Camptothecin, a naturally occurring inhibitor of the DNA-replicating enzyme topoisomerase I, demonstrated promising anti-tumor activity in pre-clinical testing; however, because of unexpected toxicity and low anti-tumor effects in the initial clinical trials, further testing was discontinued. We hypothesized that local controlled delivery of camptothecin sodium would achieve effective concentrations in brain tumors without the observed systemic side effects, thereby allowing this novel drug to be used to treat patients with malignant gliomas. To test this hypothesis, we evaluated the sensitivity of rat glioma lines and established human glioma lines to camptothecin in vitro. We found that the LD90 for the established rat and human lines was 0.3 to 1.4 microM after a 1 hr exposure and decreased to less than 0.1 microM after continuous exposure for 7 days. We loaded camptothecin into a controlled-release polymer (ethylene-vinyl acetate co-polymer; EVAc) and showed by high-pressure liquid chromatography that controlled release occurred over at least 21 days. We then tested camptothecin against 9L gliosarcoma, implanted into the brain of Fischer 344 rats. Five days after tumor implantation, animals were treated with camptothecin delivered either systemically or locally by release from EVAc. Local controlled delivery by the polymer significantly extended survival: 59% of the treated animals were long-term survivors (> 120 days) compared to 0% of controls. Systemic administration did not extend survival compared to controls. We compared the efficacy of camptothecin delivered locally with a polymer to camptothecin injected directly into the tumor. Camptothecin increased survival only when delivered locally by polymer.