Clift et al1 aimed to estimate prostate-specific antigen (PSA) contamination in the control arm of the CAP prostate cancer screening trial, concluding that there is ‘limited plausibility of deriving clear conclusions from trials of PSA screening’. However, their conclusion failed to acknowledge that the CAP intervention significantly increased prostate cancer detection: during the first 18 months following recruitment (the screening phase) there was a 5-fold increase in rate of prostate cancer detection — 10.42 per 1000 person-years in the intervention group versus 2.18 per 1000 person-years in the control group (P<0.001).2 Such a difference would be expected to lead to mortality benefits over long-term followup, but there was little evidence of any subsequent mortality reduction from earlier detection. Relying on how urinary symptoms are coded may overestimate opportunistic PSA screening.
In our analysis of 558 UK general practices,3 28% of men received a PSA test, but a raised PSA (≥3 ng/ml) was rarely followed with a prostate biopsy (6% of tests) or prostate cancer diagnosis (15%), as would be clinically expected for screening. In our trial, the corresponding figures were 85% undergoing biopsies and 34% diagnosed with prostate cancer. The CAP trial excluded London, the South East, and West Midlands. In the Clift et al paper, 21% of men were from London and PSA screening was 34% higher in London than the East Midlands; 46% higher in the South East; and 20% higher in the West Midlands. The CAP trial was overseen by independent, international Trial Steering and Data Monitoring Committees to ensure robust inference.
We believe the CAP trial conclusions remain important: single PSA screening detected significantly more prostate cancers compared with ad hoc testing but had no significant effect on prostate cancer mortality after 10-years’ follow-up.
- © British Journal of General Practice 2021
