Pharmacology of neuromuscular blocking drugs

This chapter addresses Section L2(i) from the 2017 CICM Primary Syllabus, which expects the exam candidates to have an "understanding of the pharmacology of neuromuscular blocking drugs".   For all intents and purposes, this section is the musculoskeletal system, in the sense that its representation in the exam (and, to be fair, in the daily life of an intensivist) is the most substantial among all the musculoskeletal topics. Of the past SAQs covering the musculoskeletal system, the vast majority were about neuromuscular junction blockers:

• Question 19 from the first paper of 2023 (prolonged block, and vecuronium)
• Question 13 from the first paper of 2022 (rocuronium vs. suxamethonium) 
• Question 6 from the first paper of 2021 (vecuronium)
• Question 10 from the second paper of 2020 (suxamethonium)
• Question 13 from the second paper of 2018 (roc vs cis)
• Question 10 from the first paper of 2018 (suxamethonium)
• Question 12 from the first paper of 2016 (speed of onset)
• Question 7 from the second paper of 2014 (potency)
• Question 9 from the first paper of 2014  (speed of onset)
• Question 1 from the second paper of 2013 (suxamethonium)
• Question 2 from the second paper of 2012 (suxamethonium)
• Question 6 from the first paper of 2011 (suxamethonium)
• Question 3 from the second paper of 2009 (speed of onset)
• Question 1 from the second paper of 2008 (potency)

As the reader can plainly see, there's a few examinable topics here. Suxamethonium is somewhat overrepresented, and is therefore deserving of its own chapter, but of the curare poisons, most only ever appear once, and can probably be lumped together into one uber-chapter. Some suxamethonium information therefore ends up being duplicated here, considering that questions about potency and speed of onset will usually need to be answered with reference to both depolarising and nondepolarising blockers.

Given that vec, roc, cis and sux have all been asked about at some stage, it would be difficult to produce a non-cumbersome table which covers all four, and so the greybox summary of these agents is therefore made generic to cover the class rather than to focus on single agents. 

For a broad overview that covers all the basics and does not get involved in overmuch biochemical wankery, Appiah-Ankam & Hunter (2004) is perfect, and free from the good samaritans of the BJA. On the other hand, if biochemical wankery is in fact welcome and desirable (and let's face it, reader, you wouldn't be here if it wasn't),  Lee (2001) or Wierda & Proost (1995) are excellent deep dives into structure/function relationships, Atherton & Hunter (1999), as well as Agoston et al (1992), are excellent for pharmacokinetics, and Kampe et al (2003) or Claudius et al (2009) for the side effects.

What is, and what is not, a neuromuscular junction blocker

"Muscle relaxants",  "paralysing agents" and "curare toxins" are terms we often use to describe these drugs, and we use them knowing full well that they are colloquialisms, but generally speaking they are understood and useful clinically. To refer to them as "neuromuscular junction blockers" would probably be more accurate, as this separates them from the whole host of substances that affect neuromuscular function at levels other than the neuromuscular junction. In fact, it feels normal and appropriate to digress about those right here, if only because there may not be any better place to put them:

Substances that Affect Neuromuscular Function at Levels Other than the Neuromuscular Junction

At the level of the sarcolemma: Dantrolene (interferes with calcium release) Quinine (depletes calcium store) At the level of acetylcholine synthesis (blockade of choline reuptake) Hemicholinium Triethylcholine Vesamicol At the level of the presynaptic membrane: Botulinum toxin (interfere with vesicle docking) Aminoglycoside antibiotics (inhibit calcium-induced release) At the level of the CNS: Baclofen Tizanidine Cyclobenzaprine

One could try to functionally define NMJ blockers as acetylcholine receptor blockers or nicotinic receptor antagonists, but this phrasing does not have any support for some reason. Presumably this is because nicotinic acetylcholine receptors are ubiquitous, and the agents that act at the neuromuscular junction are sufficiently distinct from (for example) ganglionic blockers which technically also block nicotinic receptors. In short, "neuromuscular junction blockers" is a term we are stuck with.

Classification of NMJ blockers

We can decide to classify them on the basis of chemistry, or mechanism of action, or ideally both. 

• Depolarising NMJ blockers open the ion channels associated with postsynaptic nicotinic acetylcholine receptors and depolarise the motor endplate for a sustained period, preventing the generation of new myocyte action potentials. Examples include: 
  • Decamethonium (1949)
  • Laudexium (1950)
  • Suxamethonium (1951)
• Non-depolarising NMJ blockers are competitive antagonists of acetylcholine at the active binding sites of the postsynaptic nicotinic acetylcholine receptor. 
  • Trisquaternary compounds
    • Gallamine (1947)
  • Benzylisoquinolinium diesters:
    • d-tubocurarine (1942)
    • Doxacurium (1980)
    • Mivacurium (1981)
    • Atracurium (1981)
    • Cisatracurium (1994)
  • Chlorofumarate diesters
    • Gantacurium (not yet available)
  • Aminosteroids:
    • Pancuronium (1964)
    • Vecuronium (1975)
    • Rocuronium (1988)

The development of these drugs was a fascinating process that is unfortunately beyond the scope of even this notoriously directionless resource. It yielded a massive variety of potential agents, the vast majority of which never saw the light of day, usually because of some crippling flaws (terrifying autonomic side effects, for example). However it is still worth driving down these culdesacs to look at the abandoned molecular structures, as there is something potentially useful to be learned from them.

Structure-function relationship of NMJ blockers

Being the ligands for the postsynaptic nicotinic acetylcholine receptor, the majority of these substances are - in some way or another - structurally similar to acetylcholine.  And acetylcholine is just an acetate molecule and a choline molecule, taking a long walk together. To borrow an image from an early version of this website:

Choline is a quarternary amine, i.e. a permanently charged N⁺ core bound to four organic molecules. In this case, three of those organic molecules happen to be  methyl groups (CH₃), which gives us the opportunity to meet the term "methonium" (methyl and ammonium, get it?). This is the business end of acetylcholine, sometimes referred to as the "head" of the molecule, and it binds to a negatively charged section of the active site of the acetylcholine receptor.

 From this methonium head, we get suxamethonium, which is basically just two acetylcholine molecules in a trenchcoat. Without revisiting a lot of the structure-activity relationship work in the suxamethonium chapter, it will suffice to remind the reader that other methoniums (methonia?) also exist, and many of them have muscle relaxant or ganglion-blocking properties, though only decamethonium is clinically useful as a paralysing agent and only hexamethonium is clinically relevant as a ganglionic antagonist.

The inclusion of acetylcholine in this lineup of rogues is interesting, as it raises a perfectly natural question: would an infusion of acetylcholine result in a depolarising neuromuscular junction block? Readers may be surprised to learn that such infusions (intravenous and intraarterial) were a favoured research method of early physiologists who were investigating the autonomic control over various vascular beds (eg.  Ellis & Weiss, 1931). None of these experiments resulted in muscle paralysis (the subjects just dropped their blood pressure by about 20mmHg)  presumably because the acetylcholinesterase present at the neuromuscular junction is extremely efficient at clearing acetylcholine. Only when this enzyme is inhibited could an acetylcholine infusion become an effective neuromuscular junction blocker, and in fact Thesleff (1955) determined that about ten times as much acetylcholine was required to produce depolarising block when neostigmine was not involved. Logically, therefore, any intravenous dose of acetylcholine large enough to overwhelm this mechanism would surely kill the subject through vagally mediated bradycardia, amidst other cholinergic horrors, and would probably be indistinguishable in its effect from the consequences of sarin gas poisoning.

Anyway. As you can see, these methonium compounds can be simple chains (for example decamethonium is just a polymethylene chain ten atoms long), or they can be weird many-headed monsters like gallamine (which has three ethonium heads hanging off a phenyl ring). This representation from a defunct pharmacopoeia is particularly Hydra-like:

Gorgeous. And even more interestingly, when the molecule gets bulkier like this, something weird happens. The drug no longer wedges the ion channel stuck in the open configuration, i.e it starts acting as a non-depolarising blocker. This relationship of molecule bulk to its function was first described by Bovet (1951), who used the terms "pachycurare" and "leptocurare" to make a distinction between fat molecules and thin molecules. CICM trainees are strongly advised against using these terms in their written paper, as it appears they have fallen into disuse (and feel faintly insensitive from a cultural perspective), but the relationship of molecule size and flexibility on its effects at the nicotinic receptor are worth remembering.

Other structure/function rules loosely apply, listed in some detail by Lee (2001). To briefly outline these:

• Bisquaternary compounds (those with two methonium groups) are said to be more potent, in terms of magnitude of NMJ blocker effect per dose. For example, Hill et al (1994) mention in passing that, if decamethonium were to be made monoquaternary by removing one of the methonium groups, its potency at the neuromuscular junction decreases by a factor of 50. The same stands for non-depolarising agents; monoquaternary drugs like rocuronium require doses around 50-100mg, whereas bisquaternary drugs like pancuronium and cisatracurium required doses in the range of 5-10mg. The exception to this rule is vecuronium, which is monoquaternary and still very potent. 
• The distance between quaternary ammonium groups appears to play a role in determining potency and the degree of ganglionic block side effects - i.e. agents with a shorter distance seem to produce more ganglionic block, and agents with an optimum distance, thought to be about ten carbon atoms, have the greatest potency.
• A short (eg. methyl) chain on one of the quaternary ammonium groups imparts potency; vecuronium has this and rocuronum does not. The longer these chains, the more
• Lipophilicity among aminosteroids is associated with lower potency (and is usually imparted by long chains hanging off the quaternary ammonium group)
• Methoxy groups (methylated -OH groups) impart potency mainly to the benzylisoquinolinium family

There are many others, for example dealing with the susceptibility of these drugs to  different liver enzymes - but one thing is clear: none of these are essential knowledge, and the CICM trainee preparing for their exams can safely ignore this section in its entirety. From a technical perspective, looking at the molecular structures of nondepolarising agents would force any reasonable person to conclude that they are sufficiently difficult that nobody will ever be asked to reproduce them for an exam. Moreover, nobody could possibly argue that the knowledge of these structure-function relationships is somehow a necessary precondition for the safe clinical use of these substances. However, that combination of complexity and irrelevance has never been a barrier on this website. Therefore, in the spirit of pointless excess, the reader will now be ambushed with the following impossibly long 1185 × 3517 pixel diagram of common NMJ blocker structures, with important functional groups identified and described:

Pharmaceutical presentation of NMJ blockers

From the point of view of an educator, there's not much to be said about the presentation of these drugs, other than to list the caveats to their storage which arise from various idiosyncratic chemical properties. Predictably, these agents are usually sold in sterile ampoules containing one adult dose of the drug (for a normal-sized adult) and are usually brightly labelled with a warning to indicate that they can paralyze people. One thing that must be mentioned is that these substances are, for the most, in some way inherently unstable, and therefore need to be refrigerated for storage. This varies: the room temperature shelf life of pancuronium is said to be three to six months, whereas in the case of vecuronium, the drug cannot even be presented as an aqueous solution, and is instead sold as a lyophilized powder which needs to be reconstituted immediately prior to use. Information about the specific storage requirements and shelf stability of these substances seems to be scattered widely, and no single peer-reviewed resource seems to summarise and explain their storage properties. One is left with charts like this, clearly meant to be stuck to the doors of pharmacy fridges. To summarise:

Storage Conditions for Neuromuscular Junction Blockers

Drug	Recommended storage temp	Stability at room temperature
Cisatracurium	2 – 8ºC	~ 48 hours (but up to 21 days for some formulations)
Pancuronium	2 – 8ºC	3 months
Rocuronium	2 – 8ºC	8-12 weeks
Suxamethonium	2 – 8ºC	30 days
Vecuronium	< 25ºC (as powder)	Years

Administration and absorption of NMJ blockers

All of these agents are minimally absorbed from the gastrointestinal tract, which is excellent because historically it has meant that it was safe to eat the meat of animals killed by curare darts. Moreover, one might be tempted to point out that the convenience of oral administration for use in the community is not a desirable property for a paralysing toxin; and that the slow unpredictable onset of effect would make oral NMJ blockade a frustrating experience for the anaesthetist and surgeon. The idea must be so preposterous that no modern data has been collected to support the throwaway line seen in textbooks, that these agents are minimally absorbed from the GI tract. What we know mainly comes from rat studies performed using tubocurarine, such as Mahfouz (1949); the author fed rats a concentrated tubocurarine mixture for some hours before sacrificing them and recovering virtually all of the drug unchanged from their GI tracts. They also seem reluctant to cross even well-vascularised mucous membranes:  Katz & Barr (1955) gave a range of these agents sublingually, and were unable to detect any clinical relaxant effect even at very high doses (up to 500 times the normal intravenous dose, eg. 50mg/kg of suxamethonium). In short, critical care trainees will only ever be expected to know about the parenteral routes of administration; of which the main ones are IV and IM.

Intramuscular administration of NMJ blockers

Yes, reader, it is possible, and in fact occasionally desirable and appropriate, to give these agents as an IM injection, and these circumstances are not limited to assassination. It is in fact a completely reasonable reaction to a situation where a challenging intubation must imminently occur in the absence of decent vascular access, such as in the case of emergency paediatric anaesthesia (Brandom & Fine, 2002). In this case, a (slightly larger than usual) dose of NMJ blocker can be administered IM. The result is a slower onset of effect, but an effect nonetheless.

There is some variability in the effectiveness of different agents when given IM, which is mainly related to their pharmacokinetic profile. Suxamethonium and rocuronium have reasonably good absorption from the site of injection (after 30 minutes, only about 5% of the rocuronium is still in the muscle, according to Reynolds et al, 1996), but other drugs are either badly degraded by local metabolic processes, or are too slow to diffuse to make it into the circulation in clinically significant quantities, or have some other peculiarity that makes them unsuitable. For example, mivacurium, a famously massive molecule, takes so long to get into the circulation that it is completely useless for intubation or the relief of laryngospasm. Atracurium has had case reports of successful intramuscular use but is generally held to be too irritant to be routinely used like that.  Tubocurarine has been used IM in tetanus, mainly for the relief of muscle spasm, but its potency was reduced so greatly that it was being safely administered as a regular dose to spontaneously breathing ward patients. In practice, though it seems all agents will have some degree of effect if given IM in a large enough dose, only suxamethonium is common.  Shaw et al (2015) list several studies where doses and onset times are documented, and recommend a dose of 4mg/kg for suxamethonium.

Why is this practice limited to paediatrics? Is it just naturally assumed that adults can be reasoned with, and talked into a cannula? Because we all know that's just simply not true. The answer may be mainly related to the pharmaceutical preparation: most of these drugs are not offered in a small highly concentrated volume, with the exception of suxamethonium (100mg/ml) and vecuronium (which is a powder that you can reconstitute into whatever volume you like). Because only a reasonably finite volume can be given as an intramuscular injection,  only about 10-20mg of rocuronium could be given, which would only be enough for a small child. At least this is the excuse given by Shaw et al (2015) to explain the lack of adult data for this drug. An inventive reader may point out that the adult body has numerous large muscles which could become the recipients of multiple intramuscular doses, but of course, in the sort of emergency where this level of cruelty becomes acceptable, the normal sane person would instead opt for the intraosseous route.

Protein binding and distribution of NMJ blockers

What follows is a table of basic physicochemical and pharmacokinetic properties that was pieced together from fragments of Atherton & Hunter (1999), Cameron et al (1995), Roy & Varin (2004), and multiple other works.

Drug	pKa	Volume of distribution (L/kg) *	Lipid solubility (octanol partition coefficient, Log)	Protein binding
d-tubocurarine	6.45	0.74	minimal lipid solubility	42%
Mivacurium	18.59	0.15-0.34	-2.46	28%
Atracurium	19.02	0.10-0.12	-3.98	38%
Cisatracurium	19.02	0.16-0.20	-3.73	38%
Pancuronium	7.4	0.20-0.23	minimal lipid solubility	82%
Vecuronium	8.97	0.20-0.27	-0.75	77%
Rocuronium	7.96	0.22-0.27	-1.68	46%
Suxamethonium	13.0	0.14	-4.15	20%

* The exact VODs are all slightly different depending on which textbook you read, and textbook VODs are all slightly different from manufacturer PI sheets. This column represents the range of reported values, for lack of any better idea. Trainees are reminded that specific values are meaninless and memorising them is a fool's errand.

The reader seduced by neat tables of numbers should be warned that this table is not a reliable reference for these numbers, and that some of this basic information was so hard to find that its reliability is called into question. For example, the statement that that the pKa of most nondepolarising agents is around 8.0 comes from a small-n sheep study performed in 1980. VODs are listed in a table in Jurado et al (2015), but the authors do not mention where they got these values from. According to a casual aside by Payne (1958), the pKa of suxamethonium, decamethonium and gallamine is over 13, but the author did not give any actual specific numbers nor mention where this data comes from. Occasionally, one can find a reassuring statement from the heart, such as where the writer of the mivacurium product monograph for AbbVie, who confessed that as a quaternary ammonium salt, "it is completely ionizable at all concentrations measured. No pKa values can be assigned". In defence of these authors,  the drugs themselves don't make things easy. For example some (like mivacurium) have multiple enantiomers, and each of them has a slightly different VOD and pKa; or the drug might constantly break down spontaneously, making a mockery of your attempts to measure its volume of distribution. The most important tl:dr here is:

• The pKa for most NMJ blockers makes them highly water-soluble
• It doesn't matter anyway, as they never need to penetrate a single lipid bilayer on their way to their site of action.
• They are largely confined to the circulating volume, with small VODs
• They are not very protein-bound, with the most protein-bound being pancuronium
• They vary a lot in terms of their lipid solubility, with vecuronium being the most lipid soluble of them all, and suxamethonium the least (being about 10,000 times less lipid soluble than vecuronium). The reader horny for log-scale graphs will enjoy this representation of octanol:phosphate buffer partition coefficients from Roy & Varin (2004):
  

Metabolism and elimination of NMJ blockers

The exam-going candidate will benefit from a kind of structured generalisation here that sacrifices accuracy in return for memorability. Thus:

• Drugs that end in "methonium" tend to be metabolised by plasma cholinesterase, and have relatively short half-lives. 
• Drugs that end in "curium" tend to be metabolised by various random plasma esterases, and generally fall apart spontaneously at normal human body temperature
• Drugs that end in "curonium" tend to be metabolised mainly by the liver, but rely on renal clearance to a clinically significant extent

To remember this series of truisms is easiest, and the truisms then make it easier to remember the variations to the rules, such as:

• Mivacurium, unlike atracurium and cisatracurium, depends considerably on plasma cholinesterase for its metabolism, and behaves more like suxamethonium clinically (i.e. rapid offset of effect)
• Atracurium and cisatracurium undergo spontaneous degradation - much more so than the rest of these inherently unstable drugs - to a degree that this accounts for the majority of their clearance
• Pancuronium is the one that is most dependent on renal clearance mechanisms- 60-80% is excreted unchanged in the kidney, and the rest is converted into modestly active metabolites by the liver. It has a longer duration of action because of its relatively high protein binding (making it resistant to renal clearance) and the active metabolites are also renally cleared, increasing the total duration of block.
• Vecuronium is cleared by both biliary and hepatic mechanism equally; where about a third is metabolised, about a third is excreted as unchanged vecuronium in the bile, and a third is cleared renally.
• Rocuronium is mainly cleared by the liver, where it undergoes surprisingly little metabolism (less than 1% of it undergoes biotransformation and the rest is just dumped into the bile). Only about 10% is eliminated renally. 
• Tubocurarine, a longer acting agent which has lost its popularity, had an extremely long half life and was significantly dependent on hepatic metabolism for its clearance.
• Suxamethonium is metabolised by butyrylcholinesterase (aka "plasma cholinesterase" and "pseudocholinesterase"), an enzyme which is not present at the neuromuscular junction. Both acetylcholinesterase and butyrylcholinesterase enzymes can hydrolyse choline esters, and they are sufficiently similar that there is a non-trivial amount of drug effect crossover. For example, neostigmine, an acetylcholinesterase inhibitor, will also inhibit the activity of butyrylcholinesterase even at normal therapeutic concentrations. According to an older edition of a CICM official textbook (Stoelting), "30 minutes after administration of neostigmine, the butyrylcholinesterase activity remains about 50% of control values". On the other hand, there is also plenty of differences between the activity of these enzymes: for example, acetylcholinesterase is unable to metabolise suxamethonium.

Drug	Metabolism	Clearance	Half-life (from various sources)	Half-life from Stoelting
d-tubocurarine	Significant hepatic metabolism (around 50%)	10% eliminated in the bile, 45% is renally excreted	Half-life 200-300 minutes	84 min
Mivacurium	Metabolised by butyrylcholinesterase	Minimal renal or hepatic clearance	Half-lives of most isomers are short (2-5 minutes)	2 min (cis-trans), 68 min (cis-cis)
Atracurium	Metabolised by plasma esterase and spontaneously degraded by temperature-sensitive Hofmann degradation	Minimal renal or hepatic clearance	Half-life 21 minutes	21 min
Cisatracurium	Metabolised by plasma esterase and spontaneously degraded by temperature-sensitive Hofmann degradation	Minimal renal or hepatic clearance	Half-life 26 minutes	17.3 min
Pancuronium	Minimal metabolism (less than 1%)	10-20% cleared by the liver, 80-90% cleared by the kidney	Half-life 110 minutes	132 min
Vecuronium	Perhaps 5% is metabolised; one of the metabolites has about 80% of the potency of the parent drug (3-desacetylvecuronium)	30-40% is cleared by the liver and into the bile (good lipid solubility); another 30% is renally eliminated as unchanged drug	Half-life 110 minutes	73 min
Rocuronium	Minimal metabolism (less than 1%)	around 90% is cleared by being excreted into bile, and the rest is cleared renally	Half-life 60-100 minutes	71 min
Suxamethonium	Metabolised by butyrylcholinesterase	Minimal renal or hepatic clearance	Half-life 1-2 minutes (duration of action 5-6 minutes)	47 seconds

The half-lives of these drugs seem to always be different depending on which textbook you read, which makes it somewhat difficult for the trainee to know exactly which pointless fact they are supposed to memorise in order to push the "marks" button in the examiner's brain. It does not help that the values are frequently very different from book to book. In order to make this even more difficult, and to rub their noses in it, the readers are offered this table of official CICM textbooks and their differences.

Half-lives of NMJ Blockers in "Official" CICM Textbooks

Drugs	Katzung (2018)*	Goodman & Gilman (2011)**	Smith (2016)	Stoelting (2015)
d-tubocurarine	>50 min	80 min	-	84 min
Mivacurium	-	15-21 min	1.8 min (cis-trans), 53 min (cis-cis)	2 min (cis-trans), 68 min (cis-cis)
Atracurium	20-35 min	45 min	17-21 min	21 min
Cisatracurium	25-44 min	45-90 min	22-29 min	17.3 min
Pancuronium	>35 min	85-100 min	69–161 min	132 min
Vecuronium	20-35 min	40-45 min	31-80 min	73 min
Rocuronium	20-35 min	36-73 min	73 (66–80) min	71 min
Suxamethonium	<8 min	6-11 min	2.7– 4.6 min	47 seconds

To sow confusion among the trainees,

* Katzung offer an "approximate duration of action" instead of a half-life

** G&G give a "clinical duration" instead of a half-life

And this is fine. This is normal and good. Consider the difficulties in assaying the concentration of a drug that is at every moment threatening to fall apart into daughter compounds from no challenge more vigorous than the gentle chemical attention of warm water molecules. It would be more disturbing and unbelievable if these textbooks all reported the same numbers, as it would be entirely implausible that every study everywhere had come to the same conclusion (among a variety of animals, human patients, anaesthetic techniques and experimental settings). In short, reader, life is an ocean of uncertainty, and as an intensivist one must come to terms with this, and learn to just be chill. 

Pharmacodynamic effects of NMJ blockers

The drugs we describe as "neuromuscular junction blockers" all exert their effect by one of two main mechanisms:

• Agonist effect at the nicotinic receptor, with sustained depolarisation of the neuromuscular junction
• Competitive antagonist effect at the nicotinic receptor, preventing the normal neurotransmitter action of acetylcholine

These have significant differences in their direct neuromuscular effects, as well as their side effects. There is probably some merit in knowing some of the details here, as both anaesthesia and ICU training programs tend to have a lot of exam questions focused on the practical uses of this group of agents. 

Pharmacodynamics of depolarising NMJ blockers

Without recapping a lot of the material covered by the suxamethonium chapter, the following short summary of depolarising block is left here to simplify revision:

• Depolarising blockers are acetylcholine receptor agonists
• They are sufficiently similar to acetylcholine, from a chemical point of view, that they are able to bind to the nicotinic receptor
• When they bind to the receptor, the pore is open, and the motor end-plate is depolarised.
• When the endplate depolarises, voltage-gated sodium channels open, and some transient muscle contraction occurs.
• These sodium channels then close, and become inactivated. They will not become active again until the membrane is repolarised.
• However, the membrane will not become repolarised any time soon, because the depolarising agents are resistant to acetylcholinesterase and persist at the receptor, keeping the pore open.
• As a result, no transmission of action potentials can occur from the motor endplate, as adding more acetylcholine into this synapse will only open more channel pores and depolarise the membrane even further, preventing the sodium channels from becoming activated. This block persists until the endplate membrane potential returns to normal and the sodium channels become active again.

In short, the inactivation of perijunctional voltage-gated sodium channels is the main reason for the failure of NMJ transmission due to depolarising block. By comparison, non-depolarising blockade is much easier to grok:

Pharmacodynamics of non-depolarising NMJ blockers

This can be summarised as "they block the receptor", and everything else said about these drugs would be a verbose embellishment of that one basic statement. 

• Non-depolarising NMJ blockers are antagonists of the nicotinic receptor at the neuromuscular junction
• When they bind to the receptor, they do not produce a conformational change that would open the channel pore, but the binding site they occupy is meant to be occupied by acetylcholine, which makes them competitive antagonists
• Their binding is transient and dynamic, with constant frequent association and dissociation of molecules, which means that acetylcholine and NMJ blocker are always in equilibrium at the receptor
• Thus, increasing the concentration of NMJ blocker at the synapse gradually decreases the number of receptors available for binding acetylcholine, which gradually decreases the amplitude of the endplate potential.
• Eventually, the endplate potential becomes so low that it does not reach the threshold required to activate the voltage-gated sodium channels, which means an action potential cannot propagate from this neuromuscular junction site.

According to Appiah-Ankam & Hunter (2004), your twitch amplitude will remain mostly unchanged until the blocker is illegally squatting in 70-80% of your receptors, and then the neuromuscular junction progressively fails until complete block is achieved at about 92% occupancy. If these numbers sound made-up (because how could they possibly be accurate for all possible agents?), the reader can be reassured that they are real, and largely consistent across different blockers. Most textbooks with the decency to do proper referencing tend to point to a 1967 paper by Paton & Waud as a means of backing this assertion. The authors measured the extent to which endplate depolarisation by classical depolarising blockers was antagonised by non-depolarising agents, to determine whether there was a "margin of safety" in neurotransmission, i.e. how much of the junction would need to be disabled before it stops working. The 70-80% value (specifically, 76%, +/- 5%) and the 92% value (specifically, 91.7%) come from this paper. Occasionally confused authors also point to the results of an animal experiment by Douglas and Barbara Waud (1972), which used the same basic technique (a method of their own devisement), reporting data from cat and dog diaphragms and largely reproducing the results observed by the earlier study. 

Phases of block 

"Phases" of neuromuscular block is a term occasionally encountered in the literature to describe the change in the effect of depolarising blockers that occurs with sustained use. Without going into too much detail for something covered much better in the suxamethonium monograph, it will suffice to mention that the term "phases" was first introduced probably by Stephen Thesleff (1955), who described the detailed physiology of depolarising block in terms of two distinct stages. The first stage was characterised by the depolarisation of the membrane during which the junction remained insensitive to any amount of acetylcholine. This was followed by a stage of repolarisation, during which the membrane recovered its resting potential, and which could be overcome with a sufficient amount of acetylcholine (for example where neostigmine was used to block acetylcholinesterase). 

Complete neuromuscular junction block looks more or less the same, no matter which agent is responsible. There is no response to electrical stimulation. Only partial block will be different, for depolarising and non-depolarising agents. The main difference will be in terms of the response to sustained twitches (train-of-four) and tetany. 

Fasciculations are not really a phase of block, but an early phenomenon witnessed clinically with the use of depolarising neuromuscular junction blockers, which has several (possibly inaccurate) theories to explain it. The "phase" of fasciculations is followed by Phase I block.

Phase I or accommodation block is the flaccid paralysis that develops as the result of acetylcholine receptors remaining open, with the resulting depolarisation of the membrane rendering the neuromuscular junction insensitive to further acetylcholine release. The main features of Phase 1 block are:

• A decrease of twitch amplitude
• An absence of fade during train-of-four testing (i.e. the low-amplitude twitches all remain the same low amplitude during testing)
• Absence of post-tetanic potentiation
• A potentiation with the use of acetylcholinesterase blockers (block deepens with neostigmine)

Phase II block occurs following large doses of depolarising blocker, or following the repeated administration of a depolarising blocker. It is characterised by electrophysiological features of non-depolarising block:

• The presence of fade during train-of-four testing
• The presence of post-tetanic potentiation
• Reversal with the use of acetylcholinesterase blockers

Desensitization block is the term occasionally used to refer to a  physiological state of acetylcholine receptor closure during which the receptor is closed, and is not susceptible to opening in response to agonists (including acetylcholine). This is distinct from a Phase I block (where it is fixed open) or a Phase II block (where it is competitively antagonised). It is thought to be a physiologically normal state which these receptors can normally occupy during their routine function (a safety feature, supposed to prevent harmful repeated depolarisation, they say), and can develop due to the presence of high levels of agonists or acetylcholine itself at the neuromuscular junction. It is characterised by:

• Increased sensitivity to non-depolarising agents
• Lack of reversal with the use of acetylcholinesterase blockers

Non-depolarising block resembles Phase II block when it is wearing off, and is characterised by:

• The presence of fade during train-of-four testing
• The presence of post-tetanic potentiation
• Reversal with the use of acetylcholinesterase blockers

And because the chapter on suxamethonium pharmacology is a much better resource for the precise mechanisms of all these features, no further attention will be paid to them here. 

Factors affecting the rate of onset of NMJ blockade

Because so many past paper questions ask about the factors that affect the speed of onset of neuromuscular junction block, the topic receives attention in a chapter all to itself, and only this brief summary will be offered here, in order to spare the reader some time:

Those speed-of-onset SAQs are:

• Question 12 from the first paper of 2016
• Question 9 from the first paper of 2014 
• Question 3 from the second paper of 2009

and for the reader short on time, this past ANZCA answer from ketaminenightmares.com offers a perfectly concise summary and structure.

Factors affecting the potency of NMJ blockers

The reader, after a short review of the grey box above, may come to the conclusion that a lot of the factors that affect the speed of onset will also affect the potency of NMJ blockers. This would be completely normal and apparently incorrect, according to examiner comments to Question 7 from the second paper of 2014 ("some confusion also existed confusing speed of onset kinetics with potentiation kinetics"). What additional material did they expect? Hard to know; but commentary from the examiners suggested that pharmacokinetic, as well as pharmacodynamic, elements needed to be included, and that this pk/pd structure would have been better than a structure that walks along the synapse and points out the relevant factors (which was apparently one of the common styles used by the candidates in that paper). Even though all of this might make the reader thing that entirely novel concepts were expected here, distinct from the speed-of-onset SAQs, the list of factors offered in the comments for Question 1 from the second paper of 2008 seems to be very familiar, and definitely contains a lot of factors that would affect both speed and potency:

"...Examples include drug interactions (anticholinesterases, aminoglycoside antibiotics, local anaesthetics, steroids, antiarrhythmic drugs, anticonvulsants (phenytoin), diuretics, magnesium, lithium), hypothermia / hyperthermia, acidosis, [k+], burn injury, allergic reactions, gender, altered elimination due to renal or hepatic dysfunction and disease states (adrenocortical dysfunction, myasthenia, myopathies, denervation injury). Also extremes of age and pregnancy."

So: to restate material already well covered in another chapter might seem wasteful, but given that the pass rate for the 2014 version of the question was 15%, the critical hit to SEO was absolutely worth it. What follows is a reorganisation of the list of factors that influences the speed of onset, modified to fit into the recommended structure:

Non-neuromuscular effects of NMJ blockers

Considering that most people would give neuromuscular junction blockers mainly to block the neuromuscular junction, it would be fair to call these effects "side effects", or even "adverse effects", as the majority of them are undesired. In the (highly likely) event that some CICM examiner one day decides to ask this question, the trainees would probably benefit from dividing the adverse effects of these drugs into categories:

What follows is an attempt to add some longform explanations for these briefly summarised effects, so as to answer the natural "why it do that?" questions which might come up in the minds of the curious trainees. Whereas of course trainees without curiosity are reminded that, apart from the brief summary above, nothing else is likely to ever be required of them, and that the details here can be safely forgotten once they are well on their way towards a lucrative career in anaesthesia.

Adverse effects related to NMJ blockade

It might seem counterintuitive to include "NMJ blockade factors" in the section discussing non-NMJ-related adverse effect profiles of NMJ blockers, but there are definitely associated and unintended effects related to paralysis itself which can ruin the mood of an otherwise peaceful theatre list. Of these, the most egregious is probably awareness during anaesthesia, a nightmarish circumstance where a patient undergoing a painful procedure may be aware and experiencing it without being able to signal their discomfort. In the ICU the situation is perhaps worse by orders of magnitude, as the patient may be paralysed for many days without relief, suspended in a dreamlike state by incompletely anaesthetic doses of opioids and sedatives. Ballard et al, in 2006, documented horror narratives recalled by survivors, in case the reader is interested in such things. 

Other immediate adverse effects of paralysis, which can be traced directly to but which are not the desired effects of muscle relaxant use, include accidental respiratory catastrophe (e.g where you become disconnected from the ventilator but are unable to breathe on your own) and the inevitable development of pressure injuries resulting from one's inability to self-reposition. These need to be listed but they are predictable and not especially interesting from an exam point of view.  A much more consequential outcome of sustained NMJ blockade specific to the intensive care environment is a rapid deterioration in muscle power and bulk, leading to profound weakness and slowed recovery. This probably requires its own subsection:

Muscle wasting and weakness due to sustained NMJ blockade

Though it is only one of the many factors contributing to ICU-acquired weakness, the use of sustained NMJ blockade is widely suspected to exacerbate the normal muscle-melting effects of critical illness. Unfortunately critical illness itself is a massive confounder, insofar as there is never a need to paralyse somebody for a week outside of the setting of critical illness, and the "pure" effects of sustained neuromuscular junction blockade on muscle weakness are therefore submerged in a highly catabolic cytokine soup. In short, the perception that NMJ blockers somehow contribute to this process are perhaps not entirely supported by the available evidence, leading to what Puthucheary et al (2012) referred to as an unfairly "cautionary and conservative approach" to the use of these agents.

What data we have to support this concern is mainly physiological and observational. Physiologically, it would make some sense that neuromuscular junction blocker should cause muscle wasting independly of any other factor, as they interrupt neuromuscular transmission, and tonic innervation is a known trophic stimulus for muscle cells. When Tomanek & Lund denervated some guineapig ankles in 1973, they observed a marked loss of muscle mass, with a decrease of individual fibre diameter down to 64% of control values over 21 days. In the same year, Riley & Allin did the same thing to some cat tails, and demonstrated that the muscle mass can be preserved with tonic (electrical) stimulation. 

Moreover, a second mechanism has been proposed, which relates the effect of curare toxins to the myopathy by blaming their steroid structure. This seems to be mentioned a lot, and the mentioners tend to refer to Griffin et al (1992) and Kaplan et al (1983), whose works are basically case reports wagging long tails of speculation. The authors referred to steroidal structures to explain the increased susceptibility to myopathy, and pointed to elevated CK levels among their patients to demonstrate that it was taking place. The finding that nonsteroidal agents such as cisatracurium are not associated with an increase in critical illness polyneuromyopathy seems to either support this hypothesis, or undermine the entire idea of NMJB-associated weakness.

Observationally, we see patients lose muscle mass in the ICU, and wherever they are sick enough to require NMJ blockade for prolonged periods, they appear to experience a greater degree of muscle wasting. But how to separate the effect of the blockers from the effects of sepsis, trauma, burns, ARDS or the effect of steroids? Some of the most widely cited papers on NMJ blocker-associated weakness come from experience with asthmatics, receiving NMJ blockers alongside huge doses of steroids (unknown quantities in Adnet et al, 2001, and around 100-200mg/day of methylprednisolone in Kesler et al, 2009).  Puthucheary et al (2012) listed a whole range of such observational studies and concluded that those that sent the strongest positive signal were also the ones that suffered from the greatest methodological flaws, and that there was only a weak and non-statistically significant signal among the noise. Overall it appears that septic patients probably get more weakness if they are exposed to prolonged NMJ blockade, but on balance one might have to agree that the ability to ventilate them properly is probably more important for their survival than the ability to decannulate their tracheostomy on day 35 instead of day 36 of ICU stay.

Adverse effects related to ganglionic stimulation by NMJ blockers

As mentioned elsewhere, nicotinic acetylcholine receptors are widespread, and it is only the chance alignment of quaternary ammonium molecules that keeps NMJ blockers from affecting all the other varieties of nicotinic receptor. And even this selectivity is imperfect, particularly at large doses, and for selected agents. Specifically, some NMJ blockers also bind to the nicotinic receptors of the autonomic ganglia. As the result, NMJ blockers can theoretically have effects on both the sympathetic and parasympathetic nervous system (and conversely, some ganglionic blockers can have a muscle relaxant effect).

There are several good references to describe these effects ( Flynn & Goldhill, 1990; Baraka, 1981; Kampe et al, 2003) but unfortunately they often disagree as to the magnitude or even the direction of the effect. It does not help that some of the studies  were performed in rats, and some in merely rat atria. For the reader whose main interest is to pass the CICM exams, a single source of truth is more important than the validity of reliability of the experimental results, which means this table from page 482 of the 14th edition of Katzung is more than enough. 

Ganglionic Effects of Neuromuscular Junction Blockers

Drug	Ganglionic effects	Reference
d-tubocurarine	Ganglionic blocker	Kim et al (1981)
Mivacurium	No direct ganglionic effects	Gursoy et al (2011)
Atracurium	No direct ganglionic effects	Ghorbanlo et al (2016)
Cisatracurium	No direct ganglionic effects	Ghorbanlo et al (2016)
Pancuronium	Mild ganglionic effect	Birmingham et al (1980)
Vecuronium	Minimal ganglionic effect	Melnikov et al (1999)
Rocuronium	Minimal ganglionic effect	Melnikov et al (1999)
Suxamethonium	Ganglionic stimulant	Ertama et al (1984)

As you can see, the main ganglionic actors are the older agents, tubocurarine and pancuronium. Of these tubocurarine is the more potent ganglionic blocking agent, as a much lower dose is required to produce the same level of block - it is about half as potent as hexamethonium, whereas the EC50 of pancuronium required for ganglionic block is about ten times greater than tubocurarine (Birmingham & Hussain (1980) have an excellent table of EC₅₀ values for ganglionic block among a range of older agents, in case anyone is interested).

Direct autonomic effects of NMJ blockers

When we say" direct" here, it mainly means "directly on the specific branch of the autonomic nervous system", but also possibly "directly on the specific receptors in question". Nicotinic and muscarinic receptors have the same endogenous ligand, and so it is not so farfetched to consider the possibility that NMJ blockers might also have parasympathetic effects; and some of them also have indirect sympathomimetic effects.

Drug	Muscarinic effects	Sympathetic effects	Reference
d-tubocurarine	Vagolytic	Sympathomimetic	McCulloch et al (1972)
Mivacurium	No muscarinic effects	nil	Savarese et al (2004)
Atracurium	No muscarinic effects	nil	Ghorbanlo et al (2016)
Cisatracurium	No muscarinic effects	nil	Ghorbanlo et al (2016)
Pancuronium	Vagolytic	Sympathomimetic	Domenech et al (1976)
Vecuronium	Mildly vagolytic	minimal	Gursoy et al (2011)
Rocuronium	Mildly vagolytic	minimal	Gursoy et al (2011)
Suxamethonium	Cholinergic stimulant	minimal	Ertama et al (1984)

The effect of tubocurarine and pancuronium on the sympathetic nervous system actually seems like a cholinergic postganglionic effect, i.e. by acting as ganglion blockers with higher selectivity for sympathetic ganglia these drugs tend to disinhibit those noradrenergic postganglionic fibres and enhance catecholamine release from their terminals. Of all the other drugs, none have a really convincing direct effect on the sympathetic nervous system, but they can certainly activate it indirectly by means of causing histamine-mediated vasodilation:

Adverse effects related to histamine release by NMJ blockers

It appears that all neuromuscular junction blockers can stimulate histamine release. According to Salvatore Basta (1992), there are at least three different mechanisms behind this. Of these, antigen-mediated release (i.e. a"an allergic reaction") is by far the least common. Another less common mechanism is the binding of the drug to some surface immunoglobulins on the mast cells, which results in a complement-mediated histamine release. By far the most common mechanism is a directly chemical stimulation of histamine release from mast cells which appears to occur on contact with the agent and which does not have any immunological basis ("anaphylactoid" is sometimes the term given to such reactions). Skin weal experiments by Galletly (1986) and venous blood measurements by Naguib et al (1995) were used to construct the table below, seeking to add some ranking or order to the list of agents so the reader may appreciate which among them is the most histaminergic:

Drug	Local histamine release *	Systemic histamine release **
d-tubocurarine	22.6	252%
Mivacurium		370%
Atracurium	17.3	234%
Cisatracurium
Pancuronium	1.0
Vecuronium	1.5	0%
Rocuronium		0%
Suxamethonium	0.1

* The arbitrary-looking numbers here are a metric used by Galletly (1986), who compared the potency required to cause a cutaneous weal to the potency required for an ED₉₅ of neuromuscular effect, as some kind of ratio. To put this in simple terms, the higher the value, the more histamine-releasy the agent.

** The arbitraty-looking percentages here are the increases in measured plasma histamine concentrations which were found by  Naguib et al (1995)

In short, tubocurarine is probably the most histaminergic agent, followed probably by mivacurium. Cisatracurium, unlike atracurium, does not appear to have any significant histamine release associated with it, within the normal clinical dose range.

So:

The net haemodynamic effects of NMJ blockers

The act of reading a large number of articles across which this information is scattered leaves one with the distinct impression that the net total of the haemodynamic effects of these agents is of absolutely no interest to anybody, because if it did then surely there would be some kind of helpful table somewhere that would summarise all these data. There is no such table, and among the authors who do publish on the subject, there is some disagreement as to the exact effects.  It does not help that everyone gives different co-analgesia and co-anaesthetic agents to their experimental subjects. What follows was mainly derived from Flynn & Goldhill (1990), for no reason other than the convenience of a large bulk of the information being gathered in the same paper.

Drug	Heart rate	MAP
d-tubocurarine	Mildly increased	Markedly decreased
Mivacurium	Mildly increased	Decreased
Atracurium	Basically unchanged	Decreased
Cisatracurium	Basically unchanged	Mildly decreased
Pancuronium	Markedly increased	Markedly increased
Vecuronium	Mildly decreased	Basically unchanged
Rocuronium	Basically unchanged	Basically unchanged
Suxamethonium	Decreased	Decreased

Idiosyncratic adverse effects of individual NMJ blockers

Apart from adverse haemodynamic effects and unexpected consequences of NMJ blockade, there are a few unique idiosyncratic complications that occur as the result of the use of these agents. Of these, the most memorable would probably have to be the side-effects of suxamethonium, discussed in detail in a chapter that covers this drug exclusively. In brief, these are:

• fasciculations
• hyperkalemia
• Increased intraocular, intragastric, intracranial pressure

The other main adverse effect is the lowering of the seizure threshold by laudanosine, the metabolite of atracurium. Fodale & Santamaria (2002) cover this thing in more detail than the CICM trainee would ever be expected to know. For the time-poor candidate, the following point-form description is all that is required:

• Laudanosine is a metabolite of atracurium and cisatracurium
• It forms spontaneously as the breakdown product of these agents and can increase in concentration during storage
• Each molecule of the parent agent is degraded into two molecules of laudanosine
• Cisatracurium produces fewer of these molecules mainly because it is more potent, and the molar concentration of the parent drug is lower
•  It undergoes mainly hepatic metabolism
• Its main consequences are neurotoxicity
• It penetrates the blood-brain barrier and can accumulate in the CSF, where it appears to have a stiumulant effect, which could potentially lead to earlier arousal from anaesthesia, or even awareness while paralysed
• An extension of these stimulant effects is the tendency of laudanosine to reduce the seizure threshold. 
• This proconvulsant effect is probably largely theoretical, as in experimental animals EEG changes seem to occur at doses 8-10 times greater than what would normally be expected from a standard infusion.