Stenting as a bridge to surgery for colorectal cancer

We read with interest the article by Jeanin van Hooft and colleagues, and congratulate them on their accomplishment, especially in an emergency setting. However, this trial raises ethical and oncological dilemmas that deserve comment. First, in 1996, the American Society of Clinical Oncology stated that overall survival and quality of life should be regarded as the main patient outcomes in clinical oncology. Quality of life is particularly important when survival is equivalent for diff erent treatments. Van Hooft and colleagues did not justify the omission of survival data in their trial. Second, the trial was stopped early because of safety concerns; there was a high perforation rate, six of 47 patients (12·8%), and low technical success rate (70%, 33 of 47 patients) in the stenting group. The investigators discussed the various reasons for these poor outcomes. However, those in favour of the use of stents could argue that better external validity of the data is needed; do 20 stent placements represent suffi cient surgical experience? What is the mean number of stents placed per month per endoscopist? The uncertainty principle, or equipoise, is a fundamental scientifi c and ethical principle for randomised clinical trials. From the results of existing trials it would seem that uncertainty regarding colonic stenting versus surgery no longer exists because, to our knowledge, three of the four previous randomised clinical trials in this discipline were stopped for safety reasons. In these three trials, the cumulative mortality rates were similar after stenting as a bridge to surgery and surgery alone (7·9% [eight of 105 patients] vs 5·7% [six of 105], respectively). Systematic reviews of stenting show a median clinical perforation rate of 4% (range 0–83%). This rate seems to be acceptable to most authors. However, if a surgeon reported such a rate of surgical perforation, he would almost certainly need to retrain. Cancer perforation can jeopardise the future of patients by changing a potentially curable disease into an incurable one. In addition to the clinical perforation, researchers have reported a high rate of silent perforation, which could compromise survival even if this risk has not yet been shown. Taking into account that stenting can lead to a high perforation rate (and the oncological consequences of that), along with the unknown oncological outcome after stenting, van Hooft and colleagues’ proposal of further randomised trials is troubling. Trials in such a setting would not respect the uncertainty principle. Before a new trial is undertaken, we should know the eff ect of stenting on survival and local recurrences, which could be achieved by establishment of a registry to include survival data for all patients receiving colonic stents. We suggest that colonic stenting as bridge to surgery should no longer be done routinely until such robust survival data are available.