The objective of the trial is to evaluate the efficacy of capecitabine in patients with metastatic hormone-resistant prostate carcinoma (HRPC), in terms of prostate-specific antigen (PSA) response and clinical benefit (decrease of pain or analgesic score) and its safety profile. In all, 25 patients with HRPC were enrolled on a phase II trial of capecitabine (Xeloda) at a dose of 1250 mg m(-2) orally twice daily on days 1-14 every 21 days. The inclusion criteria were PSA serum levels >3 x upper limit of normal, a WHO performance status 0-2, age <85 years and adequate bone marrow, liver and renal function. In patients with grade 2 or higher haematological toxicity on day 1 of the treatment cycle, therapy was first delayed, and then continued at a lower dose. Trial end points were PSA response and clinical benefit defined by quality of life (QL) data and analgesic consumption. The median age of patients was 70 years (range 54-85 years). A median of three cycles of capecitabine was administered (range 1-8). PSA response was observed in three patients (12%, 95% CI 3-31%), with times to tumour progression of 18, 21 and 35 weeks, respectively. In these patients, the response durations were 12, 17 and 32 weeks, respectively. Minor PSA regression was also seen in two further patients. The median time to tumour progression of all patients was 12 weeks (95% CI 9-15 weeks). Haematological toxicity was minor, with leukopenia grade 3 observed in one patient. There were three deaths during trial treatment, respectively, due to sepsis following mucositis and leukopenia, presumed sepsis with mucositis induced by chemotherapy and concomitant radiotherapy and cerebral dysfunction progressing to coma. Hand-foot syndrome grades 2 and 3 were observed in four patients each. Clinical benefit was observed in five patients (20%, CI 7-41%). Based on toxicity data, we recommend a lower starting dose of 1000 mg x m(-2) orally twice daily. While capecitabine has some activity in HRPC, as suggested by observed PSA responses, we conclude that it is not worthwhile to investigate capecitabine monotherapy in a phase III trial. Combinations of capecitabine with other agents, such as vinorelbine or docetaxel, may prove to be more effective.