Association of reductions in PSA screening across states with increased metastatic prostate cancer in the United States.

228 Background: While PSA screening was found to reduce prostate cancer metastasis and mortality in a large European randomized trial, PSA screening has also resulted in over-treatment of prostate cancer with significant quality-of-life implications. As a result, the US Preventive Services Task Force (USPSTF) did not recommend PSA screening in 2008 and 2012. It is unknown if reductions in PSA screening were responsible for increased metastatic prostate cancer in the United States. We test this hypothesis by associating longitudinal variations across individual states in PSA screening with their incidence of metastatic prostate cancer at diagnosis from 2002 to 2016. Methods: Age-adjusted incidences of metastatic prostate cancer at diagnosis per 100,000 men were obtained from the North American Association of Central Cancer Registries in 2002 – 2016 for each state. Survey-weighted PSA screening estimates for each state were extracted from the Behavioral Risk Factor Surveillance System, which collects this information for men at least 40 years of age every 2 years from 2002 onward. PSA screening and metastasis data were collated as a multi-panel time series and then analyzed using a random-effects linear regression model with random effects at the state level. Results: There was significant variation between states in the percent of men age >40 years who reported ever receiving PSA screening (range 40.1% to 70.3%) and in the age-adjusted incidence of metastatic prostate cancer at diagnosis (range 3.3 to 14.3 per 100,000). From 2008 to 2016, the mean percentage of men screened decreased (61.8% to 50.5%) whereas the mean incidence of metastatic prostate cancer at diagnosis increased (6.4 to 9.0 per 100,000; Bonferroni adjusted p < 0.001 for both). A random-effects linear regression model demonstrated that longitudinal reductions across states in PSA screening were associated with increased metastatic prostate cancer (regression coefficient per 100,000 men: 14.9, 95% CI 12.3 – 17.5, p < 0.001). This indicated that states with larger declines in PSA screening had larger increases in the incidence of metastatic prostate cancer at diagnosis. Variation in PSA screening explained 27% of the longitudinal variation in metastatic prostate cancer within states. Conclusions: In the context of randomized trial data demonstrating a metastasis reduction with PSA screening, our study strengthens the epidemiologic evidence that reductions in PSA screening may explain some of the recent increase in metastatic prostate cancer at diagnosis in the United States. The trend of rising metastatic disease at diagnosis is a worrisome consequence that needs attention. Thus, we support shared-decision making policies, such as the 2018 USPSTF update, that may optimize PSA screening utilization to reduce the incidence of metastatic prostate cancer in the United States.