Anatomy of the sympathetic nervous system

This chapter is relevant to Section M (i) of the 2017 CICM Primary Syllabus, which expects the exam candidates to "describe the autonomic nervous system, including anatomy" among several other things. One might, having manipulated one's share of sympathetic neurotransmitters in the course of doing this Intensive Care Specialist Job, have concluded that the anatomy of the sympathetic nervous system is by far its least important property, but CICM examiners clearly attribute some value to this knowledge, and have asked at least 1.7 written paper questions about it: 

• Question 20 from the first paper of 2015 
• Question 5 from the second paper of 2020 

One of these was 30% physiology, whereas the other was 100% anatomy, and the pass rate for the latter was only 25%, because probably this sort of thing just doesn't come up in routine clinical practice (whereas the physiology of the SNS is a common topic of discussion, what with all these litres and litres of noradrenaline sloshing around). Admittedly, anatomy of the sympathetic nervous system does have a clinical relevance, particularly in the context of spinal injury where it is interrupted by trauma, resulting in neurogenic shock; but for whatever reason this practical aspect does not appear to have been tested in the CICM exams at any level.  

In summary: 

Structure of the Autonomic Nervous System by Giorgio Gabella (1976) ended up being the best organised and most reliable resource for this largely anatomical section, and it remains relevant fifty years later because human anatomy has not changed very much on that sort of time scale. The other excellent resource is the autonomic system entry from Comprehensive Physiology by Wehrwein et al (2016)

Overall organisation of the sympathetic nervous system 

The reader, who has likely landed here from some kind of search engine result, will usually expect to be assaulted with a diagram that tries to fit the entirety of the sympathetic nervous system into a single image. This common approach gives rise to some truly intimidating posters in the foyers of medical schools and human health sciences faculties, hung there to bewilder and confound the visitor, impressing upon them the complexity and exclusivity of the eldritch knowledge wielded by their professors. Of these, here is an excellent example from Patel (2015): 

It felt appropriate to inflict this image on the reader mostly because it fits in with the ethos of this website, which is to pursue completeness at the expense of clarity. In truth it is an excellent representative of its genre and contains all of the most important elements, even though the visual design is cluttered with unnecessary images of organs (as it is not entirely clear how the educational characteristics of the diagram were improved by both labelling “bladder, penis, gonad” and depicting the pelvis). This diagram is also exceptional because it affords the author an opportunity to crow over the terms “sudomation” and “pilomation”, which do not exist in the English language according to Google, but which likely refer to the control of sweat glands (sudomotor effects) and the movement of cutaneous hair (pilomotor effects). 

Diagrams are a common feature in written exam answers, and where the CICM examiners can go with this one, nobody exactly knows, but we can be generally rather sure that they will not expect their trainees to reproduce the crossections of pelvises as a part of their response. Some kind of simplified diagram is needed, which contains all of the most important structures, and which can be embellished with additional detail if this is what is called for. A suggestion for this is presented below and probably represents some kind of passable minimum. Any less than this and one risks being one of those people who “had a simple sketch understanding of the question asked but could not add enough of the next layer to be awarded a pass mark” 

This reduces the complexity of the usual depictions of this system into a comforting linearity, at which the lecturer and student both can take a sigh of relief. The discussion that follows can now take a path from the most proximal structures down to the most distal, stopping to note the landmark attractions along the way. CICM exam trainees are as always invited to contemplate the possibility of other, equally valid structures, and to devise their own.  

Central control structures of the sympathetic nervous system 

Immediately upon looking at this topic one comes to realise that very little is actually known about the workings of the uppermost levels of control for the SNS, such that any discussion of them might be entirely pointless for the purposes of exam revision (because where would the examiners find the answer to the question they asked, if the textbooks are vague or silent?) However it is still possible to point the reader to such excellent reviews as Sun (1995) or Coote & Spyer (2018) whose main points are oversimplified below: 

• Brainstem structures, such as the rostral ventrolateral medullary vasomotor centre, raphe nucleus and the nucleus tractus solitarius, are simple control relays which provide the coupling between the afferent and efferent limbs of cardiovascular and respiratory reflexes. These appear to function fairly independently (hence autonomic) but do receive a lot of regulatory inputs from higher structures. Their connections were revealed in a clever study by Amendt et al (1979), where the investigators injected horseradish peroxidase into the sympathetic neurons in the spinal cords of cats at the level of T3, and watched its retrograde transport along axons into the abovementioned brainstem nuclei.  
• Hypothalamic structures such as the paraventricular nucleus are also described as important (pivotal, as per Coote & Spyer) to autonomic function, but the exact details remain murky, other than to point to the fact that they project to the medullary reflex nuclei but can be destroyed without much loss of function, suggesting that these mainly modulate the medullary reflexes. 
• Cortical structures appear to have direct sympathetic connections; for example the right anterior insular cortex is implicated in the control of cardiovascular reflexes, and more generally higher cortical functions can clearly produce autonomic effects (eg. emotional stress). Again, autonomic reflexes are preserved when higher cortical structures are destroyed while the brainstem is spared. 

For the intensivist (taking exams out of the equation for a second), the practical importance of these structures and relationships is twofold: firstly, some are susceptible to pharmacological manipulation (eg. by centrally acting sympatholytic agents such as clonidine), and secondly, it becomes important to know a little about them because of the effects of CNS injury on autonomic function. Without trespassing too far into clinically relevant Second Part Exam territory, the reader is teased with links to articles on dysautonomia following brain injury (Baguley et al, 2008), autonomic dysreflexia due to spinal injury (Allen & Leslie, 2018), and the use of CNC-autonomic interactions to investigate disorders of consciousness (Riganello et al, 2019). These rabbit holes can probably wait until after one has passed their First Part.  

Sympathetic nervous system in the spinal cord 

The sympathetic contents of the spinal cord is represented by the cell bodies of the preganglionic neurons, which sit in the intermediolateral nucleus. This indistinct group of cells occupies an area of the spinal grey matter in the lateral column (Bror Rexed’s lamina VII). Below, an image of a mouse spinal cord from Sengul & Watson (2012) illustrates the scale and position of this structure: 

That stain is an immunohistochemical label that targets choline acetyltransferase, the enzyme responsible for making all that acetylcholine used by these preganglionic neurons. From here, the preganglionic neurons send their projections to the ganglia. From studies on non-human mammals, it appears that these neurons send their axons down the ramus at their spinal level, i.e. they appear to be organised segmentally.

Sympathetic preganglionic fibres 

These fibres are generally described as “Type B” fibres, following the standard classification of nerve fibres. Yes, they are mostly myelinated, which gives the white ramus the white colour, but otherwise they are the slowest of the myelinated fibres. Eccles (1935) measured their velocity and reported a figure of about 12 m/sec. This is fine, as most of them are very short - 

There are several possible things these fibres can do: 

• Synapse with a postganglionic neuron at some point in the sympathetic chain (at, above or below the level of its nerve root exit) 
• Pass directly through the sympathetic chain and travel along some abdomino-pelvo-viscero-splanchnic nerve, of which there are several: 
  • The greater splanchnic nerve 
    • Originates from T5-9 
    • Synapses with cell bodies in the coeliac ganglion
  • The lesser splancnic nerve 
    • Originates from T10 and T11
    • Synapses with cell bodies in the coeliac ganglion
  • The least splanchnic nerve 
    • Originates from T12
    • Synapses with cell bodies in the renal ganglion
  • The lumbar splanchnic nerve 
    • Originates from L1 and L2
    • Synapses with cell bodies in the aortic plexus
  • The sacral splanchnic nerves
    • Originates from T12-L2
    • Synapses with cell bodies in the inferior hypogastric plexus, the superior hypogastric plexus and the aortic plexus

These are confusing, because there are other sympathetic splanchnic nerves which contain postganglionic fibres (eg. the cardiopulmonary splanchnic nerves), and apart from memorising which is which, there is no elegant system or basic principle to describe the distribution of these, to aid the recall of an exam candidate. Fortunately, there is no possible way any CICM examiner could ever be interested in this level of minutiae. Likely, the only thing one could be expected to know about these nerve fibres is that their terminals release acetylcholine at N2 nicotinic receptors of the sympathetic ganglionic neurons.

White and grey rami

The preganglionic fibres, coursing from their cell bodies in the intermediolateral nucleus, exit the spinal cord in a way that always seems to confound the illustrator. For whatever reason, the text prompt "path of preganglionic fibres in the spinal nerves" seems to generate this sort of image:

This is a representative one from Wikipedia, and is perhaps the clearest of them all. Unfortunately, this is another one of those situations where trying to represent the accurate anatomical shape and relationship of structures actually undermines the explanatory potential of the diagram. It is, however, possible to explain this in an unordered point-form list:

• Intermediolateral cell bodies send efferent fibres anteriorly into the ventral root
• From the ventral root, the fibres course into the spinal nerve, and then into the ventral (anterior) primary ramus of the spinal nerve
• From the ventral primary ramus of the spinal nerve,  preganglionic fibres exit via the white rami, so called because they are more full of myelin, and appear more pale. 
• The white rami join the sympathetic chain ganglia, where the fibres can either:
  • Synapse with a ganglionic neuron at the same level
  • Travel up or down the sympathetic trunk to synapse with a ganglionic neuron at some other level
  • Pass through the sympathetic trunk without synapsing, and carry on along a splanchnic nerve to synapse with a distant gangion somewhere peripherally
• The fibres that synapse with a ganglionic neuron send postganglionic fibres that return to spinal nerve via the grey rami, so called because they are darker and thinner, being composed of largely unmyelinated fibres.
• These postganglionic fibres then carry on with other somatic nerves to innervate their peripheral targets (vessels, skin, sweat glands, etc)

Alternatively, one could try to plot these pathways as a simple circuit flowchart, carefully treading around the edges of the pit that contains all the other anatomical illustrators in history:

Sympathetic ganglia 

For some reason, the best resources for this specific section of the sympathetic nervous system all come from the exotic out-of-print textbooks from the 1970s.  Furness & Costa (1974) and Gabella (1976) were the most interesting here, mostly as sources of interesting digressions. There are also extensive detailed articles by du Plessis & Loukas (2022), divided into Part 1,  2 and 3.  Unfortunately, any attempt to identify ICU-relevant structural information here was confounded by the realistic concern that the intensivist will rarely be in a position where they are the highest authority on this anatomy, and where somebody's life depends on their accurate knowledge of it. The act of researching detailed or definitive anatomical information here was therefore frustrated by the constant oppressive sensation of pointlessness. With this caveat, what follows is a summary of the most important notable elements:

• Paravertebral ganglia are lined up along the spinal column. There are about 24 paravertebral ganglia on each side, forming the sympathetic chains. There's generally one ganglion for each spinal level, except in the neck where there are only three (superior cervical ganglion, stellate ganglion and intermediate ganglion)
  • Superior cervical ganglion is formed by sympathetic fibres fibres of the upper four spinal levels, running up the vertebral artery. Its a mass of postganglionic cell bodies about 3cm long, and it is positioned in front of the lateral mass of the atlas and axis.
  • Middle cervical ganglion sits in front of the vertebral artery at the level of around C6.
  • Inferior cervical ganglion lies behind the origin of the vertebral artery.
    ment of the vertebral artery. It may not exist, instead being absorbed into the stellate ganglion.
  • The stellate ganglion is a fusion of the first thoracic ganglion and the inferior cervical ganglion, sitting in front of the neck of the first rib.
• Prevertebral ganglia are not necessarily bilaterally symmetrical, forming a network of nervous tissue structures ventral to the abdominal aorta. These consists of the coeliac plexus, aortic plexus and the superior hypogastric plexus, each of which consist of, contribute fibres to, a bunch of ganglia. Or sometimes textbooks may refer to the ganglia themselves, and not to the plexuses (plexi? Plexae?). As you can see, the anatomy is messy and the connections are difficult to map, which seems to have given rise to an anarchy among anatomists, such that no online resource agrees with another with regards to what is innervated by what, or how the systemic best organised. When this sort of confusion arises, the reader is directed to Last's Anatomy Regional and Applied (the revised 9th edition, not whatever the latest edition is), as this is the recommended text of the Royal Australasian College of Surgeons, and is therefore canonical with respect to anatomy. 
  • Coeliac ganglia are two semilunar masses of nervous tissue that wrap around the coeliac trunk artery,  medial to the adrenal glands and anterior to the crura of the diaphragm.  Fibres from this ganglion innervate the lower oesophageal sphincter, stomach, liver, pancreas, spleen, and about half of the duodenum (though it is not clear where the jurisdiction of the coeliac ganglion ends, and the territory of the superior mesenteric ganglion begins).
  • Superior mesenteric ganglion  sits in the retretroperitoneum at the origins of the SMA, and sends fibres to innervate basically all of the intestine, from the lower half of the duodenum down to the transverse colon (Patel et al, 2021)
  • Aorticorenal ganglia invest the renal arteries bilaterally and innervate the kidneys.
  • Inferior mesenteric ganglion innervates the rest of the colon and rectum. It wraps around the origin of the inferior mesenteric artery.

The CICM trainee will likely never encounter these structures directly, except as drug targets, and so it would be fairly pointless to make detailed notes about their anatomical relations. Some points of clinical relevance to help connect the names to familiar conditions are probably still valuable, and can be listed here to help create a cognitive scaffold for the exam candidate:

• The stellate ganglion can be blocked to relieve refractory VT, and the procedure sounds completely wild, as it involves advancing the needle alongside the trachea at the level of the cricoid cartilage, until it basically hits the body of C6. 
• The coeliac ganglion can be radiofrequency-ablated or blocked with local anaesthetic to relieve the pain of chronic pancreatitis or pancreatic cancer
• The aorticorenal ganglia can be among the targets for sympathetic renal denervation of the renal arteries, which is a method of managing intractable hypertension

Postganglionic sympathetic neurotransmission

Most of the cell bodies contained in sympathetic ganglia are adrenergic, i.e. the main neurotransmitter secreted by them is noradrenaline. There are a few exceptions:

• If one regards the chromaffin cells of the adrenal medulla as just a heavily militarised  version of postganglionic neurons, then about 80% of them they are unique in that they have the right enzymes that methylate noradrenaline to make adrenaline instead. If you are then prepared to extend the metaphor even further, then the systemic release of this catecholamine is  "postganglionic neurotransmission".
• A minority (4%) population of neurons in sympathetic ganglia are cholinergic, i.e. they release acetylcholine just like the preganglionic and parasympathetic fibres, and the postsynaptic receptors of their target organs are muscarinic. These are the sudomotor neurons, i.e. they innervate the eccrine sweat glands exclusively, and this seems to be consistent across all species of sweaty mammals. It appears they are born noradrenergic but undergo a phenotype switch when they connect to the sweat glands (Schütz et al, 2008), which appears to be coordinated by the neurons themselves and does not seem to be related to anything the sweat glands are doing.

Popular online resources suggest that there is also a clone of sympathetic postganglionic neurons that release dopamine. Specifically, these are supposed to be the neurons that supply the kidney and renal vessels.  However, it appears that their existence and activity have an almost mythic quality, or are at least sufficiently questionable that eminent authors ( Bell, 1982) phrase their discussions of them in terms of a "proposal that dopaminergic neurones might exist in the autonomic nervous system". And this is coming from the main player in dopaminergic autonomic research, nine years after the same author (Bell & Lang, 1973) offered dopamine as an explanation for the vasodilator responses they observed in the renal vessels of anaesthetised dogs, when descending central autonomic pathways were stimulated. The vasodilator response was also abolished by haloperidol (a dopamine receptor antagonist), and moreover other papers showed that dopamine receptors exist in the kidney and yet other papers demonstrated that dopamine is a vasodilator there, so the idea seemed plausible.

However it appears that direct evidence to support the existence of a discrete dopaminergic postganglionic system in the kidney was missing, and it remained missing over subsequent decades of study. Articles from the modern era sidestep the need to identify and discuss this subject by mentioning that there might be some dopamine release from sympathetic terminals in the renal arteries, but then dismissing dopaminergic innervation of the kidney as something rather irrelevant to its overall function. For the CICM exam candidate, it is worth knowing that the college-recommended textbooks do not mention this system, judging by this author's quick survey of the editions available to him. Ganong (23rd ed) does not list any dopaminergic receptors in their giant table of autonomic receptors (Table 17-1, p. 267-268), Guyton & Hall (13th ed) makes no mention of renal dopaminergic innervation, nor is it mentioned anywhere in Vander's Renal Physiology (7th ed). In short, the reader interested in passing exams should confidently ignore this controversy, as it should be possible to score marks without having to commit to either of the conflicting beliefs. 

Sympathetic postganglionic fibres 

Transmission to the ganglion is fast along some myelinated fibres (3-15 m/sec), and transmission through the ganglion is also reasonably fast (mostly because nicotinic receptors depolarise the postsynaptic membrane instead of doing sluggish metabolic things).  From there, however, the speed of transmission slows down, as all the postganglionic sympathetic fibres are small-diameter Type C fibres. Fagius & Wallin (1980) recorded a transmission velocity of around 0.74-1.69 m/sec for a selection of cutaneous and muscle vasomotor fibres in healthy subjects. 

Why are these long fibres unmyelinated, one might ask? The target organ is some distance from the ganglion, so surely the rate of transmission would be better if there was some myelin on these fibres? There is probably no educated way of answering a question like this. Indirectly, it appears that they are small and unmyelinated because each fibre innervates a very small target, and that if the axon was any thicker it would end up stimulating myelin synthesis. Voyvodic (1989) was able to demonstrate this experimentally when he managed to get some myelin growing on sympathetic postganglionic efferents by increasing the caliber of the axon. 

Sympathetic nerve endings 

"Beaded strands" is usually the description of these nerve terminals, reflecting the finding that these nerve endings do not have a discrete terminal, but rather a series of catecholamine-laden varicosities that approach their target organ without a specialised contact zone of any sort (i.e. there is no "neuromuscular junction" between these varicosities and, for example, a smooth muscle fibre). In fact often these fibres are loosely laid among their target tissues; Smolen (1988) describes distances of as much as 1-2 μm between sympathetic endings and smooth muscle cells, which is an unheard-of distance for neurotransmission. Varicosities shout in catecholamine from the mountaintops, and in the foggy distance somewhere far a herd of myocytes contract in lazy unison.

This system is clearly widely distributed (much more so than the parasympathetic network) and obviously every minuscule ramification of the vascular tree cannot be individually innervated, otherwise the proximal postganglionic sympathetic nerves would have to be thick as tree trunks.  Sympathetic nerve fibres leverage the ability of smooth muscle to transmit a wave of depolarisation, which means sparse innervation can still be expected to reliably affect all of the target tissue (given time). On the other hand, some tissues of importance are innervated more densely. For example, for some obscure reason each smooth muscle myocyte of the vas deferens is individually innervated with a branch of a sympathetic axon. Which is the beginning of a discussion about the innervation of the tissues by the sympathetic nervous system:

Extent of sympathetic innervation 

 By mass, surface area, or whatever other metric you prefer, the sympathetic nervous system supplies much more of the tissues and organs than the parasympathetic. Sympathetic nerve endings can be found virtually everywhere in the body, and for some of the tissues the SNS is the only division of the autonomic nervous system in control. This works just fine because the SNS has a constant tonic output, which means the tissue or organ in question can be modulated in its activity by the action of increasing or decreasing this output. The tissues and organs that are managed exclusively by the sympathetic nervous system include:

• The adrenal glands
• The majority of the blood vessels (where the parasympathetic nervous system only innervates the helical arteries and sinusoids of the erectile tissues in your reproductive organs and some of the blood vessels to various glands)
• The pilomotor muscles in the skin (hair follicles),
• Sweat glands

One probably also needs to mention that there are some tissues that do not receive direct innervation from the sympathetic nervous system, but which express various adrenergic receptors, which means they can still benefit from the constant tonic baseline secretion of noradrenaline (0.070 to 1.7 ng/mL, corresponding to an infusion rate of 0.6 to 15ml/hr of the standard 6mg/100ml dilution). These include:

• the vast majority of the arteriolar smooth muscle cells (as the blood vessels are sparsely innervated by the SNS, which means many cells will rely mostly on humoral signals),
• the adipose tissue (which receives some direct innervation from the sympathetic nervous system, but which mostly relies on humoral control and expresses α₁ and β3 receptors)
• Bone, which receives sympathetic innervation only via perivascular vasomotor fibres, but which expresses β2 receptors on the surface of osteoblasts and osteocytes, suggesting that some noradrenaline spillover is used for signalling (and it appears to promote bone resorption)
• In fact the entire β-2 receptor system is not directly innervated by noradrenergic sympathetic fibres, as β-2 receptors could not care less for noradrenaline, and respond only to adrenaline. These are certainly a part of the sympathetic nervous system, but are manipulated by hormone-like release of adrenaline from the adrenal medulla.

For completeness, it is worth mentioning that there are also organs and tissues which might receive some sympathetic vasomotor fibres exclusively for blood flow regulation, but which are otherwise controlled exclusively by the parasympathetic nervous system. These include the lacrimal glands, ciliary muscle of the iris (which controls accommodation) and the sublingual salivary gland. 

Another way of looking at the extent of the sympathetic innervation would be to list the thoracic spinal levels responsible for supply to each specific area, such as this table from Wehrwein et al (2016):

 Topographical Organization of Sympathetic Preganglionic Neurons

Segment	Ganglion	Effector targets
T1-T2 T1-T5	Superior cervical Superior and middle cervical	Pupillary muscles of the eye; submandibular, sublingual, and parotid glands Sweat glands and the vasculature of the head and neck; vasculature of the brain; choroid plexus; carotid body
T1-T7	Stellate and other upper thoracic	Heart
T2-T7	Stellate and other upper thoracic	Trachea; bronchii; lungs
T3-T6	Stellate and other upper thoracic	Brown adipose tissue, sweat glands, erector pili muscles, and vasculature of upper extremities
T5-T6	Stellate and other upper thoracic	Esophagus
T4-T12	Adrenal gland	Chromaffin cells of the adrenal medulla
T6-T11	Celiac	Smooth muscle and glands of the stomach; liver; gallbladder; pancreas
T8-T12	Aorticorenal	Tubules of the renal cortex; renal blood vessels; proximal convoluted tubules; glomeruli; pelvic wall
T9-T10	Celiac, superior, and inferior mesenteric	Small intestine; ascending limb of large intestine
T10-L2	Lumbar and upper sacral	Sweat glands, erector pili muscles, and vasculature of lower extremities
T11-L1	Celiac, superior, and inferior mesenteric	Transverse large intestine
T11-L2	Inferior mesenteric, hypogastric, and pelvic	Descending large intestine, colon, and rectum
T11-L3	Hypogastric and pelvic	Urinary bladder; male reproductive system (epididymis, vas deferens, seminal vesicles, and prostate glands); female reproductive system (vagina and uterus)