In order to evaluate the threshold dose of thyroid tumor-promoting effects of KA, male F344 rats were initiated with N-bis(2-hydroxypropyl) nitrosamine (DHPN; 2000 mg/kg body wt., single s.c. injection) and, starting 1 week later, received pulverized basal diet containing 0%, 0.002%, 0.008%, 0.03%, 0.125%, 0.5% or 2%KA for 20 weeks. Five rats each in the 0%, 0.125%, 0.5% and 2%KA groups were sacrificed at week 12, and 10 rats each in all groups at week 20. As an additional experiment, three groups without DHPN initiation received basal diet, a diet containing 0.5% or 2%KA for 20 weeks. The serum T4 levels were significantly decreased in the DHPN-initiated groups given 0.125%KA or more at week 12. No significant decreases in serum T3 levels were observed in the groups treated with DHPN+KA and a significant increase was evident in the 2%KA-alone group at week 20. Some rats in the DHPN+2%KA group at weeks 12 and 20 and the 2%KA-alone group at week 20 showed pronounced elevation of serum TSH. Thyroid weights were significantly increased in the DHPN-initiated groups receiving 0.5% and 2%KA at weeks 12 and 20 and in the 2%KA-alone group at week 20. Histopathologically, the incidences of focal thyroid follicular cell hyperplasias in the DHPN-initiated groups treated with 0.125%, 0.5% and 2%KA at week 20 were 5/10, 10/10 and 8/8 rats, respectively. At week 20, adenomas were observed in 7/10 rats in the DHPN+0.5%KA group and 8/8 rats in the DHPN+2%KA group, and carcinomas were observed in 6/8 rats in the DHPN+2%KA group. In the groups without DHPN initiation, only focal follicular cell hyperplasia was observed in 1/9 rats in the 2%KA-alone group. These results suggest that the no-observed-adverse effect for the thyroid tumor-promoting effect of KA is 0.03% (15.5 mg/kg/day) under the present experimental conditions, and that KA possesses weak tumorigenic activity in rats due to continuous serum TSH stimulation by a non-genotoxic mechanism.