Delayed Massive Hemorrhage after Pancreaticoduodenectomy: A New Therapeutic Approach

Abstract

riority trial require that an initial difference in outcomes between groups be designated as a clinically relevant difference. Adhering to a strict interpretation of the results of the equivalency trial, one can only make conclusions with respect to the difference in outcomes initially designated. For our study, we chose 6% as that designated difference, to avoid a potentially meaningless difference in infection rate between groups, which might be statistically different but not clinically relevant. This was balanced with the sample size requirements for a smaller equivalence threshold. In Dr Lee’s example for comparing 9% infection rate for paint-only versus 4% in paint-plus-scrub, the sample size requirements would be approximately 650 patients per group (1,300 patients total). We have contributed to the literature a controlled clinical trial that was designed to test the idea that preoperative scrubbing of skin with povidone-iodine soap adds no incremental protection against wound infection. We do not agree with Dr Lee that our results have the same interpretation as a hypothetical trial, where the actual infection frequency was 9% in the paint-only arm and 4% in the paint-plus-scrub arm. Although Dr Lee’s worst-case outcomes would be within the tolerance of our 6% equivalence threshold based on our current results and the results of previous trials, that specific outcome would be unlikely given the data that has been reported as of this writing. Admittedly, Dr Lee is correct that the interpretation of our results has not added to the possible knowledge-space: povidone-iodine scrubbing might add benefit, might have no effect at all, or might actually increase infection likelihood. Statistical reasoning does not deal in perfect knowledge states. It can lead to conclusions that speak to the likelihood of outcomes. We would argue that our data have shown that the advantage of povidone-iodine scrubbing over paint-only is minimal at best because it is the most likely interpretation. Despite the results of even the most rigorously designed single clinical trial, the thoughtful clinician will always weigh the pros and cons of altering his or her clinical practice. Results of our single clinical trial must be interpreted in the context of the four earlier clinical trials available for review, all of which are cited in our publication. Each of these was a negative trial, and each trial compared the “gold standard” scrub-plus-paint to something different and often less than scrub-pluspaint. In view of this body of literature and our clearly negative clinical trial, an objective observer who is not lost in the vagaries of statistical minutiae, can conclude that a reasonable argument can be made for abandoning scrub-and-paint.