 17 (59%) remained hyperthyroid. By comparison, of the 28 patients who received a thyroid tissue dose of 60 Gy, 17 (61%) remained hyperthyroid. No significant difference in treatment response was found (P=0.881). At 6 months, five patients in the 90-Gy group were hypothyroid, compared to two patients in the 60-Gy group (P=0.246). Overall at 6 months, non-responders to low-dose therapy had a significantly larger thyroid gland mass (respective means: 35.9 ml vs 21.9 ml) and significantly higher levels of serum thyroglobulin (respective means: 597.6 microg/l vs 96.9 microg/l). Where low-dose radioiodine treatment of Graves' disease is considered, a dose of 60 Gy will yield a 39% response rate at 6 months while minimising early hypothyroidism. No significant advantage in response rate is gained by using a dose of 90 Gy. For more rapid therapeutic effect at the expense of an increased rate of hypothyroidism, doses in excess of 120 Gy may be required. Ultrasound determination of thyroid mass and measurement of serum thyroglobulin levels may be predictive of those patients who will be less responsive to low-dose therapy.