This study was designed to investigate the relationship of serum prostate-specific antigen to prostatic size and hormonal stimulation. Seven patients with benign prostatic hyperplasia were treated for six months with nafarelin acetate and then followed for an additional six months. Nafarelin acetate is a potent luteinizing-hormone-releasing hormone agonist which causes reversible testosterone deprivation resulting in involution of the prostate. During therapy and follow up, serum prostate-specific antigen correlated with: 1) serum testosterone (p less than 0.001); 2) quantity of prostatic epithelium (p less than 0.001); and 3) prostatic size (p less than 0.05). Before therapy, serum prostate-specific antigen (mean +/- SD) was 0.43 +/- 0.2 ng./ml. per gram of epithelium. This did not change significantly after six months of androgen deprivation (0.48 +/- 0.36), although the ratios of prostate-specific antigen to testosterone and to prostatic size each changed significantly. Despite testosterone levels in the castrate range at six months, five of seven patients had serum prostate-specific antigen concentrations above the female range and three of seven patients had prostatic biopsies containing columnar epithelium which stained positively for prostate-specific antigen. These results demonstrate that serum prostate-specific antigen is related to prostatic size, prostatic epithelial weight, and testosterone stimulation. However, prostatic size is not a good predictor of serum prostate-specific antigen because there is tremendous variation in the relative amount of epithelium in a prostate; in this study the ratio of prostatic size to epithelial weight varied threefold. Furthermore, although testosterone determines prostatic size and amount of prostatic epithelium, it may not totally control prostate-specific antigen production.