A different kind of unconsciousness
— and two drugs given alongside rather than instead.

• Ketamine’s block is non-competitive and use-dependent, because its site is inside a channel that has to open first. Calling it competitive gets the pharmacology backwards.
• Dissociative anaesthesia is a distinct state, not deep sleep. The cortex is depressed while the limbic system is stimulated, and the reflexes that remain must not be assumed protective.
• Ketamine is a direct myocardial depressant wearing a sympathomimetic disguise. Remove the catecholamines and the disguise comes off.
• Midazolam’s ring opens below pH 4 and closes above it — water soluble in the ampoule, lipophilic in the patient. It is the only benzodiazepine that manages both.
• Dexmedetomidine’s haemodynamic response is biphasic, and the early hypertension and the later hypotension have two different mechanisms in two different places.

Ketamine’s dominant action is non-competitive antagonism at the NMDA receptor, binding the phencyclidine recognition site inside the open channel. Lesson 1 sets out what follows from that location — the antagonism is non-competitive and unsurmountable, the block is use-dependent because the channel must open first, and the mechanism itself changes with concentration, from a reduction in opening frequency at sub-anaesthetic doses to blockade of both open and closed channels at higher ones.

What that produces at the network level is worth stating separately: the most important consequence is inhibition of NMDA-mediated glutamatergic input to the GABAergic system, which changes excitatory activity in the cortex and limbic system and ends in unconsciousness. Ketamine does not achieve unconsciousness by adding inhibition; it achieves it by removing an excitatory input, which is why the resulting state looks so unlike the others.

Ketamine also acts at a number of other sites, and the honest account gives them different weights rather than listing them as equals.

The primary site of ketamine’s central action is the thalamoneocortical projection system. It selectively depresses neuronal function in parts of the cortex — the association areas particularly — and in the thalamus, while simultaneously stimulating parts of the limbic system, including the hippocampus. The result is described as a functional disorganisation of non-specific pathways in the midbrain and thalamic areas: the electroencephalogram shows dissociation between the thalamocortical and limbic systems, with the α rhythm replaced by θ and δ activity.

Contrast that with a GABAergic agent. Propofol, thiopentone, etomidate and midazolam all produce a state that resembles natural sleep — quiet, still, eyes closed, progressive depression of everything at once. Ketamine produces a state in which some systems are more active than they were before. That is what the word dissociative is doing: it is not a synonym for deep, it describes two parts of the brain that are no longer talking to each other.

Ketamine is also the one agent in the module that does not meet the module’s own definition: it does not induce anaesthesia in one arm–brain circulation time, with central effects becoming evident at about 90 seconds after an intravenous dose. That is a consequence of what has to happen — a channel population has to open before it can be blocked — rather than of poor delivery.

The carbon carrying both the chlorophenyl ring and the methylamino group is asymmetric, so ketamine exists as R(−) and S(+) enantiomers. Most ampoules contain the racemate; the S(+) enantiomer, esketamine, is marketed separately in some countries.

Why a single enantiomer behaves differently from its mirror image at all is stereoselectivity, and it is taught in the isomerism lesson rather than repeated here. One clinical note that is not pharmacology but belongs somewhere: an S(+)-ketamine nasal spray has been approved for treatment-resistant depression, which is why the drug now appears in contexts far outside the operating theatre.

Ketamine is unlike every other induction agent in producing sympathetic nervous system stimulation. Circulating adrenaline and noradrenaline rise, and with them heart rate, cardiac output, blood pressure and myocardial oxygen requirement. It does not appear to precipitate arrhythmias.

There are two mechanisms behind the sympathetic effect, and giving both is better than giving one:

1. Central sympathetic outflow increases. Ketamine acts centrally to raise sympathetic tone.
2. Reuptake of catecholamines into postganglionic sympathetic nerve endings is inhibited. This is a cocaine-like action, and it means the catecholamine already released stays in the synapse longer.

The consequences for specific patients follow from the same argument. Ischaemic heart disease: tachycardia and hypertension raise myocardial oxygen demand while shortening diastole, so ketamine is unwise as the sole agent — and this is where the S(+) isomer’s preservation of ATP-sensitive potassium channels becomes an argument in its favour. Vascular aneurysm: the sudden rise in pressure is the specific hazard. Haemorrhagic shock with intact sympathetic reserve: ketamine is the agent of choice, and the mechanism is why.

• Respiratory drive is largely preserved, and the respiratory rate may increase. This is the property that makes ketamine the agent for a patient whose ventilation you do not want to take over.
• Bronchodilatation. Ketamine relaxes bronchial smooth muscle, consistent with its muscarinic antagonism and its sympathomimetic action, and is genuinely useful in asthma and in severe bronchospasm.
• Laryngeal reflexes are relatively preserved — with the caveat from section 03 that preserved is not the same as protective.
• A patent airway is often, but not always, maintained. Increased muscle tone in the jaw can itself precipitate airway obstruction, which is a mechanism specific to this drug: the airway is lost through too much tone rather than too little.
• Salivation and respiratory secretions increase, sometimes markedly. An antisialagogue — glycopyrrolate for preference, since it does not cross the blood–brain barrier and will not add to the central anticholinergic picture — is conventionally given with it.
• Apnoea has been reported, particularly with rapid injection, so this is a favourable profile rather than a guarantee.

Analgesia occurs at considerably lower blood concentrations than loss of consciousness — the pain threshold is raised at plasma concentrations of 0.1 μg/mL or more, while general anaesthesia requires 0.6 to 2 μg/mL. Two things follow. There is a considerable period of analgesia after emergence from a ketamine anaesthetic; and sub-anaesthetic doses can be used for analgesia alone.

The analgesia is not a single mechanism. NMDA antagonism at the spinal cord inhibits dorsal horn wide-dynamic-range neuronal activity and prevents central sensitisation; ketamine occupies opioid receptors in brain and spinal cord, with the S(+) isomer carrying μ activity; it inhibits monoamine reuptake; and it activates descending inhibitory monoaminergic pathways. Functional imaging shows a dose-dependent reduction in activation of the secondary somatosensory cortex, insula and anterior cingulate cortex — reduced connectivity both in the regions that sense pain and in those that process its affective component.

The clinically important consequence is that ketamine prevents opioid-induced hyperalgesia and attenuates acute opioid tolerance, because the NMDA receptor is central to both. That is the pharmacological basis of the opioid-sparing effect and of its use in chronic and neuropathic pain.

Norketamine, the principal metabolite, is itself active — variously reported at 20 to 30%, or a third to a fifth, of the potency of the parent compound. It is less potent than ketamine, not more, and it contributes to the prolonged analgesia after a bolus or an infusion.

Emergence phenomena

Vivid and unpleasant dreams, hallucinations and delirium may follow a ketamine anaesthetic. Four things reduce the incidence or severity, and all four are worth knowing because they are among the few genuinely actionable pieces of advice in the module:

• Concurrent benzodiazepine or opioid. This is the most effective measure, and it is why midazolam and ketamine are so often given together.
• Recovering undisturbed. Emergence phenomena are less common in patients left quiet during recovery — a nursing intervention rather than a pharmacological one.
• Age. They are less common at the extremes — in the young and in the elderly.
• The S(+) isomer, which produces less intense phenomena — though, as section 04 notes, no less frequent ones.

Ketamine and intracranial pressure

Routes and bioavailability

Ketamine is unusually flexible, because it is water soluble and highly lipid soluble at once. Bioavailability is 93% parenteral, but only 20–30% oral because of extensive first-pass metabolism, and 40–50% intranasal, which partly bypasses the liver. It has been used rectally and, in preservative-free form, epidurally and caudally. The preservative in the standard preparation is chlorobutanol, which is neurotoxic, so that formulation must never be given neuraxially.

Two further adverse effects complete the picture. Nausea and vomiting occur more frequently than after propofol or thiopentone. Chronic high-dose recreational use causes severe interstitial cystitis which may ultimately require cystectomy, and hepatotoxicity has been reported after repeated prolonged infusions for chronic pain.

Where it binds. Midazolam acts at the benzodiazepine site at the α/γ interface of the GABA-A receptor, producing positive allosteric modulation and nothing else — as lesson 1 sets out, it increases the frequency of channel opening in the presence of GABA and cannot open the channel alone. The ceiling effect follows directly: there is a maximum beyond which more drug cannot increase inhibition, because the drug can only amplify what the brain is already producing.

Which effect you get depends on which α subunit the receptor carries. Sedation, anterograde amnesia and the anticonvulsant action are mediated by α1-containing receptors; anxiolysis and muscle relaxation by α2-containing receptors. That is why the effects can be partly dissociated, and why the amnesic effect is more potent than the sedative one — a patient can be awake, conversational and entirely amnesic for the conversation.

Amnesia is anterograde, not retrograde. Midazolam prevents the formation of new memories from the time it is given; it does not erase memories already laid down. Describing it as retrograde is a common error and reverses the clinical meaning entirely.

Co-induction is the use of a small dose of one agent to reduce the dose of another, and midazolam is its commonest example. Its interactions are not all the same kind, and distinguishing them is the substance of the topic:

The metabolite is the other thing that distinguishes midazolam. CYP3A4 and CYP3A5 produce 1-hydroxymidazolam, which has similar sedative activity to the parent compound — its receptor affinity is about 60% of midazolam’s, so it is less potent but far from inert. It is normally conjugated rapidly and excreted renally, so it matters not at all after a single dose. It matters a great deal in two situations:

• Prolonged administration, where the metabolite accumulates.
• Renal impairment, where the conjugated metabolite is not cleared, and sedation is prolonged and unpredictable. This is the specific reason midazolam behaves badly in intensive care.

The elderly are affected twice over, and it is worth separating the two: a pharmacokinetic change — reduced clearance, and a larger volume of distribution with obesity prolonging the elimination half-life — and a genuine pharmacodynamic increase in sensitivity. Cirrhosis reduces clearance further by reducing metabolism.

Lesson 1 sets out the pathway: reduced noradrenergic output from the locus coeruleus releases the ventrolateral preoptic nucleus, which increases GABA and galanin release onto the tuberomammillary nucleus, which reduces histamine release into cortex and subcortex. The clinical properties all descend from having taken that route rather than the GABA-A one.

Organ failure. Clearance falls with hepatic impairment — to about 74%, 64% and 53% of normal in mild, moderate and severe impairment respectively — so the dose should be reduced. Renal impairment does not alter the pharmacokinetics, though the sedative effect may be stronger in severe renal disease. Weight-based dosing is only properly justified in the non-obese; fat-free mass may be more appropriate in obesity, and that remains under investigation.

The haemodynamic response to dexmedetomidine is biphasic, and it is biphasic because two different α2 receptor populations in two different places are being stimulated with different concentration–response relationships.

In a phase III trial of 401 patients the reported incidences were hypotension 30%, hypertension 12% and bradycardia 9%. All three are common; none is rare enough to treat as an idiosyncrasy.

A choice of induction agent is rarely a free one. In almost every case a single feature of the patient dominates, and the reasoning runs in a fixed order: identify what must be preserved, identify which agent’s mechanism threatens it, and choose from what remains. The flow below is that order.

Two of the rows are worth a further sentence because they are the ones most often answered by reflex.

Malignant hyperthermia susceptibility is a question about what is not a problem. No intravenous induction agent triggers it; the trigger list is the volatile agents and suxamethonium. Total intravenous anaesthesia is the technique, and every drug in this module is available. Answering it as though a choice had to be made between the intravenous agents misses the point of the question.

The shocked patient is the one where the reflex answer — ketamine — is right for a reason that can fail. Ketamine supports the circulation indirectly, through sympathetic stimulation and catecholamine reuptake inhibition. In a patient whose catecholamine reserve is already exhausted there is nothing left to stimulate, the direct myocardial depressant action is unmasked, and the drug chosen for its haemodynamic support can cause collapse. Etomidate is the alternative when that is a real concern, with the adrenal question from lesson 3 attached to it. There is no agent here without a cost; the skill is knowing which cost this patient can pay.