So we carry it. We use it (1) to augment analgesia at 0.2 mg/kg; (2) for conscious sedation at 0.5 mg/kg, and (3) for RSI at 1.5 mg/kg * Although I would consider reducing the dose slightly if they're severely hypotensive or sympathetically-driven.
We pretty much use it for all our RSI. The opinion of our medical directors is that it's acceptable in closed head injury and status epilepticus, despite historical concerns. We used to have the option for fentanyl / midazolam and only use ketamine for hypotensive patients, but we've now been directed to use ketamine for pretty much everything, with the exception of perhaps a malignantly hypertensive patient.
When you use ketamine, you have to be aware that it doesn't have a traditional dose-response curve. You tend to see a "staircase" effect, where you give a small dose and get analgesia (with a bit of altered mental status), then hit a state of dissociation, then hit an anesthetic level. It seems that you hit each plateau at a fairly arbitrary level.
Anecdotally:
* I have limited experience with using it for pain control in subdissociative dosing.
* I've used it a few times for conscious sedations in bad ortho injuries, including a pediatric femur fracture I discussed in one of the earlier threads. If we're pacing we tend to use aliquots of fentanyl in preference (there's some theoretical concerns about negative inotropy -- ketamine has a complex pharmacology, it's an indirect sympathomimetic, but a negative inotrope), and for cardioversions, if they have a pressure we use fentanyl / midazolam. It's worked very well.
* For RSI, I love it. Etomidate isn't an option here. It's nice to have an agent with a very rapid onset. I've seen a lot of RSI's go sideways, because someone gave fentanyl and midazolam, then didn't wait long enough for them to fully take effect. In practice, you can give the ketamine, wait 30-60 seconds, then push the succinylcholine. Our current guidelines are to use fentanyl / midazolam for ongoing sedation (rocuronium is an option for paralysis, but rarely needed), if they are normotensive, or repeat ketamine if they remain (or become) hypotensive. I've talked with some physicians who feel that the effects of ketamine are insufficient in isolation, and that sometimes the addition of fentanyl offers better analgesia.
* We don't currently prophylacticaly treat with benzodiazepines to prevent emergence reactions. This hasn't been a problem for me yet. I haven't heard of any bad situations from others, but this is no guarantee that this hasn't happened.
* We still use haloperidol / midazolam for chemical restraint. I'm curious about the use of ketamine, but have seen some fairly high intubation rates in the literature, e.g. 14/51 in
http://www.ncbi.nlm.nih.gov/pubmed/25455046
