PROSTATE-SPECIFIC ANTIGEN (PSA) AND PSA DENSITY - RACIAL-DIFFERENCES IN MEN WITHOUT PROSTATE-CANCER

Background: Many physicians now use serum prostate-specific antigen (P SA) to screen for prostate cancer in asymptomatic men. Whether or not a prostate biopsy should also be performed depends on an accurate defi nition of what constitutes a normal PSA value. Until recently, studies conducted to establish normal serum PSA values have involved study po pulations that have included few African-American men. Purpose: We sou ght to compare serum PSA levels and PSA density (i.e., serum PSA level /prostate volume ratio) in African-American and white men without hist ologic evidence of prostate cancer. Methods: We reviewed the medical r ecords of 826 consecutive men who underwent one or more prostate biops ies at the Veterans Affairs Medical Center in Shreveport, LA, from Jan uary 1993 through December 1995. In this retrospective review, we reco rded patient's age, race, serum PSA level, digital rectal examination result, ultrasound-determined prostate volume, indications for biopsy, and biopsy results. Data from a total of 752 consecutive men who were either white or African-American and whose indication for biopsy incl uded a serum PSA of greater than 4.0 ng/mL and/or an abnormal digital rectal examination were analyzed. To examine possible differences in s erum PSA level, PSA density, prostate volume, and patient age, the two -sided Student's t test was employed. Multivariate linear regression a nalysis was used to determine if serum PSA levels were associated with the patient's age, race, or prostate volume in men without prostate c ancer. Results: Of the 752 men included in this analysis, 254 had hist ologic evidence of prostate cancer and 498 did not. Of the 498 men wit hout prostate cancer, 367 (74%) men were white and 131 (26%) were blac k. There were no racial differences ire age or calculated prostate vol ume. Serum PSA levels and calculated PSA density, however, were signif icantly (both P<.0001) higher in African-American men than in white me n. A multivariate linear regression analysis indicated that race and p rostate volume were independent variables associated with serum PSA le vel. For African-American and white men, serum PSA values of greater t han 4 ng/mL were associated with prostate cancer with sensitivities of 89.5% and 81.9%, respectively, and specificities of 38.2% and 52.3%, respectively. Conclusion: Among biopsied men without histologic eviden ce of prostate cancer, African-Americans have a significantly higher P SA level and PSA density than similarly aged while men. Implications: Published criteria for normal PSA level and density have been derived primarily from white men and may not be directly applicable to other p opulations. Race-specific data are needed to fully optimize PSA as a t umor marker in racial populations that are at high risk for prostate c ancer death.