We previously have shown that urine components capable of stimulating ornithine decarboxylase activity of urothelium can enhance rat urinary bladder carcinogenesis, and that alpha-difluoromethylornithine (DFMO), an irreversible inhibitor of ornithine decarboxylase, suppresses carcinogen-initiated rat urinary bladder carcinogenesis. The present investigation was conducted to determine whether DFMO's suppressive effect is stage specific during carcinogenesis and whether the suppressive effect lasts with its continued use. Following initiation with 0.05% N-butyl-N-(4-hydroxybutyl)-nitrosamine in drinking water for 6 wk, male Fischer 344 rats initially weighing 125 to 150 g were randomly divided into two groups, the first receiving 0.2% DFMO in drinking water ad libitum and the second receiving tap water only. Groups of animals were killed at regular intervals until the completion of the experiment at 75 wk. The effect of DFMO was evaluated by monitoring the incidence of tumors, the mean number of tumors per rat, the mean volume of individual tumors, and the mean total tumor volume per rat. The results showed that continuous treatment with DFMO significantly reduced tumor formation until 60 wk (P less than 0.017). The effect was only of borderline significance (0.017 less than P less than 0.035) at 75 wk. Discontinuation of DFMO treatment at 40 wk resulted in the loss of protective effect in all comparisons except for the borderline effect on the tumor number and total tumor volume per rat. DFMO had no significant effect on the incidence or development of preneoplastic early lesions. Mucosal polyamine (spermidine and spermine) levels were reduced and correlated well with the reduction in tumor growth, suggesting that the reduction in tumor growth rate by DFMO may be due to its ability to reduce polyamine levels in urothelium. There were no side effects attributable to DFMO treatment. DFMO may be a useful chemopreventive agent to retard the recurrence of human superficial bladder cancer.