Population-based assessment of prostate-specific antigen testing for prostate cancer in the elderly

Objectives: To perform a population-based analysis to characterize the effect of prostate-specific antigen (PSA) testing on oncologic outcomes in men diagnosed with prostate cancer. Materials and methods: We used the Surveillance, Epidemiology, and End Results-Medicare-linked data to identify 98,883 men diagnosed with prostate cancer from 1996 to 2007. We stratified frequency of PSA testing as none, 1 to 2, 3 to 5, and≥6 tests in the 5 years before prostate cancer diagnosis. We used propensity scoring methods to assess the effect of frequency of PSA testing on likelihood of (1) metastases at diagnosis and (2) overall mortality and prostate cancer-specific mortality. Results: In adjusted analyses, the likelihood of being diagnosed with metastatic prostate cancer decreased with greater frequency of PSA testing (none, 10.6; 1-2, 8.3; 3-5, 3.7; and≥6, 2.5 events per 100 person years, P<0.001). Additionally, greater frequency of PSA testing was associated with improved overall survival and prostate cancer-specific survival (P<0.001 for both). Conclusions: Greater frequency of PSA testing in men 70 years of age or older in the 5 years before prostate cancer diagnosis is associated with lower likelihood of being diagnosed with metastatic prostate cancer and improved overall and prostate cancer-specific survival.