Sedation and analgesia V: Atypical sedatives

The fourth and last in a multipart series on the basics of ICU analgesia and sedation. Start with Part I: PrinciplesPart II: OpioidsPart III: Non-opioids, and Part IV: Typical sedatives.

Last time, we talked about some of the cornerstones of ICU sedation: propofol and benzodiazepines. Let’s finish up by discussing a few useful, but somewhat more idiosyncratic options.

Dexmedetomidine (Precedex)

This one is fun.

The first challenge with dexmedetomidine is the name you can’t pronounce (or spell). The secret lies, as it occasionally will, in organic chemistry:

  1. Take etomidine.
  2. Methylate it. Now you have medetomidine, an alpha-2 agonist similar to clonidine.
  3. Now take the more useful D-isomer, yielding dexmedetomidine.
  4. Wasn’t that easy?

The history of this started before my time, when clinicians prescribing clonidine as an anti-hypertensive noticed that it also yielded a gentle, useful sedative effect. Unfortunately, its antihypertensive (aka hypotensive) effects tended to preclude it from being used primarily in this role. So the eggheads got to work isolating a more precise form that emphasized the sedation while minimizing the hemodynamic impact—and thus came the IV agent dexmedetomidine.

Thus also came a lot of cost. The patent on this drug actually expired in 2016, but to my knowledge, a generic has not yet reached market. So this is not a cheap bottle of chill.

However, it may be worth the money, because it yields a very unique sort of sedation. Unlike propofol’s irresistible depression in consciousness that “takes away everything,” Precedex instead offers a calming hand which is relatively physiologic, and leaves the patient almost… normal. It’s the all-natural, grass-free, organic alternative.

At lower doses, it is simply an anxiolytic. At high doses, it induces a state that resembles normal sleep—and like sleep, it usually leaves patients rousable. They can be awoken, examined somewhat reliably, and allowed to fall back to sleep (a useful ability for neuro patients).

When it comes to delirium, it stands at the opposite end of the spectrum from benzos. A growing body of evidence suggests that dexmedetomidine tends to promote less delirium than other sedatives, and may even be an effective delirium treatment. This is a bit remarkable, since delirium tends to beget sedation (for safety) which begets more delirium, and a way to exit that cycle is welcome.

As one last difference from most sedatives, it usually does not suppress the respiratory drive. While it can be a good choice for intubated patients, even non-intubated patients can safely sit on a Precedex drip with little concern for their breathing. This makes it an especially appealing choice when patient anxiety is making extubation a challenge.

Certainly a useful arrow in the quiver. The disadvantages?

  • The other side of being a “gentle” sedative is that it can be too gentle. Some very agitated, wall-shaking patients simply won’t be controlled, even at maximal doses.
  • It is fairly hemodynamically stable. However, it can induce some degree of hypotension, mostly by causing bradycardia. Other than arbitrary dose ceilings, this phenomenon is often what limits dosage.
  • While it is purported to offer a degree of analgesia, it is not all that potent, and patients may still need a separate analgesic. And while it may induce some amnesia, it is not nearly as reliable as benzos.

Infusions can start around .1 mcg/kg/hour, with an upper limit determined institutionally, usually in the ballpark of 1.0-1.5 mcg/kg/hour. It can be initiated with a bolus for more rapid effect, but since bolusing is far more likely to cause significant bradycardia, this has become an uncommon practice.


Most often, etomidate—a sedative/amnestic with no analgesic effects—is reserved for procedural sedation, particularly RSI. Why?

It is a potent sedative with fast onset and short action. Thus, a push of etomidate yields very dense sedation lasting only a few minutes, followed by rapid emergence. This is exactly the desired profile for procedures requiring only a brief period of anesthesia (intubation, cardioversion, joint reduction). Dosing is around .3 mg/kg (classically about 10–30 mg).

Of course, propofol also has similar pharmacokinetics. The key difference? Propofol, as we’ve learned, is a reliable cardiovascular depressant. In hemodynamically unstable individuals—i.e. the shocky, underresuscitated septic or hemorrhagic patient—it can cause them to crash. You may be able to counteract this with some pressor (such as a push of phenylephrine), but you could also just choose a more hemodynamically neutral drug like etomidate.

(Of course, we must remember that even though etomidate has little intrinsic hemodynamic impact, it is still a sedative, and can still cause parasympatholytic-mediated hypotension in a patient with compensated shock. But it’s safer than slugging someone with the giant marshmallowy fist of propofol.)

The other risk with etomidate is inducing adrenal suppression. In bench research it has been shown to consistently depress cortisol levels, something sick patients are already susceptible to. This is particularly concerning in sepsis, one of our most common foes.

This phenomenon has led to etomidate, once used in continuous drips, becoming relegated to a purely procedural agent. Even in that role, it worries some of us, and its use remains controversial. On one hand, adrenal suppression has been demonstrated after even a single dose. On the other hand, this has not been shown to worsen outcomes and may simply be a lab abnormality. On the gripping hand, maybe it’s one more blow our sick patients don’t need.

Especially when you could just intubate them using…


We’ve already learned some of the secrets of ketamine, the trendiest dissociative analgosedative to reach a drug box near you.

As discussed, there is not (yet) much data on using ketamine for prolonged infusions at sedative doses. (Or at analgesic doses, for that matter.) We know that it packs a potent kick for short-term analgesia and sedation, but what if we leave it running for days? Will it have useful opioid-sparing effects? Will it reduce delirium—or will its famous dissociative trips put the patient at greater risk for it? We don’t know.

Which is not to say it can’t be used. But the main role for an ICU ketamine infusion at this point is probably either in patients with severe, refractory pain, or for sedation in patients hemodynamically unable to tolerate anything else.

This helps skirt one of its challenges, which is titration. Fixed, low-range analgesic dosing is fairly easy. However, if you increase the dose to produce ketamine sedation, you run the risk of entering the undesirable middle-range “partial dissociation” dose window which is a setup for agitation and psychological distress.

As a result, you probably cannot put a stable patient on a low-dose sedative ketamine infusion to “take the edge off their anxiety.” That would be a pharmacological oxymoron. The only truly safe way to offer ketamine sedation would be to go big and cause full dissociation. However, this violates our Second Rule of using the least possible sedation and leaving patients conscious whenever possible.

In a critically ill patient in florid shock, however, a low-to-moderate dose ketamine drip, combined with the natural sedative effects of their hypoperfusion, may be enough to keep them deeply sedated while you resuscitate. As their mental status improves, you can either drop it down to a pure pain dose, or just discontinue it and switch to a more conventional agent.

Ketamine is also probably one of the better options for intramuscular sedation. While the phenomenon of the strong, wildly combative, dangerous individual punching out nurses is more common in the ambulance and ED than the ICU, it does happen, and if IVs are not available, IM chemical restraint is the safest answer. Haldol, benzos, or the venerable B52 (50 mg Benadryl, 5 mg haloperidol, and 2 mg lorazepam, given IM from the same syringe) are all options, but IM ketamine has become popular due to its potency and relative safety from a cardiopulmonary perspective. (Ocasionally someone will worry about it raising the BP in a hypertensive patient, but usually these patients are hypertensive because they’re agitated, and improve with sedation, regardless of the type.) IM dosing is quite high, however, anywhere from 4–13 mg/kg (often 400–500mg), which requires a concentrated formulation to keep the injected volume manageable.


Older “typical” antipsychotics, usually haloperidol (Haldol), and the newer generation, most commonly quetiapine (Seroquel), are handy agents. Their antipsychotic effects are not what we’re after, not that anybody would ever ask us to manage a schizophrenic’s outpatient drug regimen anyway. However, these drugs do provide good, somewhat titratable sedation, which tends to be fairly safe; hemodynamic or respiratory compromise are infrequent. Haldol can be given orally or IM, but in the ICU tends to be delivered IV (usually 1–5 mg) for acutely agitated patients. Seroquel, on the other hand, has no parenteral formulation, so is only used orally (or down a feeding tube); as a result, it is less useful for acute episodes, but handy for chronic maintenance, and when administered once or twice a day on a scheduled basis (usually 25–100 mcg/dose), it has a calming effect. It’s not a bad sleep aid, either. Other agents such as olanzapine/Zyprexa or ziprasidone/Geodon are also around.

The appeal: these do not seem to promote delirium to the same extent as benzos. Despite what you may think, however, they are probably no better than anything else at treating delirium, the cure to which is not sedation but time and demedicalization.

The downside: QT prolongation. Unlike with many “QT prolonging drugs” (i.e. pretty much all drugs ever), this is a real effect. A QT interval should be checked prior to starting antipsychotics, if possible, and certainly after initiation at any significant dose. Their dopamine antagonism should also be remembered when treating patients with Parkinsonism and neuroleptic malignant syndrome.

A few other antipsychotic and idiosyncratic psychiatric meds are sometimes wielded in very hard-to-control patients, such as valproic acid, but in most cases these are best left to psychiatry.

This concludes our review of basic ICU sedation and analgesia. Questions? Thoughts? Put them in the comments below!