Résumé : Objectives. To evaluate the short- and long-term effects of neoadjuvant hormonal treatment in locally confined prostate cancer. Methods. We report the preliminary results of 354 patients (199 with a clinical T2 tumor and 155 with a clinical T3 tumor) of whom 164 randomly received neoadjuvant total androgen deprivation using a luteinizing-hormone-releasing hormone (LHRH) analog (goserelin) plus flutamide for a period of 3 months. Results. Serum prostate-specific antigen (PSA) levels and prostatic volume decreased from a mean of 19.9 ng/mL and 37.7 cm3 to a mean of 0.8 ng/mL and 26.5 cm3 after 3 months of neoadjuvant therapy. 'Clinical down-staging' was seen in 32% in the neoadjuvantly treated group. 'Pathological downstaging' percentages were 6% and 16% in the direct radical prostatectomy group and neoadjuvantly-treated group, respectively (P <0.01). In patients with clinical T2 tumors, a significant difference in number of positive margins was shown in favor of the neoadjuvantly treated group (P <0.01). In patients with clinical T3 tumors, a significant difference could not be detected (P = 0.14). In 215 patients with a mean follow-up time of 15 months, the calculated 95% confidence intervals of mean time of PSA progression-free survival were 26 to 35 months in the neoadjuvantly-treated group and 28 to 37 months in the direct radical prostatectomy group, indicating no significant differences between treatment groups. However, follow-up time is currently too short to draw definite conclusions. Conclusions. These early data confirm high understaging percentages in clinical staging. The clinical relevance of the statistically significant smaller numbers of patients with positive margins in the neoadjuvantly treated group with a clinical T2 tumor will have to be confirmed when further follow-up allows an accurate evaluation of time to PSA progression, local recurrence, and distant metastases. Presently, neoadjuvant therapy is not advisable outside clinical research settings.