Do Treatment Differences between Arms Affect the Main Outcome of ERSPC Rotterdam?

Purpose We assessed differences in treatment between the screening and control arms of ERSPC Rotterdam and studied whether possible treatment differences could explain the positive study outcome.
Materials and Methods In ERSPC Rotterdam men 55 to 74 years old were randomized to a screening arm of 21,210 and a control arm of 21,166. Treatment after diagnosis was at the discretion of the care provider chosen by the patient. Initial treatment was compared in 4 risk groups. The relation between prostate cancer incidence and prostate cancer mortality was assessed by risk group by correlating the incidence RR and the mortality RR. A direct relation would have supported a stage shift as the main cause of changes in prostate cancer mortality.
Results Initial treatment differed between the arms in the low, intermediate and high risk groups but not in the metastatic group. The RRs of prostate cancer incidence and mortality per risk group were related 1:1 (regression line slope 1.00, 95% CI 0.30-1.74). Of changes in prostate cancer mortality 94% could be explained by changes in prostate cancer incidence. This made treatment differences unlikely as the reason for the observed decrease in prostate cancer mortality.
Conclusions Differences in treatment between the ERSPC Rotterdam screening and control arms were unlikely to explain the differences in prostate cancer mortality. Results are instead consistent with a decrease in prostate cancer mortality as the result of a favorable stage through screening.