Thymidine (dThd) has been shown to increase the activity of BCNU in mice, possibly due to its ability to inhibit poly(ADP-ribose)polymerase (PADPRP), an enzyme thought to be active in DNA repair. The present phase I study characterized the pharmacokinetics and toxicity of dThd combined with BCNU. Sixty patients with refractory malignancies were infused with escalating doses of dThd from 7.5g/m2/day to 105.5 g/m2/day for 48 hr, along with 100 mg/m2/day of BCNU for 2 doses. Further dose escalation of dThd was limited by large fluid volumes required; therefore, the BCNU dose was escalated to a maximum of 160 mg/m2/day for 2 days. Plasma dThd concentrations were determined using high-performance liquid chromatography. At doses above 37.5 g/m2/day, steady-state concentrations of dThd approached or exceeded 1 mM, a concentration that nearly completely abolished BCNU-induced PADPRP activity in preclinical studies. Myelosuppression was consistent with BCNU dose but was not apparently increased by the coadministration of dThd. One patient had a partial response to therapy. Both the lack of effect of increasing dThd doses on BCNU-induced myelosuppression and the low response rate suggest that the schedule of drug administration was not optimal to inhibit PADPRP, or that PADPRP may not be essential in repairing BCNU-mediated DNA damage in humans.