BERKELEY, CA (UroToday.com) - The investigators were surprised to see only a very small increase in the relative reduction of prostate cancer mortality from 20% to 21% but were delighted that the level of evidence increased from 0.04 in 2009 to 0.001 in 2011.
Attempts are ongoing to try to explain the reasons why the mortality curves do not separate more during the added period of two years. Our data so far suggest that in spite of excluding prevalent prostate cancer prior to randomization, there are men who have rather advanced prostate cancer at the time of entry into the study and that we are witnessing an effect of the treated natural history of these cases which seem to die around this period of follow-up. Obviously, the group is very curious how further follow-up will change this situation.
The abstract cites a significant rate ratio of 0.62 for the time period 10-11. Obviously, this translates into a relative risk reduction during that time period of 38%. Again, the investigators were surprised about this finding. Even with a median follow-up of 11 years, follow-up must be assumed to be incomplete during the period immediately preceding the median. The effect of possible incomplete follow-up during this period cannot be excluded at this time. The data may change with more follow-up.
The abstract reports that 1,055 men needed to be invited for screening and 37 cancers needed to be diagnosed and treated in order to prevent one death from prostate cancer. These data are based on a population with follow-up restricted to 11 years. On page 985 of the manuscript we state that if we consider all follow-up in a non-truncated analysis the number needed to invite (NNS in the previous publication) and to diagnose (NNT in the previous publication) is 936 and 33. These figures show a marked improvement compared to the NNS of 1,410 and the NNT of 48 in the 2009 publication, which amounts to 31% for the NNT figures.
Obviously, our manuscript does not address the downsides of screening except for showing a large difference in incidence between the screen and control arm, suggesting overdiagnosis. The difference between 6,963 PC diagnosed in screening and 5,396 diagnosed in the control arm amounts to 22.5%. Obviously, it remains unclear what proportion of this difference can be earmarked as overdiagnosis. As in previous publications and comments, I have to state that the analysis of quality of life and quality of life adjusted life years is still pending. Our manuscript is under review.
Finally, I should like to address a remark made in the editorial comment which is also brought forward by a number of other prominent comments on our study: ERSPC does not show a difference in overall mortality. Our study has not been designed for this purpose. In our power calculation, which is cited in all our reports, the endpoint and the resulting power are clearly determined. Our study group wishes to contribute to the understanding of screening with respect to saving lives by reducing the proportion of prostate cancer deaths. This is what we are doing. In this way we are contributing to a common goal of all public health services in most western countries, the decrease of the proportion of prostate cancer deaths contributing to overall mortality.
Written by:
Fritz H. Schröder, MD as part of Beyond the Abstract on UroToday.com. This initiative offers a method of publishing for the professional urology community. Authors are given an opportunity to expand on the circumstances, limitations etc... of their research by referencing the published abstract.
Prostate-cancer mortality at 11 years of follow-up - Abstract
