Just Added!

New Videos with Amal Mattu, MD

Watch NowGo

Ketamine and Post-Induction Hypotension Again?

May 28, 2021

Written by Clay Smith

Spoon Feed
After ketamine use became widespread for prehospital RSI in an Australian EMS system, there was an associated increase in the rate of post-induction hypotension.

Why does this matter?
We covered the NEAR articles that raised concern about ketamine and post-induction hypotension. EMCrit said – not so fast. Yet, here is another study that has found a similar association with hypotension and ketamine.

Seems like I have seen this before…
This was a retrospective study based in Australia that looked at 1,759 patients with presumed head injury who were intubated in the prehospital setting pre- and post-allowance of ketamine for induction in 2015: 792 patients pre-2015 and 827 patients after 2015. Pre-2015, 93-94% of patients were induced with fentanyl and midazolam. From 2015 to 2020, 92% received ketamine, progressively more and more each year. There was an initial increase in ketamine use >50% in January 2015 after the EMS policy change allowed it for RSI. They found an association with a 5% increase in post-intubation hypotension (SBP <90) immediately following more widespread use of ketamine in 2015, and the rates of hypotension increased over time as did the use of ketamine. Pre- and post-induction SBP dropped an average of 7.8 mm Hg after ketamine was introduced in 2015. This analysis is limited by its retrospective nature and may be confounded. However, even in the 2009 KetaSED RCT, there was a 5mm Hg drop in SBP in the ketamine group compared to the etomidate group, though this was not statistically different. Hypotension and ketamine is a concerning association that keeps popping up in the literature. Where there’s smoke, there’s [an association with] fire.

Another Spoonful
You should listen to Scott Weingart’s take on the NEAR studies that looked at ketamine vs etomidate and hypotension. I added some commentary to my coverage of the NEAR articles that you may want to read as well.

The association of ketamine induction with blood pressure changes in paramedic rapid sequence intubation of out-of-hospital traumatic brain injury. Acad Emerg Med. 2021 Mar 24. doi: 10.1111/acem.14256. Epub ahead of print.

14 thoughts on “Ketamine and Post-Induction Hypotension Again?

  • This is a quasi-experimental study, so confounding is unlikely to be a problem! that’s a strength of the study, so it’s not your typical retrospective study and the results are very believable.

    • Agree. I think it is likely that ketamine causes a dip in BP in very ill patients. Scott Weingart is correct that the NEAR studies (and this study) don’t prove causation. Also, the Jabre study did not show a statistically significant difference in SBP between the etomidate and ketamine groups, though this was not the primary outcome and the study was not designed to detect a difference or non-inferiority in this outcome. To truly test the impact on BP of etomidate or ketamine for induction, one would need to construct a randomized study powered to detect a difference (or non-inferiority) in BP as a primary outcome. Another question is what significance a small, transient decrease in BP has on patient-centered outcomes. There are at least three studies on ClinicalTrials.gov that are going to give us more information about this in the future.

      • It is a false belief that only randomized trials can "prove" causation. There are observational methods called quazi-experimental methods and natural experiments that can approach RCT for causality. For example, a well conducted instrumental variable analysis can show causality, and it is not a randomized experiment. Ditto with regression discontinuity designs. And this study discussed here is a type of regression discontinuity called "interrupted time series". It is not a historically controlled trial! As such, strong causal claims can be made here about the impact of ketamine on BP in the prehospital RSI population studied.

        While Weingart is sort off correct that NEAR studies don’t prove that ketamine causes blood pressure drops, we cannot ignore the NEAR studies either, especially since they seem to agree with the Jabre trial. And this study too agrees with the Jabre trial. Weingart is incorrect when he states that there was no BP difference found in the Jabre trial. There was, just not significantly. Absence of evidence is not evidence of absence…

        Clay, you ask "Another question is what significance a small, transient decrease in BP has on patient-centered outcomes?" Well this current study did not just show a small transient decrease in BP, but a significant 5% increase in hypotension in a TBI population! That worries me.

        Clay do you have a link to these new trials you spoke of? I am intrigued, high time someone does a trial!

        • PF, I hear what you’re saying. I searched on ClinicalTrials.gov and typed the search terms – ketamine and etomidate and induction – in the "Other terms" box and got three hits – a completed study by Driver et al (no difference in hypotension but with much lower recruitment than planned) and two others. I agree that it is not best to conclude that because there was no statistically significant difference in BP between etomidate and ketamine in the Jabre trial that, as Scott put it, "ketamine is at least as stable as etomidate." That may be true. He may be correct. But it may be that the sample size was not large enough to detect a significant difference in that outcome measure. As an aside, it would be ideal if your comments were not anonymous. It is important for readers to know who is writing. That way you can look us up and find out what qualifies (or in my case – disqualifies) us. I am just a guy whose niche is knowledge translation – trying, and sometimes failing, to get this stuff right. I have a high "SOFA score" because all I do is armchair quarterback the real heroes here – the patients who participate in these trials and the researchers who do all the hard work 🙂 . But I respect your anonymity if that’s what you prefer. Again, thanks for taking the time to comment. It makes JF so much better.

          • Clay, I see your point. I am Pieter Francsois Fouche the first author of this paper. I commented because I objected to the characterization of our paper as just another retrospective observational study, and that no causal inference are justified. This is not true. Our paper is a natural experiment, and strong causal claims can be made. Its an okay rule of thumb that "RCTs can prove causation and observational studies cannot". The reality is more nuanced than that. I believe that ketamine cause more hypotension and blood pressure drops compared to induction with midazolam/fentanyl.

          • First of all, let me just say congratulations for publishing such an important paper. Second, thanks again for taking time to comment. I see what you are saying, and I think this warrants more discussion. I hope we all have the chance to talk this out.

          • Clay, you state "Scott put it, "ketamine is at least as stable as etomidate." That may be true. He may be correct". Weingarts conclusion that ketamine is at least as stable as etomidate just does not follow logically from the Jabre study. He is wrong here. His non-sequitur arises from a failure to understand that a lack of statistical significance is the same as proving no difference. Not true! The Jabre did find a difference, it just didn’t reach statistical significance, which means that we cannot be very sure about that result. That’s all. it doesn’t mean that there was no difference. A lot of people listen to guys like Weingart, so he has an obligation to say the truth. And he did not do this in his podcast that you referred to.

          • I think we are tracking together on this. I am just not willing to be quite as firm on this point. You may be right as well.
            In fact, I lean more toward your interpretation of the Jabre study than Scott’s. However, if it’s one thing I have learned about the literature, it’s that it is hard to arrive at the truth. And I want to give people a lot of grace as they process and interpret the literature. I have seen so many high profile reversals over the years: rest for concussion, opioids for ACS, excessive oxygen, Pedialyte vs dilute apple juice for dehydration, therapeutic hypothermia vs targeted temperature management, and the list goes on and on. I think it is healthy for us to discuss and debate these things publicly, especially when we have a large number of people consuming what we put out there into the public sphere. And I really appreciate your point of view.

          • Many people think the Jabre trial proves that ketamine is safe. It doesn’t. I just shows convincingly that ketamine is safe compared to etomidate. But how would ketamine compare to other induction agents? Our paper strongly suggests there is potential problems when comparing to midazolam/fentanyl. We need a trial. A trial is the gold standard. But be careful to ignore observational evidence, they serve to warn us of problems! I agree that it is hard to arrive at the truth. And I welcome debates on this important topic! Good job with your blog Clay, I am a fan.

      • I hope you don’t mind, but I don’t agree with with you said on non-inferiority designs. Non-inferiority trials are NOT a design used to detect differences between groups, but "… is appropriate for evaluation of the efficacy of an experimental treatment versus an active control when it is hypothesized that the experimental treatment may not be superior to a proven effective treatment, but is clinically and statistically not inferior in effectiveness". To show a difference, a classic parallel trial would do the job.

        • Good point. My statement was poorly worded. What I was trying to say, with few words, was that it depends on the underlying assumptions as to how a person might design a trial. One way to design it would be to do it the usual way – to detect a difference – if one assumed etomidate was superior to ketamine in its hemodynamic stability (or vice versa). Another study design, if one assumed the two drugs were similar (lacking a clinically important effect on BP) with regard to hemodynamic stability, would be a non-inferiority approach.

  • Weingart made an error in assuming that the lack of statistically significant differences between ketamine and etomidate in the Jabre trial is the same as proving that there is no a actual differences. Which is not correct. The trial did find a 5 mmHg difference but it did not reach significance probably due to the small sample size of the study.

  • Pieter,

    You make some strong claims here that I think would benefit from a fleshed out discussion. Why don’t you come on the show to discuss. Reach out to me at emcrit.org/contact

    Tried to email you with the address on the study, but it doesn’t go through.

What are your thoughts?