Previous work has shown that treatment with thyroid hormone (T3) decreased the incidence of rat hepatocellular carcinoma (HCC). The present study was designed to determine whether the inhibitory effect of T3 on HCC development was limited to early steps of the carcinogenetic process or, whether a similar effect could also be exerted by starting T3 treatment at later stages. Hepatic nodules were induced in Fischer rats by a single dose of DENA, followed by a 2-week exposure of the animals to 2-AAF and partial hepatectomy. Rats were then divided into 3 groups: group 1 was maintained on basal diet: group 2 was fed a diet containing 4 mg/kg T3 for a week, every month/7 months, starting 9 weeks after DENA administration: group 3 was exposed to cycles of T3 starting 8 months after initiation. Results demonstrate that inhibition of HCC development was essentially similar in rats exposed to T3 starting either 9 weeks or 8 months after initiation (50% inhibition compared to control rats). We have previously shown that T3-induced nodule regression and HCC inhibition occurred in spite of its mitogenic effect. Therefore, we next wished to determine whether a similar antitumoral effect could be exerted by other liver mitogens, such as peroxisome proliferators. Rats exposed to the initiation-promotion protocol described previously, were subjected to 11 cycles of a T3 or a ciprofibrate-supplemented diet, each cycle consisting of 7 days/month: the incidence of HCC and lung metastases was determined 13.5 months after initiation. Results showed that although treatment with T3 strongly inhibited HCC development (only 31% of T3+ rats showed HCC vs 91% of controls), rats given ciprofibrate developed the same number of HCC as T3-untreated rats. In conclusion, the results of this study showed that the anticarcinogenic effect of T3 is maintained also when treatment begins late in the process, and its antitumoral property appears to be specific and may not be shared by other liver mitogens.