I just read the RSI trial ("Randomized Trial of Sedative Choice for Intubation" - hereafter I will just call it the Random Sedation for Intubation trial). In this trial enrolling 2365 patients undergoing emergency intubation in 14 EDs and ICUs at 6 centers in the USA, the authors compared use of etomidate to ketamine for induction of anesthesia for laryngoscopy, using 28-day landmark mortality as the primary endpoint, a key secondary endpoint being hypotension ("cardiovascular collapse"). Wanna guess what they found?
Earlier trials (e.g., VASST and KETASED) raised concerns of adrenal insufficiency associated with etomidate in the ICU. This concern was first born in the 1980s when continuous etomidate infusions were being used in ICUs. Despite these possible adrenal effects, etomidate has become the drug of choice in most ICUs in the USA because of its observed salutary effects on peri-induction hypotension compared to alternative agents which were/are widely used in the operating theatre (previously sodium thiopental, now mostly propofol with or without supplementary use of sedative-hypnotics and opioids).
The authors found that there was no difference in mortality (not even a trend) and that the long-suspected salutary effects on hemodynamics are real: there was a statistically significant and clinically relevant 5% reduction in "cardiovascular collapse" (defined as any of the following occurring up to 2 minutes after intubation: SBP<65mmHg, new or increased dose vasopressors, or cardiac arrest). I think calling these outcomes "cardiovascular collapse" is a bit dramatic. Frankly, I don't care if I have to give some vasopressors, or increase their dose in the peri-intubation period. I do care if the patient has a cardiac arrest; there were 12 of those (1%) in the ketamine group, and 10 in the etomidate group (0.8%) a difference that was not significant - not even close. Table 3 has all the outcomes, and I really don't see anything suggesting any other important differences between the groups.
There are two tacks one could take with these data, one applying a decision making principle I have discussed on the sister blog about policies: if everything else was the same, and etomidate has less hypotension, a policy of routine use of etomidate is justified to accrue the small incremental benefits of not having to worry about hypotension quite as much. But just like my wife's policy of shopping at Costco (for its alleged cost savings), the policy can occasionally be violated for the sake of other desiderata. When she's in a hurry and only needs a couple of things, she can go to the corner store without much of a ding on the household finances. That is, the gain of less hypotension with etomidate is not so great as to forsake ketamine (or other agents) altogether.
Another tack, one that I'm likely to take, it to just choose whatever induction agent I think is appropriate for the circumstance, anticipating the possibility of hemodynamic instability with whatever agent is selected. "What's a little levophed between us girls?"
A final note about this trial and others like it: in essence the authors hypothesized (for power calculations), a 5% difference in landmark mortality between these agents. What drug do we give that, compared to another drug in the ICU, that changes mortality by 5%? (Almost) none, I propose. While this chosen delta is commendable, and I cannot accuse the authors of delta inflation like in the days of old, the prior probability of finding a 5% difference in landmark mortality in the ICU is very low. After all, what is the proposed causal mechanism of the effect? That a couple of minutes of hypotension (or transient adrenal insufficiency) translates to downstream mortality at 28 days? Seems highly unlikely. But alas, now that we have the results of the RSI study, we can rest assured: our choice of induction agent is not killing patients.
No comments:
Post a Comment