Association between one-time prostate-specific antigen (PSA) test with free/total PSA ratio and prostate cancer mortality: A 30-year prospective cohort study.

Abstract

To explore if there is a long-term association between baseline prostate-specific antigen (PSA), including free/total PSA ratio and long-term (30-year) risk for prostate cancer death. In all, 1782 men were screened for prostate cancer through PSA analysis. Some years later, frozen plasma samples were used to calculate the ratio of free to total PSA (f/t PSA). At 30-year follow-up, baseline PSA and f/t PSA were compared with recent data extracts from the Swedish Cause of Death Registry and Swedish Cancer Registry. PSA values and f/t PSA values were treated as continuous variables in a multivariable analysis and also stratified according to their distribution and useful clinical thresholds. Risk of death from prostate cancer after 30 years of follow-up was significantly increased with a higher baseline PSA level, with the hazard ratio being 1.04 (95% confidence interval 1.03-1.09) per increase of one unit of PSA. Adding f/t PSA increased the model's ability to discriminate (concordance index 0.84-0.88). Men with PSA levels <1.0 ng/mL had a very low long-term risk of prostate cancer death (1.2% risk). An f/t PSA ≥ 0.25 extended the low-risk range to PSA < 2.0 ng/mL (1.5% risk). Prostate-specific antigen testing can be carried out less frequently or can be discontinued in men aged 55-70 years if their PSA levels are <2.0 ng/mL and the f/t PSA is ≥0.25.