This is discussion forum for physicians, researchers, and other healthcare professionals interested in the epistemology of medical knowledge, the limitations of the evidence, how clinical trials evidence is generated, disseminated, and incorporated into clinical practice, how the evidence should optimally be incorporated into practice, and what the value of the evidence is to science, individual patients, and society.
Friday, August 2, 2013
Sause for the Goose, Sauce for the Gander: Low Tidal Volume Ventilation in the Operating Theatre
PIBW is based on height, not weight.
Following my usual procedure, I read the title and abstract of the methods of this article on Intraoperative Low Tidal Volume Ventilation in this week's NEJM, and I made a wager with myself on what the outcome would be. Because there are both biological plausibility and biological precedent for low tidal volume, and because it is one of the few interventions in critical care in which I have supreme confidence (yes, you can conclude that I'm biased), my prior probability for this intervention is high and I wagered that the study would be positive. If you have not already done so, read the methods in the abstract and make your own wager before you read on.
This trial is solid but not bombproof. Outcomes assessors were blinded and so were post-operative care providers, but anesthesiologists administering tidal volumes were not. Outcomes themselves, while mostly based on consensus definitions (sometimes a consensus of collective ignorance), are susceptible to ascertainment and misclassification biases. The outcome was a composite, something that I like, as will be elaborated in a now published letter in AJRCCM. A composite outcome allows an additive effect between component outcomes and effectively increases study power. This is essential in a study such as this, where only 400 patients were enrolled and the study had "only" 80% power to detect a reduction in the primary outcome from 20% to 10%. As we have shown, detecting a difference of this magnitude in mortality is a difficult task indeed, and most critical care studies seeking such a difference are effectively underpowered. How many effective (in some aspect other than mortality) therapies have been dismissed because of this systemic underpowering in critical care research is anybody's guess.
A curious selection of part of the primary composite is the sepsis component. Post-operative sepsis can arise from any source, and the biological plausibility of, say, a reduction in urosepsis or wound infection from intraoperative low tidal volume would be biologically dubious. Fortunately, as seen in Table 3 and in the KM curves in the Supplementary Appendix, the pulmonary complications component of the primary outcome carried the bulk of the weight of the composite.
The intervention in this trial was multifaceted, which will lead to the usual groans that we don't know if there's a magic ingredient (Low tidal volume? PEEP? Sighs/recruitment maneuvers?) or if the whole recipe is necessary for the effects to be manifest. As always, there were inclusion and exclusion criteria and a large proportion (40%) of the patients had pancreaticoduodenectomy, so some will moan about external validity and generalizability too.
In spite of these nuances, the primary results of the trial speak for themselves. There were highly clinically significant and statistically robust reductions in post-operative pulmonary complications (and other outcomes). Moreover, the caveats of the study design appear to be trivial. Because this intervention has few to no costs, and no apparent downsides, we appear to get a free lunch here. My biggest worry is that those who need this free lunch the most will not show up to the luncheon. Old habits die hard. So the next time you see your surgeon and anesthesia colleagues, make sure they know that what's sauce for the goose is now sauce for the gander too. And this sauce is free.
[The chart for tidal volume based on height can be found here.]