Patients with histologically proven carcinoma of the prostate (n = 186) were initially assessed and followed up according to the standardized protocol of the British Prostate Study Group, urologists from which contributed patients to this investigation. These patients were given either endocrine therapy or orchidectomy as first line treatment; the ratio of the number of patients receiving these two treatment was similar in each group of subjects compared for survival. Prognostic indices were derived for all patients and for those classified according to the presence (M1) or absence (M0) of metastases. The prognostic indices were derived from clinical and hormone data obtained at initial presentation. Whereas the degree of tumor differentiation and plasma testosterone concentrations were significant prognostic factors in both M0 and M1 disease, growth hormone was only significant in M1 patients, where age was also of borderline significance; elevated growth hormone, higher Gleason grade, younger age, and lower testosterone indicated a poorer prognosis in M1 patients. These findings indicated the feasibility of selecting a poor prognostic group of patients that may derive benefit from a more aggressive therapy.