Watchfulness may not be as good as surgery for prostate cancer

Editor—Gottlieb's news story that watchful waiting is as good as surgery in prostate cancer raises three issues.¹

Firstly, the conclusion would be possible only if the study had been specifically designed as a non-inferiority or equivalence trial and the resulting confidence limits lay within a prespecified and widely accepted interval.

Secondly, the primary end point is given as death from prostate cancer. The result of the study was roughly a halving of the risk of death from prostate cancer during follow up, the 95% confidence interval for the hazard ratio being 0.27 to 0.91. Taking the results at face value, this is direct evidence that radical prostatectomy is actually superior to watchful waiting—contrary to the headline.

Finally, because all cause mortality was specified as a secondary end point, the fact that the overall mortality was similar cannot be taken as evidence that the treatments are equivalent—especially as all cause mortality is also known to be an insensitive measure of efficacy when the target disease is responsible for only a small fraction of all deaths in the study population. This is why screening trials use cause specific mortality rather than all cause mortality as an end point.

Clearly other factors (quality of life and costs) need to be considered before radical prostatectomy can be routinely offered instead of watchful waiting, but these were not the primary aim of the study. I wonder whether such a misleadingly negative statement would have been acceptable if a treatment effect of the same magnitude had been observed in a breast cancer trial, for example?