The dilemmas of prostate cancer screening

To the Editor: Recent articles in the Journal reflect the continuing polarisation of the debate on prostate-specific antigen-based screening and the extent to which major clinical trials reveal whether lives are saved by intervention and/or watchful waiting.1,2All clinical trials have limitations, and the ERSPC (European Randomized Study of Screening for Prostate Cancer)3 and the United States PLCO (Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial) are no exception.3 ,4 The calls by two senior authors of the ERSPC trial to evaluate only high-quality studies and to avoid pooling heterogeneous data are commendable.1 However, to then argue that only their trial satisfies these criteria is scientifically questionable, since this multi-country trial is compromised precisely by this action.

In the ERSPC, prostate cancer mortality data from Finland, Italy, Switzerland, Belgium and Spain showed minimal mortality differences between the screening and control arms, and were congruent with the PLCO data, in which no lives were saved by intervention. The data from Sweden and the Netherlands, however, were clearly divergent from other European countries and yet were inexplicably pooled with them. Further, no consideration was given to the equally plausible alternative of an increase in prostate cancer mortality in some control cohorts owing to their differential treatment. This would also yield fewer deaths in the screening arm, but not because of lives saved by intervention. It is therefore still premature to so boldly claim that significant numbers of lives are saved by screening. In terms of avoiding harm and maximising hope, is it not prudent to more closely and carefully scrutinise such crucial aspects of clinical trial data?