Effects of Causal and Noncausal Components of Screening on Prostate Cancer Mortality: A Mediation Analysis from the Dutch, Finnish, and Swedish Centers of the European Randomized Study of Screening for Prostate Cancer

BACKGROUND AND OBJECTIVE: The European Randomized Study of Screening for Prostate Cancer (ERSPC) has shown that prostate-specific antigen (PSA)-based screening reduces prostate cancer (PCa)-specific mortality (PCSM). However, the mediating role of a stage shift and curative treatment in the causal pathway from screening to reduced PCSM has not been elucidated. METHODS: In the ERSPC trial, men were randomly allocated to either a screening arm, with regular PSA measurements, or a control arm. During 16-yr follow-up, 8046 men from Finland, 3701 from the Netherlands, and 1382 from Sweden were diagnosed with PCa and were aged 55-69 yr at the time of randomization. A directed acyclic graph was used to define the impact of screening on the PCa-specific restricted mean survival time (RMST) after randomization, mediated by risk group and treatment at diagnosis. KEY FINDINGS AND LIMITATIONS: The absolute increase in RMST over 16 yr for men in the screening arm versus the control arm was 0.13 yr (95% confidence interval [CI] 0.05-0.22) in Finland, 0.48 yr (95% CI 0.24-0.62) in the Netherlands, and 0.44 yr (95% CI 0.22-0.68) in Sweden. This effect was mainly attributable to risk group at diagnosis, and not through treatment at the time of diagnosis. A limitation of the study is that we only included the three largest ERSPC centers. CONCLUSIONS AND CLINICAL IMPLICATION: The reduced PCSM achieved by early PCa detection via PSA-based screening is caused by more favorable prognostic features and subsequent improved treatment outcomes. Our findings confirm that the observed results are a causal effect of screening.