The universal predictive criteria for neonatal overt thyrotoxicosis requiring treatment were examined in 108 neonates (including a pair of twins) born to mothers with Graves'' disease (36 patients under treatment with antithyroid drugs [group A] and 71 in remission [group B]). Anti-thyroid-stimulation hormone (TSH) receptor antibody activity was measured by both radioreceptor assay (TSH-binding inhibitor immunoglobulin [TBII]) and biologic stimulation assay (thyroid stimulation antibody [TSAb]). For generalization of the predictive criteria, the expression of TBll activity was standardized using standard serum made taking units of Medical Research Council long-acting thyroid stimulator, standard B as a reference, and expression of TSAb activity was standardized using bovine TSH as a standard. TBll activity was positive in 22 mothers at delivery, and TSAb activity was positive in 18. In 12 cases, both activities were positive. Both the TBll and the TSAb activity of maternal serum at delivery correlated well with that of the cord serum. Neonatal thyrotoxicosis occurred in 9 of 108 neonates (8%), of whom five (5%) had clinical overt symptoms requiring antithyroid drug treatment. In all nine cases the TBll and TSAb activities were both positive, but no neonate without TBII or TSAb activity developed thyrotoxicosis. The prediction rate (42%) of neonatal overt thryotoxicosis was higher when both TBII and TSAb were measured than when only TBII (23%) or TSAb (28%) was measured. Clinical overt thyrotoxicosis could be predicted in five of six neonates (83%) of mothers when the cutoff levels of antibody activites were increased to a TBII activity of above 8 U/ml and TSAb activity of above 1.0 TSH .mu.UEq. As a more useful predictive criterion, a new index was introduced, the binding-stimulation (BS) index, which means the product of the TBII and TSAb levels. Neonatal overt thyrotoxicosis could be predicted in all cases irrespective of the antithyroid drug treatment when the value of this new index was above 15. Methimazole did not always cause neonatal transient hypothyroidism even wen given at a hgih dose or when the BS index was smaller. These data indicate that: prediction of neonatal thyrotoxicosis is more accurate and general when both TBII and TSAb are measured and the expressions of their activites are standardized; neonatal oert thyrotoxicosis can be predicted with certainty when the BS index is above 15, irrespective of the antithyroid drug treatment; and the occurrence of neonatal transient hypothyroidism due to methimazole does not depend simply on the dose of this antithyroid drug or the BS index.