Between January 1992 and September 1993, 813 patients with stage D2 prostate cancer were enrolled in a multicentre, double-blind (for antiandrogen therapy) trial and randomised to antiandrogen therapy with Casodex (bicalutamide, 50 mg once daily) or flutamide (250 mg three times daily) and to luteinising hormone-releasing hormone (LHRH) analogue therapy with Zoladex (goserelin, 3.6 mg every 28 days) or leuprolide (7.5 mg every 28 days). Time to treatment failure was the primary efficacy endpoint. At a median follow-up time of 49 weeks, there was a significant (p = 0.005) difference between groups in time to treatment failure in favour of Casodex plus LHRH analogue. Overall, 168 (42%) of 404 patients in the Casodex plus LHRH analogue group and 218 (53%) of 409 patients in the flutamide plus LHRH analogue group reached a treatment failure endpoint. Although a cause-specific treatment-failure analysis was not performed, the difference between groups in treatment failure attributed to adverse events (mainly diarrhoea) was evident primarily in the first 7 months of therapy. The difference between groups in treatment failure for objective progression was most evident after 1 year of therapy. With further follow-up (median time of 95 weeks), the result for time to treatment failure, although no longer statistically significant, were consistent with the previous finding of an improvement in time to treatment failure associated with Casodex plus LHRH analogue therapy. With a median of 95 weeks of follow-up, 34% of deaths had occurred. The survival analysis was not dissimilar between the 2 groups. At 49 weeks median follow up, the incidence of diarrhoea was significantly (p < 0.001) lower among patients in the Casodex plus LHRH analogue group. Diarrhoea led to withdrawal from therapy for 2 patients in the Casodex plus LHRH analogue group, compared with 25 patients in the flutamide plus LHRH analogue group. In conclusion, Casodex plus LHRH analogue is well tolerated and effective with an improvement in time to treatment failure over flutamide plus LHRH analogue. Survival was not dissimilar between the 2 treatment groups.