PSA screening and deaths from prostate cancer after diagnosis--a population based analysis

Abstract

Background: The United States Preventative Health Task Force recently recommended prostate specific antigen (PSA) screening be abandoned, believing the results of prior studies failed to show benefits that outweighed risks. Prior analyses did not include a complete 10 year follow-up in their analyses.

Methods: SEER rate sessions were used to obtain for U.S. White and Black men age-adjusted incidence rates for prostate cancer, in total and by loco-regional and distant (D2) spread for 1983-2009, as well as for prostate cancer diagnoses with associated prostate cancer deaths within 10 years of diagnosis (incidence based mortality rates) for 1983-1999. The SEER-Stat Program was used to tabulate rate estimates and calculate standard errors. The Joinpoint Regression Program was used to provide estimates and 95% confidence intervals (CI) of annual percent changes (APC) and times at which APC changed (joinpoints), as well as to test for parallelism to see if APC's differed between groups of rates.

Results: All analyses showed a 1991-1993 joinpoint, consistent with an impact of PSA screening. Between 1991 and 1999, incidence based mortality rates showed a decline for Whites of 10.9% (CI 9.2%-12.7%) and for Blacks of 11.6% (CI 9.7%-13.4%); incidence based mortality and D2 spread rate curves were similar (P > 0.05, test for parallelism).

Conclusion: Incidence based mortality declined by about 10% per year between 1991 and 1999 in a fashion similar to that of D2 spread, but not loco-regional spread or overall, incidence.

Keywords: PSA; mortality; prostate cancer; screening.