Potential effects of age-specific reference ranges for serum prostate-specific antigen.

Abstract

Age-specific reference ranges for serum prostate-specific antigen (PSA) may improve this test for detecting prostate cancer. We have analyzed PSA levels from 10,024 men to determine the potential effects of these reference ranges. PSA levels (ng/ml) were grouped by patient age for comparison between standard (all ages: PSA < or = 4.0) and age-specific (< or = 49 years: PSA < or = 2.5; 50-59 years: PSA < or = 3.5; 60-69 years: PSA < or = 4.5; > or = 70 years: PSA < or = 6.5) reference ranges. Serum PSA correlated significantly with age (r = 0.33; p < 0.001). Fewer men > or = 60 years had elevated levels when age-specific reference ranges were applied (1,373 vs. 1,967; p < 0.001). Prostate biopsies and prebiopsy PSA levels from 865 men were reviewed. Sensitivities and specificities were calculated using both reference ranges. A significant increase in specificity with the age-specific reference ranges was seen for men > or = 70 years (58.6 vs. 34.2%; p < 0.001). There was, however, a concomitant decrease in sensitivity (77.6 vs. 91.7%; p < 0.001). We conclude serum PSA increases with age and we support the concept of age-specific reference ranges. However, the specificity of this test remains low, illustrating its limitations for prostate cancer detection.