In order to assess the efficacy and toxicity of dexamethasone as a single agent without the concomitant infusion of Adriamycin and vincristine (VAD), an ECOG pilot study was initiated using 40 mg by mouth daily for 4 days every week for 8 weeks. Patients who responded were then maintained on the same treatment, but at 2 week intervals. Of the 32 patients evaluable for response, three were completely refractory to all prior therapy. All patients had advanced disease and 26 had received multiple prior treatments. There were 13/32 (40%) objective responses by ECOG criteria. Of the 28 patients evaluable for subjective response, i.e., significant decrease in performance status and/or bone pain, eight (28.5%) responded. Of the 34 patients evaluable for toxicity, 19 patients (55%) had moderate to severe side effects, including nine who had central nervous system effects, three who had gastrointestinal bleeding, two who had pulmonary emboli, one with psychosis, and four who had serious infections with one death. Median survival for the entire group was 19 weeks, with 31 weeks in the responders and 9 weeks in the non-responders. Although high-dose dexamethasone is capable of producing a significant number of partial responses (40%), it is associated with excessive toxicity. Less frequent administration of the dexamethasone at 2 week intervals was well tolerated in the maintenance of partial response, but has not been studied for efficacy in induction of remission.