Objectives. To evaluate the relationship between low prostate-specific antigen (PSA) levels that are considered normal and the long-term risk of prostate cancer.

Methods. The relative risk of, and cumulative probability of freedom from, prostate cancer by PSA level and age decade was evaluated in male participants of a longitudinal aging study, the Baltimore Longitudinal Study of Aging (National Institute on Aging). The relative risk was estimated from a Cox proportional hazards regression model for men aged 40 to 49.9 (n = 351) and 50 to 59.9 (n = 445). The disease-free probability was determined by Kaplan-Meier survival analysis.

Results. The relative risk of prostate cancer for men aged 40 to 49.9 was 3.75 (range 1.6 to 8.6) when the PSA level was at or greater than the median (0.60 ng/mL) compared with men with PSA levels less than the median. This risk was similar for men aged 50 to 59.9 when comparing those with PSA levels greater than and less than the median (0.71 ng/mL). At 25 years, the cumulative probability of freedom from prostate cancer for men aged 40 to 49.9 was 89.6% (range 81% to 97%) and 71.6% (range 60% to 83%) when the PSA level was less than and greater than the median, respectively. The 25-year disease-free probability for men aged 50 to 59.9 was 83.6% (range 76% to 91%) and 58.9% (range 48% to 70%) when the PSA level was less than and greater than the median, respectively.

Conclusions. The association between the baseline serum PSA level and the subsequent risk of prostate cancer suggests that the biologic events that predispose to prostate cancer begin early in middle age. Men who have baseline PSA levels that are “normal” but reflect a higher risk of prostate cancer may be the most appropriate candidates for future prevention trials. Those men with the lowest risk of prostate cancer on the basis of the baseline PSA measurements are unlikely to benefit from frequent PSA surveillance in an effort to detect prostate cancer early.