Dexmedetomidine

In summary:

If one were limited to recommending only one article to the revising exam candidate, one would have to recommend Weerink et al (2017), as everything required to pass the CICM dexmedetomidine questions is in this paper. Additionally, one merely needs to enter its name into the a search engine to get a half-dozen articles titled simply "Dexmedetomidine", all of which would probably contain different phrasing of the same information. 

Chemical class and chemical relatives of dexmedetomidine

Dexmedetomidine is an imidazoline derivative that acts as a central α2 catecholamine receptor agonist. This sedating effect places it to the side of the more classical agents such as the benzodiazepines and barbiturates, because GABA is not involved in any of the sedation.  Its closest chemical relatives are medetomidine and detomidine, which are veterinary anaesthetic agents (specifically, dexmedetomidine is the S-enantiomier of medetomidine).  Together with xylazine and clonidine, the 'tomidines form a clique of centrally acting presynaptic adrenoceptor agonists, united by their eccentric sedating mechanism.

It feels like this group of drugs should have some special name, but "imidazoline derivatives" is all you tend to see in the literature. That is unfortunate, as this purely chemical classification does not really reveal anything about their mechanism of action, and also envelops many other pharmacologically active substances, most of which have absolutely no sedating properties (for example many are antifungal drugs or detergents), and some of which have to completely opposite properties (eg. oxymetazoline, a vasoconstrictor used as a nasal decongestant). The sedative imidazoline drugs are characterised by their tendency to bind to the imidazoline receptor, a CNS receptor site that does not seem to have anything to do directly with their α2 agonist activity, and which has no apparent endogenous ligand. Or, rather, none of the proposed endogenous ligands have any imidazoline in them (Bousquet et al, 2020), and the receptors themselves are not very well understood, aside from the fact that their activation seems to have a sedative and analgesic effect with a depression of sympathetic nervous system function.  

Pharmacokinetics of dexmedetomidine

The pharmacokinetics of this drug are probably less interesting from an exam perspective than the pharmacodynamics and side effects, but they are probably still worth revising, considering especially that it is usually administered as a prolonged infusion (i.e. there's plenty of material here to discuss, including the context-sensitive half time and the various chemical properties of dexmedetomidine which contribute to its pharmacokinetic characteristics) 

Administration and absorption of dexmedetomidine

Dexmedetomidine is only licensed for intravenous use. However, that has not stopped people from experimenting with alternative routes of administration for "uncooperative children or geriatric patients", which sounds like a terrible parenting style. Antilla et al (2003) convinced twelve suggestible young men to undergo a series of different dexmedetomidine challenges (orally, buccally, IM) and found that though first-pass effect decreased oral bioavailability to 16%, the drug was well absorbed through the oral mucosa (82% of the dose ended up in the circulation). It can also be given as a nasal spray (84 mcg per squirt, in the experiments by Iirola et al, 2011).

Solubility of dexmedetomidine

The pKa of dexmedetomidine is 7.1. It is freely soluble in water under normal conditions, which makes it extremely convenient to work with (as it does not require anything eggy or oily to make it injectable). At the same time it is also highly lipid-soluble, with a log octanol/water partition coefficient of 2.8. For those people who do not routinely throw such numbers around in casual conversation, anything with a coefficient greater than 2 is considered to be highly lipophilic, and you can generalise by saying that all the popular CNS-active agents generally have this as one of their properties. Propofol, the champion of lipid solubility, has a coefficient of 4.33, which means that, when left to its own devices in a jar of octanol and water, its concentration in the octanol would be 4.33 orders of magnitude greater than its concentration in the water layer.

Distribution and protein binding of dexmedetomidine

The high lipophilicity would make one think that dexmedetomidine should distribute rapidly and widely into tissues, and that is in fact what is observed. The distribution half-life is only six minutes in healthy volunteers. Of course, this is rarely the use case. After a prolonged infusion, at steady state, there is some equilibration of compartments, and the apparent volume of distribution is about 1.3-2.5L. When in the circulation, most of the drug molecules (96%) are bound to albumin and α1-glycoprotein. Interestingly, in critically ill patients who are usually hypoalbuminaemic, an increased volume of distribution is observed, as there is less total dexmedetomidine left in the circulating volume (nothing for it to bind to). The dose-response relationship in these scenarios is preserved, because presumably there is still enough free drug to do the job.

Elimination and metabolism of dexmedetomidine

The liver is the destination for most dexmedetomidine molecules, with a very high extraction ratio and minimal excretion as unchanged drug. The high hepatic extraction ratio makes it a flow-dependent phenomenon (i.e the liver will remove most of the drug from the blood delivered to it, which means the rate-limiting factor is the rate of blood flow).  There is no single dominant mechanism of biotransformation- it is metabolised by a number of pathways, including N-glucuronidation and hydroxylation by CYP450. None of the metabolites had any clinically significant activity, and they are all mostly eliminated in the kidneys. The presence of renal failure does not appear to limit the use of dexmedetomidine, and the half-life in critically ill patients (coming off a high dose infusion at steady state) seems to be about 2.2-3.7 hours (Iirola et al, 2011).

Context-sensitive half time of dexmedetomidine

The usual poster child for this concept in the CICM exams is fentanyl, but it is not inconceivable that some viva question author might want to use dexmedetomidine one day, and we should be prepared. Fortunately, this topic was well covered by Iirola et al (2011), whose excellent graphs are reproduced here without any permission whatsoever. In the experiment, the rate of infusion was 0.7 μg/kg/h. 

As you can see, in at least one of the patients, after the infusion was ceased some sedating effect persisted for at least 6-12 hours (for some frame of reference, the 1.0 ng/ml concentration is generally associated with being rather heavily sedated, and 0.10 ng/ml would be a barely perceptible change in the level of alertness). This is the effect of tissue distribution, which produces a large cache of active drug in the body, ready to redistribute slowly back into the circulation as the infusion is ceased. Obviously, the longer the infusion, the larger a cache you will accumulate, and the longer the duration of washout. In order to illustrate the effect of infusion duration on the context-sensitive half time, one might be expected to draw some graphs, and here is a representative set from a computer model exercise by Iirola et al (2012):

Mechanisms of action

Like with the other drugs of the imidazoline group, dexmedetomidine has very clear and well-studied effects on the presynaptic α2 noradrenaline receptors, as well as poorly understood effects on imidazoline receptors and opioid receptors (which are probably still quite clinically important).

α2 receptor effects of dexmedetomidine

The α2 receptor is a G-protein coupled receptor (G_i to be precise) that exists on the presynaptic membrane of noradrenergic neurons. That's right, you can't even call this neurotransmission properly because the noradrenaline released by the terminal bouton binds to receptors on the same bouton, and the post-synaptic neuron doesn't know anything about this.

There is a good review of these mechanisms in Khan et al (1999). In short, the effect of binding to these receptors can be broadly described as "inhibitory", or "regulatory". It is a process that can be explained equally well through diagrams or point-form lists.

• The G_i-protein, when activated, will inhibit the activity of adenylyl cyclase
• This leads to a decrease in cAMP
• cAMP controls the phosphorylation state of many proteins, and this definitely has some effect, but  in this case the most important effect of G_i-protein activation would be the opening of G_i-protein associated potassium channels.
• These channels, when activated, produce potassium efflux which hyperpolarises the cell membrane and therefore decreases the likelihood that it will depolarise
• If the presynaptic terminal does not depolarise, it will not release noradrenaline. 
• Ergo, by hyperpolarising the presynaptic membrane, α2 receptor activation creates negative feedback which suppresses further noradrenaline release from the presynaptic nerve terminal.

So, that is how imidazoline drugs are supposed to exert their sedative effect. The suppression of noradrenergic activity in the locus coeruleus, a regulator of arousal and consciousness, is supposed to produce the sedating effects by decreasing the noradrenergic signals to the rest of the nervous system, suppressing the level of consciousness. It is an elegant hypothesis that is almost certainly untrue. When Segal et al (1988) blocked noradrenergic neurotransmission in rats, the sedation they experienced from dexmedetomidine remained essentially unchanged. Whatever the mechanism of sedation is here, we can at least say that it is not wholly dependent on α2 receptor agonist activity.

The analgesic effects of dexmedetomidine are also supposed to be mediated by this mechanism. There is said to be some α2 receptor binding that takes place in the spinal cord and antagonises pain transmission, as well as some ill-defined pain pathway modulation in the higher CNS. This effect is real, as adventurous anaesthetists have given it intrathecally with laudable effect, and Zhang et al (2013) found that there are plausible molecular mechanisms underlying this (related to the inhibition of spinal ERK1/ 2 signalling pathways).

Now, you'd think that all this negative feedback for the release of noradrenaline would have some sort of impact on the autonomic nervous system. That, in fact, was the belief for the longest time. When it first became available in 1966, the antihypertensive effects of clonidine were attributed to the inhibition of noradrenaline release by the medullary vasomotor centres. Again this turned out to be wrong. When Ernsberger et al (1988) injected lots of different imidazolines (with different affinities for the α2 receptor) directly into the brainstems of rats, the haemodynamic effects turned out to be completely unrelated to the α2 receptor affinity. Instead, they correlated well with their affinity for the imidazoline receptors, which is a convenient segue into...

Imidazoline receptor effects of dexmedetomidine

Imidazoline receptors are so named because imidazoline-type drugs tend to bind to them, and not because there is some natural imidazoline substance acting on them normally as an endogenous neurotransmitter. In fact we have no idea what ligand is supposed to bind to them, which does not seem to have stopped vast fields of speculative literature from flowering. Chemically, none of the endogenous ligand candidates have any imidazoline ring structure in them, and they are virtually unknown in pharmacology (harmane, harmalan, agmatine, not exactly household names). Three imidazoline receptor subtypes have been identified, which were unimaginatively labelled 1, 2 and 3.

Bousquet et al (2020) is the most comprehensive review of these receptors, and covers the state of the art in over twenty pages. Needless to say, this is well in excess of what would be expected from senior intensivists and anaesthetists, let alone junior trainees trying to grind through their barrier exams. For the latter, the following (briefest possible) summary will probably still be surplus to need:

• I₁ receptors seem to mediate the hypotensive effects
  • They are G-protein coupled receptors (we do not know which G-protein) and they  decrease intracellular cAMP, though there are multiple other signal transduction systems involved
  • The net effect is a hyperpolarisation of the adrenergic neuron which occurs in parallel  with, but independently from,  the negative feedback occurring from α2-receptor binding
• I₂ receptors seem to mediate the analgesic effects
  • Weirdly, they are expressed on the outer membranes of mitochondria, and on allosteric sites of monoamine oxidase B (MAO-B).
  • Binding these receptors seems to modulate (potentiate) the analgesic effect of opioids and to reduce body temperature, but it is unclear how any this happens.
• I₃ receptors seem to be somehow involved in insulin secretion
  • Activiating these receptors seems to reduce the rate of potassium efflux through K_ATP channels, which decreases the membrane potential and leads to more depolarisation and therefore more insulin release

Considering how little is known about these receptors, and how little progress has been made in the decades since their discovery, one might expect this topic to be unattractive for the CICM examiners, as they would have trouble discriminating between correct and incorrect answers. 

Clinical effects of dexmedetomidine

Sedating effects of dexmedetomidine

The unique and highly sought-after effect of dexmedetomidine is the "cooperative sedation" which it produces at low doses. With increasing dose, cooperation gives way to drowsy indifference, which according to EEG studies is electrically indistinguishable from normal N2 phase sleep (Zhang et al, 2015). In a study by Ebert et al (2000), healthy volunteers demonstrated a predictable loss of awareness at a certain dose, which the investigators unfortunately measured in terms of plasma dexmedetomidine concentration (ng/ml). Because most  people will not have access to infusion pumps with in-built pharmacokinetic models which allow precise targeting of plasma concentrations, the following graph had to be slightly altered to include useable dose data from Fujita et al (2013):

How the hell it does this, nobody knows. Authorities (eg. Weerink et al, 2017) shrug that "The exact mechanisms are not fully understood at the moment, although it is known that receptors, other than those acting on the gamma-aminobutyric acid system, play a role". Textbooks and less discerning papers (including CICM past papers) will often uncritically repeat the α2-related noradrenergic locus coeruleus hypothesis, but as mentioned above, this is hardly settled.

Protection against delirium

Given the natural organic gluten-free quality of the "sleep" produced by dexmedetomidine, it would be logical to expect that it should decrease the incidence of delirium, when compared to other, more brutally industrial ICU sedatives. This effect is occasionally seen, but the reader must be warned that the magnitude of this effect seems to vary inversely with the size and methodological quality of the study. Small studies in cardiac surgical patients (Maldonado et al, 2009) and comparisons to lorazepam (MENDS trial, 2007) demonstrated significantly more delirium-free days in the dexmedetomidine group. However,  huge 4000-patient gigatrials from the ANZICS CTG behemoth have been performed, for some reason primarily looking for a mortality difference (they found none), and demonstrated only one day of difference in delirium-free days between intervention and control (23 vs 24 days). In short, it is still safe to say that adding dexmedetomidine decreases delirium in the marketing literature, so long as one does not go into too much detail.

Analgesia with dexmedetomidine

It is often said to have analgesic effects, but their magnitude is debated, and not everybody is convinced that they exist. In an uncharacteristic avoidance of unnecessary detail, the reader is referred to Weerink et al (2017) where studies dealing with this property are dissected. To summarise, the analgesic effect of dexmedetomidine on its own is probably minimal and what analgesic requirement reduction we see may be mainly due to the decreased anxiety and perception of pain (consider that to have their pain measured, the patient needs to be sufficiently motivated to report it). 

Airway reflexes with deep dexmedetomidine sedation

One of the the main touted advantages of this drug is its ability to profoundly sedate people without the need for complex airway management, freeing more time for the anaesthetist to explore the stock market on their laptop. This is for real.  Mahmoud et al (2017) found that airway responses with dexmedetomidine sedation resemble those observed during natural sleep. In other words, if you routinely obstruct your airway in your own bed due to OSA, you will still have apnoeic episodes while under the effects of dexmedetomidine in the ICU.

Respiratory drive and ventilation with dexmedetomidine

Even when Ebert et al gave their one brave patient a truly insane dose of dexmedetomidine (a plasma concentration of 14.3 ng/ml, probably corresponding to something like 4.5-5.0 μg/kg/hr), their respiratory drive remained unaffected, even as they were completely unrousable. Interestingly, even their CO₂ remained stable, within about 3-4mmHg of baseline. Writers tend to include warnings that dexmedetomidine is often used with other sedating agents (the respiratory depressant effects of which would be potentiated), and that the response to hypercapnia is depressed among the elderly, in whom heroic doses of dexmedetomidine would therefore have unpredictable respiratory effects. Of course, one would never see that, because heroic doses of dexmedetomidine in the fragile elderly would have totally predictable haemodynamic effects, as will follow:

Haemodynamic effects of dexmedetomidine

Dexmedetomidine and all the other imidazoline derivatives have peripheral α2-receptor effects on the vascular smooth muscle, which tend to produce vasoconstriction.  The result of this is a transient increase in blood pressure, which is often so short-lasting that one might miss it in the operating theatre (it will sneak in between cycled non-invasive blood pressure measurements). As the drug crosses the blood-brain barrier, it begins to exert its mysterious effects on central sympathetic nervous system control, and a reduction in blood pressure and heart rate is seen, with some studies reporting a 60-80% decrease in circulating catecholamine levels. The reduced heart rate logically decreases the cardiac output, but there is apparently little effect on contractility, at least within a sane dose range.