OBJECTIVE To define the characteristics of serum prostate-specific antigen (PSA) in a population of healthy men without clinically evident prostate cancer, but who are at risk for developing the malignancy. The influence of patient age and prostatic size on the serum PSA concentration was assessed in order to use PSA more appropriately to detect clinically significant prostate cancer at an early, potentially curable stage. DESIGN Prospective, community-based study. PARTICIPANTS Between December 1989 and March 1991, 2119 healthy men aged 40 to 79 years from Olmsted County, Minnesota, were entered into a prospective study to assess the natural history of benign prostatic hyperplasia. Of these, 537 (25%) were randomly chosen to participate in a detailed clinical examination that included a serum PSA determination (Tandem-R PSA assay), digital rectal examination, and transrectal ultrasonography. Four hundred seventy-one (88%) completed the prostatic evaluation and had no evidence of prostate cancer by any of these three diagnostic tests; these men formed the study population on which all analyses were performed. MAIN OUTCOME MEASURE Serum PSA concentration, prostatic volume, and PSA density (serum PSA level/prostatic volume) as a function of patient age. RESULTS The serum PSA concentration is correlated with patient age (r = .43; P < .0001) and prostatic volume (r = .55; P < .0001). Prostatic volume, in turn, is directly correlated with patient age (r = .43; P < .0001), whereas the PSA density value is only weakly correlated with patient age (r = .25; P < .001). For a healthy 60-year-old man with no evidence of prostate cancer, the serum PSA concentration increases by approximately 3.2% per year (0.04 ng/mL per year). The recommended reference range for serum PSA (95th percentile) for men aged 40 to 49 years is 0.0 to 2.5 ng/mL; for 50 to 59 years, 0.0 to 3.5 ng/mL; 60 to 69 years, 0.0 to 4.5 ng/mL; and 70 to 79 years, 0.0 to 6.5 ng/mL. CONCLUSIONS The serum PSA concentration is directly correlated with patient age and prostatic volume, the latter of which also is directly related to age. Thus, rather than rely on a single reference range for men of all age groups, it is more appropriate to have age-specific reference ranges. These age-specific reference ranges have the potential to make serum PSA a more discriminating tumor marker for detecting clinically significant cancers in older men (increasing specificity) and to find more potentially curable cancers in younger men (increasing sensitivity).