Controlled-release polymers have facilitated the interstitial delivery of drugs within the central nervous system. In the present study, dexamethasone was incorporated into ethylene-vinyl acetate polymers, which were then implanted adjacent to a 9L gliosarcoma in the brain of Fischer 344 rats. The effect of interstitial delivery of dexamethasone on peritumoral edema was assessed and compared to the effect of dexamethasone delivered systemically. Eighty-five rats underwent intracranial implantation of the 9L gliosarcoma. Five days later, the animals were randomly assigned to one of four treatment groups: Group 1 received intracranial implantation of controlled-release polymers containing dexamethasone; Group 2 received intraperitoneal implantation of controlled-release polymers containing dexamethasone; Group 3 received serial intraperitoneal injections of dexamethasone; and Group 4 received sham treatment. The animals were sacrificed 3 days after initiation of therapy and their brains were removed for measurement of the water content (edema) in the tumor-bearing and contralateral hemispheres. Brain and plasma samples were analyzed by reverse-phase high-performance liquid chromatography to determine the tissue and plasma concentrations of dexamethasone. Measurement of the release kinetics of dexamethasone from the ethylene-vinyl acetate polymers in an in vitro system showed that the drug was released in a controlled, tapering fashion. During the first 3 days of controlled release in vitro, 330 micrograms of a total content of 7.5 mg of dexamethasone was released into the medium. Analysis of tissue for drug levels demonstrated, however, that the interstitial delivery of this fractional amount of dexamethasone within the brain resulted in levels 19 times higher than those achieved by administering the full dose of 7.5 mg systemically over a 3-day period. Conversely, the systemic administration of dexamethasone resulted in plasma levels 16 times higher than those measured in the interstitial delivery of dexamethasone in the brain. Brain-water content determinations showed that the interstitial controlled release of the fractional amount of dexamethasone within the brain was as effective in controlling peritumoral edema as systemic administration of the full dose by serial intraperitoneal injections. The study demonstrates the following: 1) controlled-release polymeric carriers deliver biologically active dexamethasone in a sustained fashion; 2) very high concentrations of dexamethasone in brain tissue can be achieved using interstitial polymer-mediated drug delivery while minimizing plasma concentrations of this drug which are sometimes associated with serious systemic side effects; and 3) peritumoral brain edema can be effectively treated by the interstitial delivery of dexamethasone directly within the tumor bed.