Previous studies demonstrated that single doses of systemic cyclophosphamide (CY) as low as 0.5 mg/kg are effective in preventing bladder tumor implantation in a rat model. In an effort to determine if the urinary bladder represents a unique site of CY activity, a series of experiments were performed to define the mechanism by which low-dose CY prevents bladder tumor implantation. Potential sites for CY antitumor activity include direct tumor cytotoxicity resulting from serum delivery of drug to the tumor; tumor cytotoxicity resulting from tissue drug levels at the site of implantation; altered tumor cell adherence to the urothelial injury site; nonspecific urothelial cytotoxicity resulting from urinary excretion of the CY metabolites; tumor cell-specific cytotoxicity resulting from urinary excretion of the CY metabolite acrolein; and second-pass cytotoxicity resulting from urinary excretion of the active form of CY. Experiments were performed to determine if a single predominant site of activity could be defined. Cyclophosphamide levels at the site of tumor implantation appeared to be the most important determinant of antiimplantation activity. Only tumor recipients pretreated with CY had a significant decrease in bladder tumor implantation. In vivo and in vitro assays measuring the effect of blood-borne drug delivery directly to the tumor failed to demonstrate cytotoxic activity. Tumor cell adherence assays measuring in vitro adherence of CY-treated tumor cells and in vivo adherence of tumor cells in CY-treated recipients showed no difference in comparison to control groups. Interval histological comparison of CY-treated and control bladders failed to demonstrate any difference. Urinary levels of acrolein did not contribute to antiimplantation activity. Preimplantation CY doses prevented tumor development in a s.c. implantation model, thereby excluding a second-pass effect resulting from urinary drug excretion. These data suggest that the bladder is not unique in its response to systemic low-dose CY administered for the prevention of implantation-mediated tumor recurrence. Low-dose, perioperative chemoprophylaxis may be applicable to other tumor systems in which intraoperative tumor dissemination is felt to contribute to recurrence risk.