In a prospective, randomized study of 135 newly diagnosed patients with hyperthyroidism due to Graves' disease we compared the effect on remission rates of additional triiodothyronine (T3) with conventional antithyroid drug therapy. To this end 114 patients were followed for at least 12 months (15.7+/-4.9, mean+/-s.d.) after the discontinuation of any therapy. After return of thyroid function to normal (8.5+/-7.4 weeks, mean+/-s.d.) patients were maintained on antithyroid medication for 9.0+/-2.5 months. They were then randomly assigned to one of three groups: group 1 (n=44) stopped methimazole, groups 2 (n=39) and 3 (n=31) continued with exogenous T3 (not exceeding 75 microgram/day in any patient) for a further 6 months either with (group 2) or without (group 3) a fixed dose of 10mg methimazole daily. The T3 dose was kept variable to keep TSH suppressed (<0. 1mU/l), which could be achieved in 82% of patients on 100% of their monthly visits. No serious side-effect requiring the discontinuation of the study occurred in any patient. Total T3, TSH-receptor antibodies and some previously suggested potential predictors of relapse including thyroid size by ultrasound, 24h urinary iodine excretion, history of cigarette smoking and ophthalmopathy were determined at the outset of the study and subsequently every 6 months (and total T3 every 4 weeks). No significant difference (P>0.05, Chi square) was seen in relapse of hyperthyroidism after a mean follow-up of 16 months (range: 12-31 months; groups 1:52%, 2:44% and 3:42%) in an area of low-to-moderate iodine intake (prevalence of 24h urinary iodine excretion <100 microgram/24h: 17 and 25% at two different measurements respectively). Concomitantly, no predictor of recurrence of disease could be identified, irrespective of treatment modality.