Carbimazole therapy can induce insulin autoantibodies (IAA) in Japanese Graves' disease patients, a phenomenon possibly linked to their immunogenetic profile. This phenomenon is not observed in Caucasians. We assessed IAA levels in 114 North Indian Graves' disease patients before and after carbimazole therapy (mean duration 6.2 +/- 3.9 months). The functional significance of IAA was assessed in 46 of them by first phase (sum of +1 and +3 min) insulin response to intravenous glucose (IVGTT) and an oral glucose tolerance test (OGTT) undertaken before commencement of the carbimazole therapy. IAA were measured using a radiobinding assay and expressed as the assay precision unit, S.D. scores (S.D.S), over healthy controls. Before treatment 22 of 114 (19.3%) patients were IAA positive (mean +/- S.D., 5.9 +/- 3.2 S.D.S). After carbimazole therapy a further 11 (9.6%) showed positive for IAA (mean +/- S.D., 3.5 +/- 1 S.D.S). Of the 22 patients who were IAA positive before treatment, 12 became negative after carbimazole therapy. The fasting insulin and first phase insulin responses were similar in IAA positive and IAA negative Graves' disease patients (mean +/- S.D., 61.7 +/- 35.9 versus 88.3 +/- 46.6 pmol/l, P = 0.123 and 1127 +/- 696 versus 1033 +/- 430 pmol/l, P = 0.716, respectively). The OGTT results were comparable in the IAA positive and the IAA negative groups. Thus, North Indian Graves' disease patients, who resemble Caucasians in their HLA haplotypes, behave like Japanese in their tendency to become IAA positive with carbimazole therapy. A subset of the patients who were IAA positive before treatment also demonstrated negative IAA (12/22) after carbimazole therapy.