Elevated prostate-specific antigen levels up to 25 years prior to death from prostate cancer.

OBJECTIVES We tested the hypothesis that prostate-specific antigen (PSA) levels would be higher among prostate cancer deaths as compared with controls over time in the 25-year follow-up of the Multiple Risk Factor Intervention Trial of participants ages 35-57 at entry. METHODS The initial stored serum samples were collected in 1973-1975 and the mean length of follow-up to prostate cancer death was 17 years. RESULTS There were 63 prostate cancer deaths and 63 controls matched by age, clinical site, and length of follow-up. The mean PSA level for prostate cancer decedents was 2.84 ng/ml as compared with 1.10 ng/ml for controls (P = 0.002 for difference). There were nine men who died from prostate cancer and no controls with PSA levels > 4 ng/ml. Risk of prostate cancer death increased with increasing PSA levels, with increased risk observed even at moderate levels of PSA. Many of those with high PSA levels in 1973-1975 died from prostate cancer many years after the elevated PSA. CONCLUSIONS PSA levels measured from blood obtained before the introduction of widespread PSA testing were a strong predictor of prostate cancer death over 25 years of follow-up. Studies of prostate cancer etiology and chemoprevention need to focus on middle-aged or younger men with longer follow-up.