The gradual reduction of plasma testosterone in middle-aged and older men from mid-life onwards coincides paradoxically with the time when there is progressive growth of the prostate, a highly androgen-dependent organ. The growing interest in androgen therapy for older men makes it essential to understand the effects of exogenous testosterone on the non-diseased prostate, yet few studies are available. The present study examined prostate volume, prostate-specific antigen (PSA) and lower urinary tract symptom (IPSS) score in 207 men, aged 40–83 years, presenting with clinical features of age-related androgen deficiency [sexual and/or urinary dysfunction, elevated lutenizing hormone (LH)] who were treated for 6 months with oral testosterone undecanoate (TU). Men were divided into two groups, group 1 (n=92, plasma testosterone levels > 13 nmol/L) were treated with 80 mg daily; group 2 (n=115, plasma testosterone levels < 13 nmol/L) were treated with given 120 mg daily. Before treatment and after 1, 3 and 6 months of treatment, prostate volume was measured by ultrasound and hormones [testosterone, dihydrotestosterone, oestradiol, LH, follicle-stimulating hormone (FSH)] and PSA were measured. Within 1 month of treatment, the elevated blood LH levels were markedly decreased in all men in group 1, as well as most men in group 2. Group 2 was subdivided into men whose LH levels were suppressed (n=95, group 2a) and those whose LH levels did not suppress (n=20, group 2b). Men in group 1 and 2a had marked decreases in prostate volume, PSA and lower urinary tract symptom (IPSS) scores whereas no significant changes were observed in group 2b. Groups 1 and 2a also had more striking suppression of LH, FSH, dihydrotestosterone and oestradiol whereas group 2b had no significant increases in blood testosterone concentrations. These findings suggest that exogenous testosterone in middle-aged and older men with some clinical features of age-related androgen deficiency can retard or reverse prostate growth and that elevated plasma LH may be a useful index of severity of age-related androgen deficiency.