Cardiac reflexes

This chapter is probably relevant to Section G3(vi) of the 2017 CICM Primary Syllabus, which expects the exam candidate to "describe the cardiac reflexes".  This (totally reasonable) expectation seems to have caught many of us off-guard. The pass rates for the two SAQs on this topic (Question 2 from the second paper of 2013 and Question 17 from the second paper of 2010) were 22% and 27%, respectively, which is surprising because they were neither ambiguously phrased nor unfairly esoteric. Nor is there a lack of representation in textbooks, which tend to focus much attention on this topic. For example, the cardiac reflex chapter in the official college textbook spans approximately 26 pages.

Owing to the massive excess of fascinating material in this topic, this chapter has expanded to such a massive scale that a table of contents became essential. The most convenient mechanism of presenting it was this short summary, at the end of which the time-poor candidate is invited to stop reading, as the rest of the chapter is an exploration of the sort of apocrypha that really puts the "deranged" in "Deranged Physiology".

The college reminds us to use peer-reviewed resources for our revision. For the absolute cream of the peer-reviewed crop, one can blow $735 on a hardcover copy of Zucker & Gilmore's "Reflex control of the circulation", the 1079-page 2020 edition. That probably represents some sort of gold standard of the physiological literature on this subject. From this lofty apogee, the next highest tier consists mainly of specialised articles which deal with each single reflex individually, with no overview papers to occupy the middle ground between these extremes. Hainsworth (1996) and Mancia (1985) are possible exceptions, but neither is by themselves enough for what CICM examiners wanted, nor are they available for free. And if we are going to be discussing resources you have to pay for, you may as well pay for the official syllabus textbook (Pappano & Weir), which covers this topic in Chapter 5 (Regulation of the Heartbeat). The sorry state of the following summary reflects the fact that the author had to crawl a whole series of different references to get some understanding of what these reflexes did and how they behaved.

Definition of cardiac reflexes

In their answer to Question 2 from the second paper of 2013, the examiners defined cardiac reflexes as:

"...fast-acting reflex loops between the heart and central nervous system that contribute to regulation of cardiac function and maintenance of physiologic homeostasis."

This definition is sufficient for exam purposes, not only because it was used by the examiners, but also because there really is no better summary statement to open with. In case anybody is wondering where it originated, one can find a suspiciously similar phrase on page 393 of the 7th edition of Miller's Anaesthesia. Looking more broadly, the term "cardiac reflex" or "cardiovascular reflex" does not seem to have a widely accepted pattern of use, which opens it for definition by nonexperts.  Judging from the way the college answers have been set up, they had preferred us to organise our thinking according to the following framework:

• Sensor and stimulus
• Afferent nerves
• Processor
• Efferent nerves
• Effector

So, if you had to cobble together your own definition, it might sound something like this:

Cardiac reflexes are fast-acting centrally mediated negative feedback mechanisms which maintain homeostatic control of cardiovascular variables. They sense changes in heart rate blood pressure and arterial oxygenation through peripheral receptors, and respond by altering myocardial function and peripheral vascular resistance.

However, one hastens to caution the reader against making use of non-peer-reviewed resources for their revision, and points to the first definition which was quoted by the examiners, and which is therefore definitive.

Stretch mechanoreceptors in general

To be able to react to changes in cardiovascular parameters, you'd have to be able to detect them somehow. The human organism monitors its own blood pressure by means of mechanosensitive receptors which are usually afferent autonomic nerve terminals sitting in the tunica adventitia of large blood vessels and cardiac chambers. These receptors respond to mechanical deformation: basically, they are activated when they are stretched because the wall of the vessel is being distended. 

Chapleau et al (2006) go into extensive detail as to how this crude mechanical stimulus is converted into an action potential. In short, these nerve endings feature a weird coiled structure with bulbous endings, which is covered in mechanically sensitive sodium channels (DEG/ENaC, degenerin/epithelial sodium channels). Those channels are a rather ubiquitous feature of animal neurology, and are responsible for all sorts of mechanotransduction all over the place (eg. when you palpate a Caenorhabditis elegans,  they feel you through these receptors).  Here, immunofluorescence images from Drummond et al (1998) demonstrate such a nerve ending in a rat aorta, glowing with green fluoro-tagged DEG/ENaC proteins:

The sodium current through these channels is not an all-or-nothing phenomenon: rather, it appears that some stretch always activates some current, but most of the time this phenomenon is localised to the nerve ending and goes no further. However, when the stretch is sufficiently ... stretchy... the sodium current increases to the point where the membrane potential reaches the threshold of local voltage-gated sodium channels. Then, the membrane really depolarises, and the mechanoreceptor sends an action potential up its afferent nerve. The more stretch there is, the more frequently these receptors fire. 

The baroreceptor reflex

If one realigns one's definition of what's important in medicine with what is examinable in the CICM primaries, then the baroreceptor reflex takes priority over all the others. Two past paper questions ( Question 16 from the second paper of 2014 and Question 8 from the second paper of 2007) had asked about this reflex arc specifically, with the bar set quite high ("This is a core topic and a detailed knowledge was expected"). Ergo, an entire chapter has been dedicated to the exploration of baroreceptors and their relationship to the clinical management of ICU patients. In order to simplify revision, some brief summary of their activity and function is listed below, but the bulk of the discussion occurs elsewhere.

So, the baroreceptor reflex in a nutshell:

• Stimulus:  Pressure (stretch)
• Sensors: Stretch-sensitive mechanoreceptors in the carotid sinus and aortic arch
• Afferent nerves:
  • Vagus carries afferent fibres from the aortic arch
  • Glossopharyngeal nerve carries fibres from the carotid sinus
• Processor: Nucleus of the solitary tract and the caudal ventral medulla
• Efferent nerves: 
  • Sympathetic fibres to the heart and peripheral resistance vessels
  • Vagal efferents to the cardiac ganglion (heart rate)
• Effector:  Myocardium, SA and AV nodes, vascular smooth muscle

Bainbridge reflex

In an effort to banish all eponyms from medical literature, one might decolonise this phenomenon by calling it the "atrial stretch reflex", which would immediately suggest to you what it does and how it works. With this, of course, no offense is meant to Francis Arthur Bainbridge (1874-1921), an upstanding individual whose RCP biographer for some reason decided to burn him as someone who was an excellent teacher "in spite of an unimpressive appearance". He scored this eponym when he discovered a relationship between atrial distension and heart rate (Bainbridge, 1915). Here, a tracing of heart rate and blood pressure from his original paper is reproduced, with endearing handwritten axis labels:

His own conclusion describes this reflex in a manner which is fairly similar to the modern definition:

"Increased venous filling of the heart... leads to a rise of venous pressure and to acceleration and dilatation of the heart; the arterial pressure rises slightly or remains steady.  The quickening of the rate of the heart is reflex in origin since it no longer occurs after division of the vagi and cardiac accelerator nerves..."

In short, giving volume increases the heart rate. This makes some sort of logical sense; as you increase the rate of flow into the ventricles, the rate of flow out of the ventricles should also increase, and there's really only two ways this can happen (increase the stroke volume or increase the heart rate). 

However, the attentive reader will hasten to point out that this is in fact exactly the opposite of what we see in the setting of (for example) hypovolaemic shock. These patients are tachycardic, you give them fluid, and their tachycardia improves, i.e. giving volume is conventionally expected to decrease the heart rate. This is of course correct. The tachycardia in this case is mediated by the baroreceptor reflex, which seems to completely override the Bainbridge reflex. This interaction of two opposing reflexes has produced a situation where many early authors disagreed with Bainbridge, as they themselves were either unable to reproduce his findings, or their experiments generated completely opposite findings. 

In short, the ability to experimentally demonstrate the Bainbridge reflex largely depends on what the baroreceptor activity and heart rate of the subject has been prior to the fluid bolus.  From this, we come to the conclusion that the Bainbridge reflex is somehow "weaker" than the baroreceptor reflex, and the latter tends to overrule the former in the autonomic forum of the central nervous system. This is actually something one can demonstrate mathematically, by comparing species. Boettcher et al (1982) looked at the Bainbridge responses in dogs, baboons and humans, demonstrating something of a "phylogenetic progression". The investigators gave a generous fluid challenge to representatives of each species, increasing their LV diastolic pressure by 30%. In response to this, dogs had a brisk reflex and increased their heart rate by 106%, whereas baboons only incremented by 38%, and human volunteers by 21%. From this, Boetcher et al concluded that the Bainbridge reflex must have been dampened over an evolutionary time scale, becoming less important than the need to adjust cardiovascular parameters in response to posture. In other words, we humans owe the physiological prominence of the baroreceptor reflex to our insistence on standing and walking bipedally. 

Logically, the presence of the Bainbridge reflex (which increases heart rate in response to volume loading) should be matched by the presence of a "reverse Bainbridge" reflex which decreases the heart rate in response to decreased preload. This can be observed during spinal anaesthesia, where blood pressure cardiac output and venous return are markedly decreased. In these scenarios, the patient can also become profoundly bradycardic. Sure, one might say that this happens because the sympathetic supply to the heart is interrupted - but that would make absolutely no sense, considering that one usually does not block the cervical sympathetic ganglia when anaesthetising a woman for a caesarian. But you still see this bradycardia with lumbar spinal anaesthesia. In any case, the sympathetic nervous system is a minor player in the control of heart rate - when O'Rourke and Greene (1970) produced a C8 level spinal block in their normovolaemic volunteers, their heart rate decreased by only 10% (from 100 to 90). Ergo, the bradycardia seen in the setting of routine spinal anaesthesia must be attributed to an excess of vagal stimulus, produced by the reverse Bainbridge reflex (Carpenter et al, 1992). From this, it follows that the bradycardia should be exacerbated by blood loss, or changes in posture, which is in fact exactly what is seen experimentally. When Bonica et al (1972) were investigating the effects of simulated blood loss on epidural anaesthesia in healthy volunteers, the results can only be described as "Stuff You Could Only Get Away With In The Seventies". The epidural lignocaine group developed "severe cardiovascular depression, necessitating immediate treatment and termination of the study, in five of seven subjects... with brief periods (6-12 sec) of vagal arrest in two subjects".

So, from all this, the conclusion follows that under most circumstances, in hypervolemia the Bainbridge reflex is dominant, whereas in hypovolemia the baroreceptor reflex takes precedence. It is extremely unlikely that anybody will ever be asked about this balance of reflexes in any sort of exam, but if they are, it will almost certainly require the discussion of this famous diagram from Vatner & Boettcher (1978). Here, it is represented with minor modifications for clarity (for some reason, clarity required it to be stretched by about 200%). The investigators tortured some dogs with flow and pressure catheters, abruptly changing their circulating volume and observing the effect:

Anyway, to restore some much-needed exam focus, the following description of the Bainbridge reflex can be offered to the exam candidate who just needs something quick to cut and paste into their notes:

• Stimulus:  Pressure (stretch)
• Sensors: Stretch-sensitive mechanoreceptors in the atria and pulmonary arteries (Type B receptors)
• Afferent nerves: Vagus carries afferent fibres 
• Processor: Nucleus of the solitary tract and the caudal ventral medulla
• Efferent nerves: 
  • Sympathetic fibres to the heart
  • Vagal efferents to the cardiac ganglion (heart rate)
• Effector:  SA node and (maybe) AV node
• Effects:
  • Increased RA pressure produces an increased heart rate;
    unless the patient is hypovolaemic and tachycardic (in which case baroreceptor responses take precedence).

How sensitive is this reflex? For the dogs in experiments by Henry & Pearce (1956) a 100ml bolus of 6% bovine albumin was enough to change the left atrial receptor firing rate, but 50ml was not (and similar volumes needed to be extracted from the circulation to elicit the opposite response). 

The mechanoreceptors for the Bainbridge reflex are thought to be scattered somewhere in the atria, proximal central veins and pulmonary arteries (Nonidez, 1937).  Ballin & Katz (1942) applied distending stimuli to all of these regions individually and demonstrated that they do indeed participate in the Bainbridge reflex. These receptors ("atrial Type B receptors") are low-pressure mechanoreceptors which respond to stretch during diastole, in contrast to the Type A receptors which are activated by the changes in atrial wall tension during systole (Paintal, 1953).

The actual reflex arc here is less well defined than the baroreceptor reflex arc. Afferent fibres carrying signals from the stretch receptors travel up the vagus nerve, to terminate (presumably) in the NTS, and then efferents (also vagal) must descend to the cardiac ganglion to innervate the sinoatrial and AV nodes. But this is speculation. The exact mechanisms of this reflex, including central integration and control, do not seem particularly well-established. Detailed descriptions of the reflex pathway cannot be found in either Hakumäki (1987), in his tired-sounding "Seventy Years of the Bainbridge reflex", nor Crystal & Salem (2012) in their otherwise superb review.  The effector of this reflex appears to be solely the SA node; though the sympathetic nervous system is thought to have some input into the control of heart rate here, peripheral vascular resistance is not altered by the Bainbridge reflex.

Chemoreceptor reflex

Though regional hypoxia is thought to be one of the metabolic regulatory stimuli for local vasodilation, systemic hypoxia causes systemic peripheral vasoconstriction. Observe, this heavily modified graph from Pelletier & Shepherd (1972).

From the reference to "hind limb" pressure measurements, one may deduce that this was an animal study. Pelletier and Shepherd perfused the isolated carotid bodies of anaesthetised dogs with increasingly more and more hypoxic blood, recording systemic blood pressure along the way. At a dismal PaO2 of 35 mmHg, the systolic blood pressure spiked up to almost 300 mmHg, driven mainly by peripheral vasoconstriction. This is also observed in humans, and is thought to be one of the mechanisms contributing to hypertension and LV hypertrophy in chronic obstructive sleep apnoea (Cooper et al, 2005). 

Now, the chemoreceptor reflex causes an excess of sympathetic activity which is something you would conventionally expect to increase the heart rate. However, what is observed in practice is a profound bradycardia and AV block. Many of the readers will recall observing this phenomenon during the shared trauma of unsuccessful intubation. Korner (1965), at a PaO₂ of 26 mmHg, found that the heart rate and the cardiac output both decrease, a phenomenon which disappeared with vagotomy. Textbooks also usually tend to mention a hideous experiment by Berk & Levy (1977) performed on a quadriplegic patient "3 days after his cervical spinal cord was transected by a bullet", functionally at the level of C3-4. When this guy was disconnected from the ventilator "for required suctioning of his trachea, but which was not performed", he became moderately hypoxic (PaO₂ around 45 mmHg) and his heart rate dropped from 65 to 25. Then, they gave him atropine to block the vagus nerve, and demonstrated that hypoxia no longer produced bradycardia. Then, they disconnected him from the ventilator and continued to insufflate tracheal oxygen, to demonstrate that hypoxia was the most important stimulus, rather than hypercapnia or lung stretch. There is no mention anywhere in the paper of the patient or his family consenting to any of this, but this was the 70s and he was black. 

Anyway.  In the same way as all the other reflexes, this one can be broken down into easily remembered constituent parts:

• Stimulus:  low PaO₂ and/or high PaCO₂
• Sensors: carotid body glomus and aortic body glomus
• Afferent nerves: 
  • Glossopharyngeal nerve carries afferent fibres from the carotid sinus
  • Vagus (aortic nerve) carries afferent fibres from the aortic glomus
• Processor: Nucleus of the solitary tract and nucleus ambiguus
• Efferent nerves: 
  • Sympathetic fibres to the heart and peripheral smooth muscle
  • Vagal efferents to the cardiac ganglion (heart rate)
• Effector:  SA node, AV node, peripheral vascular smooth muscle
• Effects: 
  • Primary effects:
    • Vagal effects: bradycardia
    • Sympathetic effects: hypertension
  • Secondary effects: 
    • Increased preload due to increased ventilation, and thus activation of the Bainbridge reflex, which increases heart rate
    • Activation of pulmonary stretch receptors, and thus activation of the Hering-Breuer reflex, which increases heart rate

You'd be tempted to add tachycardia to the list of primary sympathetic effects, but in reality it is usually absent, as the chemoreflex and the baroreflex interact by restraining each other's responses. Chemoreflex-mediated tachycardia, which would normally be the effect of sympathetic stimulation, is dampened by the vagal effect of chemoreflex and baroreflex activation. Where rats had vagotomies, the sympathetic efferent effects of hypoxia produced a 30% increase in the atrial rate (Hashimoto et al, 1964), which demonstrates what the chemoreflex would do if it ever got off the leash.  Conversely, where the baroreflex is activated, the sympathetic response to hypoxia is diminished, or sometimes completely abolished (Somers et al, 1991); i.e. people whose baroreflex was triggered by the administration of phenylephrine did not develop more hypertension when they were also forced to inhale a low-oxygen gas mixture. 

Moreover, there are some secondary effects which need to be mentioned, and which also counteract some of the primary cardiovascular effects of chemoreceptor activation. Consider: the cardiovascular system does not signal in a vacuum, when the oxygen is low. There are also central respiratory control centres to think of, and they are far from quiescent in that scenario. With even moderate hypoxia, the respiratory drive is markedly increased, which results in increased lung inflation (larger tidal volumes). This produces a more negative pleural pressure during inspiration, which in turn increases the venous return, produces atrial stretch, and triggers the Bainbridge reflex, producing an increase in heart rate. The same thing happens due to the activation of the Hering-Breuer reflex, which occurs when the lung is stretched. The net rate-increasing effect of these secondary reflexes tends to counteract the vagal influence of the primary chemoreceptor reflex, keeping the heart rate relatively stable. 

All these complex interactions make the net cardiovascular effect of the chemoreceptor reflex difficult to describe. Even the offical CICM exam textbook (Pappano & Weir) report that "The primary reflex effect of arterial chemoreceptor excitation is to stimulate the medullary vagal center and thereby to decrease heart rate", and then some pages further along in the same chapter "Moderate degrees of systemic hypoxia characteristically increase heart rate, cardiac output, and myocardial contractility". It is unlikely that any critical care trainee anywhere will ever be asked to describe this reflex in any great detail, but if they are, one sentence regarding the multiple contradictory influences on this reflex would probably have some value. It might sound something like: 

"The chemoreceptor reflex is also affected by negative feedback loops via the Bainbridge and Hering-Breuer reflexes, which counteract the primary vagal effects to maintain a stable heart rate"

Cushing reflex

This famous neurosurgical reflex is named after Harvey Cushing only because he never gave any credit to any of the predecessors who discovered the same reflex decades earlier. Even when it was pointed out to him, he mentioned them only once, without distinction, and only “from a bibliographical standpoint”, misspelling most of their names. Though it feels dirty reward him for this by the continuing use of the eponym, the international scientific community is for some reason reluctant to abandon it, eschewing the more descriptive term "intracranial baroreflex".

This is powerful sympathetic response which occurs in response to increased intracranial pressure, and consists of the following triad:

• Hypertension
• Bradycardia
• Irregular respiration

From the nature of this response, people (Cushing included) could be forgiven for assuming that this is some sort of natural reflexive attempt by the compressed brain to mitigate the drop in its perfusion pressure. This has been debated. Evolutionary biologists point out that, in order to arise through natural selection, this reflex would need to produce viable breeding survivors following intracranial catastrophe. Occasional trepanation notwithstanding, palaeolithic neurocritical care services would probably not produce enough of those survivors to account for the presence of this reflex among all humans, and moreover it is present even in fish. The "reflex", these people argue, is really just an effect of the vasomotor centre being electrically active because it is being depolarised by ischaemia or mechanical distortion, effectively the same phenomenon as the burst of VT caused by somebody poking the ventricle. in 1960, Gracchenkov et al wrote:

"...By itself the ischemia of the medulla oblongata is not a physiological regulator but a pathological factor disturbing the activities of the nerve cells of the vasomotor center, leading to the destruction of the nerve cells. The slight and brief rise in the general blood pressure seen... is an agonal premortal phenomenon"

In response, others have found some support for the neuroprotective function of this reflex. For example, Schmidt et al (2005) were able to demonstrate an "early" intracranial baroreflex, which clearly did not arise as the result of any disastrous brainstem distortion. They funnelled normal saline into the ventricles of patients with hydrocephalus, raising their intracranial pressure and lowering their CPP. This resulted in a reproducible increase in arterial blood pressure, which was exactly enough to compensate for the increase in ICP, with the CPP maintained thereby:

So, that looks very much like a reflex designed to maintain cerebral perfusion by increasing the mean arterial pressure, and it appears to operate outside of the "catastrophic" ICP zone typically associated with unsurvivable evolution-proof head injuries. But how does it work?

Probably the most lucid and readable exploration of that question was published by CJ Dickinson (1990), and what follows owes its completeness to his extensive bibliography, filled though it is with self-references ("I comprehensively listed in my monograph [9] the previous...", etc etc).  The article, though now over thirty years old, illuminates the lack of progress since the early 20th century, as many questions about this reflex remain unanswered (or, as Dickinson had put it, "in our understanding of circulatory control there is a yawning gap").  For example, there is no acknowledged "intracranial baroreceptor" site. We simply have no idea how intracranial pressure is transduced into autonomic activity. It clearly somehow is, as multiple studies have been able to establish a connection between sympathetic output and ICP. Schmidt et al (2018), for one example, demonstrated (in mice and men) that a rising ICP produces a predictable and proportionate increase in efferent sympathetic activity (which is detectable even before there is any rise in arterial blood pressure).

How that happens remains a mystery, and it may not even be a baroreceptor response in the conventional sense, as it is seen in cerebral ischaemia even in the absence of an ICP elevation (eg. hypertension which follows ischaemic stroke). The college examiners even mention this in their answer to Question 17 from the second paper of 2010 (for them the Cushing reflex is a "result of cerebral medullary vasomotor centre ischemia")  Qureshi et al (2008) reviewed the evidence and concluded that some "direct injury to inhibitory or modulatory brain regions" was responsible for stroke-induced hypertension, but could not commit to any specific interpretation of the largely contradictory data. On the other hand, there must surely be some pressure-sensitive component, because the reflex acts incredibly rapidly. When Rodbard & Stone (1955) flushed a lethal intracranial dose of saline into the brains of rabbits, their arterial pressure response could not have been quicker, making it unlikely that ischaemia alone was responsible (it would not have had time to develop):

One strong candidate for a mechanosensitive "intracranial barosensor" is the rostroventral and lateral medulla.  It is implicated in the literature purely because poking yourself in the medulla tends to produce a strong hypertensive response. To put it into more scientific-sounding words, Jannetta et al (1985) inflated balloons at the cerebellopontine angle of baboons, just to the left to the nerve roots of the nith and tenth cranial nerves, and observed that the animals became very haemodynamically labile as the consequence. Something similar happened when Dampney et al (1982) tortured rabbit medullae with electricity. Thompson & Malina (1959) literally poked the floor of the fourth ventricle with a little metal stick, finding that the threshold of pressure to elicit a sympathetic efferent response was only about 8-20 mmHg.  Clearly this area has some potent vasomotor function, these authors concluded, and it responds to physical pressure, so... connect the dots.

Efferent output is clearly via sympathetic ganglia, and severing these connections tends to abolish the reflex entirely. The central processing centre is not clearly established, but all those rostral medullary neurons send projections to the sympathetic neurons of the spinal cord, as well as to the nucleus ambiguus. McBryde et al (2017) make a great effort to connect these dots for us, and largely fails, because over the last 100 years we have not really achieved very much in our attempts to characterise this reflex arc. But, of one had to pool all this speculation into a short summary, it would look a bit like this:

• Stimulus:  Intracranial pressure or cerebral ischaemia
• Sensors: god knows what; mechanosensors in the rostral medulla?
• Afferent nerves: 
  • Fibres from the medullary mechanosensory areas, to sympathetic ganglia
  • Fibres from cerebral hemispheres, which exert descending inhibitory control on the medullary vasomotor sensor
• Processor: Rostral ventrolateral medulla
• Efferent nerves: 
  • Sympathetic fibres to the heart and peripheral smooth muscle
• Effector:  SA node, AV node, peripheral vascular smooth muscle
• Effects: 
  • Primary effects: hypertension and tachycardia
  • Secondary effects: baroreflex-mediated bradycardia

Bezold-Jarisch reflex

This reflex is always mentioned in the college answers to SAQs, which means its exam importance is actually greater than its actual physiological importance.  In short, this reflex produces bradycardia vasodilation and apnoea. Warltier et al (2003) have a good review of this "pharmacological curiosity". It was discovered when von Bezold & Hirt (1867) injected some kind of toxic plant alkaloid into the bloodstream of a cat. Toxins, it turns out, as well as various mechanical stressors, are sensed by the walls of the cardiac chambers in the same way as pain is sensed by the skin, through unmyelinated C-fibres. The afferents travel up the vagus nerve to synapse somewhere in the medulla, presumably in the nucleus of the solitary tract, exerting a cardiodepressant effect (Lee et al, 1972). Efferent vagal and sympathetic responses both vasodilate the peripheral circulation and slow the heart rate, producing hypotension and decreased cardiac output. Here, an original dog recording from Zucker & Cornish (1981) demonstrates the magnitude and timeframe of this reflex, stimulated by a direct intracoronary injection of veratrum alkaloids:

Now, at this stage the enraged reader will surely be ready to point out that there is no endogenous source of veratrum alkaloids or capsaicin, so what could possibly be the normal role for this reflex?

According to Warltier et al (2003), it remains a mystery. Various possible physiological regulatory process and pathological phenomena have been associated with this reflex, and investigators have variously claimed that it participates in the response to cardiac ischaemia, vasovagal syncope, the response to shock, and the regulation of arterial blood pressure. The latter probably represents some sort of genuine homeostatic role for this reflex, as many of those C-fibres are mechanosensory and respond to stretch stimulus. From a clinically relevant ICU-centric point of view, this can be discussed in terms of the response of the cardiovascular system to haemorrhage. It appears that in haemorrhagic shock (where the cardiac chamber stretch is decreased) the vasoconstriction is greater than it would be if the shock was purely due to decreased cardiac output (eg. in cardiogenic shock). This added vasoconstrictor stimulus is attributed to the inhibition of the Bezold-Jarisch reflex. In short:

• Decreased circulating volume decreases the ventricular/atrial stretch 
• This loss of stretch decreases the firing rate of the neurons of the C-fibre mediated vagal afferent limb of the Bezold-Jarisch reflex
• This removes the inhibitory effect of the Bezold-Jarisch reflex on the heart rate and the sympathetic tone
• At the same time, activation of the baroreceptor reflex stimulates tachycardia and vasocostriction
• The additive effect of the two reflexes is therefore a greater vasoconstriction and a greater increase in heart rate

There are several other relevent scenarios where one might come across this reflex, brought up by Mike Cadogan in the associated eponymology entry, include the bradycardia and AV block seen during and inferior MI, or the reperfusion thereof. Apparently it is though to produce some sort of cardioprotective effect, as the reflex vasodilation of the coronaries and the bradycardia presumably decreases myocardial oxygen consumption and secures coronary blood flow. How the rest of the organism benefits from the resulting cardiogenic shock is not clear. Additionally, we should be grateful to the reader who has pointed out the role of this reflex in the bradycardia and hypotension associated with spinal anaesthesia. With the sympathetic nervous system chemically severed, venous blood pools in the lower extremities and splanchnic circulation, emptying the left ventricle and activating the reflex (Campagna & Carter, 2003). Even more curiously, this reflex bradycardia can often be abolished by the pre-spinal administration of ondansetron (Sharma & Maravi, 2020), suggesting that the pathways involved are serotonergic, and even more specifically mediated by 5HT3 neurotransmission.

To summarise this reflex further:

• Stimulus:  multiple and heterogeneous stimuli, including:
  • Mechanical: pressure and stretch (thus, inotropy preload and afterload)
  • Chemical: veratrum alkaloids, ATP, capsaicin, snake venom, and various other venoms from the animal kingdom
• Sensors: Heterogeneous sensors distributed in all cardiac chambers
• Afferent nerves: 
  • Unmyelinated C-fibres of the vagus
• Processor: Nucleus of the solitary tract, likely involving serotonergic transmission
• Efferent nerves: 
  • Sympathetic fibres to the heart and peripheral smooth muscle
  • Vagus nerve, via the cardiac ganglion
• Effector:  SA node, AV node, peripheral vascular smooth muscle
• Effects: hypotension (vasodilation) and bradycardia

At this stage, one might bring up one other interesting point. Observe: this reflex reacts to chamber stretch by slowing the heart rate. An astute observer might point out that this is the exact opposite of what the Bainbridge reflex does. How do we reconcile the contradictory efferent outputs of reflexes which react to the same stimulus?  Well, it appears we don't have to. They don't completely cancel each other out, but one acts as the physiological restraint of the other. Specifically, it appears the Bainbridge reflex dominates, and the Bezold-Jarisch reflex acts as the moderator of the increased heart rate.  Hajdu et al (1991) created a dog model where the Bainbridge reflex could be stimulated by inflating a balloon in the atrium, at the same time as the Bezold-Jarisch reflex was stimulated by infusion of veratrum alkaloids. When both reflexes were simultaneously stimulated, a tachycardia was produced, but the rate was considerably lower than what is seen with a Bainbridge reflex on its own.  

Oculocardiac reflex

This is the inconvenient reflex which occasionally produces fatal cardiac arrest during routine eye surgery. It is fortunately not usually known by its eponym (Aschner–Dagnini reflex), but is still probably inaccurately named - "trigeminocardiac" or "trigeminovagal" would probably be more accurate as it can arise from stimulation anywhere in the trigeminal nerve territory, and it appears to be a vagal phenomenon, which is almost completely abolished by atropine.  The much better name for it, coined in North America on the basis of North American coins, is the "Five & Dime reflex", which is a hundred times more useful than "oculocardiac" or "Aschner–Dagnini" because it clearly spells out the cranial nerves involved. Most readers of the modern era will not recognise the other meaning of the name, which refers to either a popular chain of thrift stores from early 20th century, or an eight-minute cartoon from 1933 about the scandalous behaviour of an urban rabbit.

In summary, this reflex is as follows:

• Stimulus:  pressure to the globe of the eye, or traction on the eye muscles
• Sensors: mechanosensitive stretch receptors in the facial muscles, especially periorbital muscles, and in the globe of the eye
• Afferent nerves: 
  • Long and short ciliary nerves, to the trigeminal nerve, via the Gasserian ganglion, to the sensory nucleus of the trigeminal nerve, and from there via short internuclear fibres to the NTS.
• Processor: Nucleus of the solitary tract
• Efferent nerves: 
  • Vagus nerve, via the cardiac ganglion
• Effector:  SA node, AV node
• Effects:
  • Vagal: bradycardia
  • Sympathetic: systemic vasoconstriction, cerebral vasodilation

So, the globe of the eye and retroorbital tissues have stretch receptors, and stimulating them can produce a vagal stimulus so potent that the heart may stop, to the point of death. Nothing could illustrate this phenomenon better than the original recordings from Aserinsky & DeBias (1963), who poked the eyes of twenty anaesthetised dogs:

That solid "eyes pressed" bar is forty seconds long. During this time, the dog experienced twenty-eight seconds of asystolic arrest. When contemplating this, a reasonable person will find it difficult to suppress the thought that it cannot possibly have any functional purpose. What possible use could there be for this phenomenon, a reflex that threatens to kill you?  Well. It appears that these pressure and pain stimuli are cross-talk between sensory fibres, and the reflex appears to be intended to sense immersion in water.  Which brings us to...

The diving reflex

Schaller et al (2009) hypothesise that the oculocardiac reflex must be grouped with the diving reflex as an "oxygen-conserving reflex", implying that they confer some sort of survival benefit in the setting of asphyxia. Both reflexes, upon stimulation of trigeminal sensory nerve afferents, produce a cardiovascular reaction which conserves oxygen and delivers it to the brain. At the same time, apnoea develops (see the "resp" curve go flat in the abovementioned recording from Aserinsky & DeBias), to prevent you gulping water, and cerebral vessels dilate to promoting flow of residual oxygenated blood to the brain. Overall, it sounds like a useful reaction to suddenly being submerged. This is the reflex you rely on to slow the heart rate when you flub a wet towel over the face of a person in SVT (it's one of those vagal manoeuvres), the wet towel presumably being viewed as somewhat more polite than sustained ocular pressure. 

This "diving reflex" is much better developed in infants, aquatic birds and diving mammals such as seals, but is actually present in one form or another in all vertebrates. Panneton (2013) reviews the comparative biology from the perspective of a veteran laboratorian ("...in our hands, the response is seen in 100% of rats, 100% of the time"). Some vertebrates are better equipped to manage this than others. For example, Panneton discusses pinnipeds whose "elastic and bulbous" descending aorta acts as an excellent windkessel reservoir to maintain steady flow during those log diastolic intervals between slow heartbeats. The data suggest that this reflex is not one single neural arc, but rather the combined effect of three separate processor nuclei and efferent processes acting in unison to respond to the same afferent stimulus.

To summarise:

• Stimulus:  trigeminal nerve sensory distribution
  • Pressure to the globe of the eye, or traction on the eye muscles
  • Pain in the trigeminal nerve distribution
  • Temperature (cold) 
  • Chemical stimulus of the anterior ethmoidal nerve (noxious)
• Sensors: Pain, temperature, chemical and mechanosensitive stretch receptors in the trigeminal nerve distribution
• Afferent nerves: 
  • Branches of thee trigeminal nerve, via the Gasserian ganglion, to the sensory nucleus of the trigeminal nerve, and from there via short internuclear fibres to the NTS.
• Processor: 
  • Nucleus of the solitary tract: vagal response
  • Rostral medulla: sympathetic response
  • Ventral medulla: apnoea
• Efferent nerves: 
  • Vagus nerve, via the cardiac ganglion
  • Phrenic nerve
• Effector:  SA node, AV node, respiratory muscles
• Effects:
  • Vagal: bradycardia
  • Sympathetic: cerebral vasodilation, systemic vasoconstriction
  • Respiratory: apnoea
  • The net effect is to prevent aspiration and to maximise the blood flow to the central nervous system at the expense of the skin, muscle and splanchnic organs.

Barcroft-Edholm reflex and vasovagal syncope

Speaking of counterproductive-sounding reflexes, this one is spectacular. Henry Barcroft and O.G Edholm discovered it in 1944 and became eponyms; for some reason the other co-authors of their paper (John McMichael and E.P. Sharpey-Schafer) got nothing. In fact, the same phenomenon was earlier described by Ebert Stead & Gibson (1941), and they also never get any credit.

Anyway. The year was 1944 and the British public were donating blood in vast amounts to sustain the war effort, draining captured German soldiers having been rejected as a distasteful option.

Approximately 550ml of whole blood was donated at a time. It became clear very rapidly that some proportion of these donors became faint, and though the overall numbers were not great (~ 3%), because the total numbers were in the tens of thousands, this equated to entire piles of collapsed donors. Moreover, the more blood you removed, the more likely the collapse. In order to explore this phenomenon in more detail, Barcroft and Edhold exsanguinated healthy members of the Friends Ambulance Unit and the Belfast Medical Students Association. By rapidly withdrawing about 1000ml of blood, they caused these resting donors to experience an abrupt vasovagal collapse, characterised by a simultaneous fall of peripheral vascular resistance and precipitous drop in heart rate:

As one can clearly see, at the point where about 1L of blood was lost over 10 minutes or so, the patient's heart rate and peripheral resistance abruptly dropped, with the subject rendered unconscious by the resulting fall in perfusion. As Ebert Stead & Gibson (1941) described it,

"With the onset of collapse both the systolic and the diastolic pressure fell precipitously, and the heart rate became strikingly slow, ranging between 36 and 40 beats per minute. ...An ashen gray pallor developed, and they retched, perspired profusely and responded slowly, if at all, to commands. One subject became unconscious."

Given the combination of vascular and vagal effects, the term "vasovagal" was used to describe this syncope. What is the mechanism here? It turns out, the response to haemorrhage and the response to sudden catastrophic news are mediated by the same mechanisms, well detailed in Jardine et al (2018).  Emotional, orthostatic and hypovolemic syncope are all variations on the same theme. To quote directly from  Alboni et al (2008), "very little is known about the afferent part of the vasovagal reflex (i.e., the step from trigger to autonomic control and central processing)" but like everything else in this chapter, the efferents are obviously the vagus nerve and the sympathetic nerve.

To bring things back to the first paragraph of this section, one has to wonder: what could possibly be the point of a reflex which, under duress, knocks the legs out from under you just as your need for cardiovascular mojo is maximal? Others have also wondered. Representative papers are crowned by an excellent review from Alboni et al (2008) and an editorial by van Dijk & Sheldon from the same issue of Clinical Autonomic Research. In summary, two main theories are favoured: the conflict hypothesis and the clotting hypothesis. The clotting hypothesis suggests that mammals have developed this protective reflex to slow the circulation in response to haemorrhage, practising permissive hypotension in response to trauma in order to minimise blood loss and maximise the possibility of clot formation. The conflict hypothesis suggests that in the palaeolithic our common ancestors were violent arseholes and "loss of consciousness triggered by fear-circuitry activation might have conferred a survival advantage on non-combatants" such as women and children. Furthermore, other vertebrates also have similar reflexes, for example the phenomenon of "alarm bradycardia" which integrates into the tonic immobility responses of prey animals. In short, from observing these flailing threads of debate, one might surmise, like van Dijk & Sheldon, that "there are perhaps already too many theories for the number of facts" in this area. 

Unfortunately, this factless reflex with all the uncertainties and blanks in its arc was the subject of Question 19 from the second paper of 2017, where the  college asked us to "describe the physiology of a vasovagal syncope". Somewhat unfairly, the college examiners insisted that "the mechanism should have been described including the various receptors involved". So, If one made an attempt to summarise this reflex, one might end up with this sort of thing:

• Description:
  • ​​V​​​​asovagal syncope, or neurocardiogenic syncope, is a transient loss of consciousness due to global cerebral hypoperfusion, which occurs as the result of an autonomic reflex response to various stimuli
• Stimulus:  
  • For "true" vasovagal syncope:
    • Emotional distress
    • Orthostatic changes (decreased preload with changes in posture)
  • For "situational" neurocardigenic syncope:
    • Micturition
    • Increased intrathoracic pressure:
      • Defaecation
      • Cough, sneeze, laughter, playing a brass instrument 
      • Following exercise
    • "Carotid sinus syndrome"
• Sensors: Central (descending) as well as peripheral
  • Mechanoreceptors located in the wall of the left ventricle, the aorta, atria and the pulmonary trunk
  • Numerous other strecth receptors, eg. splanchnic, bowel, 
• Afferent nerves: Unknown! Presumably, both central nervous system and peripheral sensory nerves are involved
• Processor: Unknown! Presumably at some stage the nucleus of the solitary tract and the nucleus ambiguus are involved.
• Efferent nerves: 
  • Vagus nerve, via the cardiac ganglion
  • Sympathetic nervous system
• Effector:  SA node, AV node, peripheral smooth muscle
• Effects:
  • Vagal: bradycardia
  • Sympathetic: systemic vasodilation (mainly muscles)
  • Vasovagal syncope is thought to have four distint phases:
    • phase 1: early stabilization (by normal baroreceptor reflex)
    • phase 2: circulatory instability (baroreflex vasoconstriction)
    • phase 3: terminal hypotension (bradycardia, cerebral hypoperfusion, systemic vasodilation)
    • phase 4: recovery

Interestingly, though the official CICM answer draws a distinction between vasovagal syncope and orthostatic hypotension ("which is a failure of the autonomic reflex response"), orthostatic causes are listed in the 2018 ESC guidelines for the management of syncope. In fact they separated vasovagal syncope into "orthostatic" and "emotional", alongside several other forms of reflexive neurally mediated syncopes. 

Respiratory sinus arrhythmia

Sinus arrhythmia is the name given to the totally normal variation in heart rate which occurs cyclically in response to normal respiration. Specifically, it increases during inspiration. One might assume that there might be some sort of crudely mechanical phenomenon behind this, as the lung bellows interact with the cardiac pump. On spontaneous inspiration,  the negative intrathoracic pressure delivers an increased blood volume to the right heart, and so (surely by means of the Bainbridge reflex) the heart rate increases slightly to compensate. However, this is probably not the case.

In fact, though various homeostatic regulatory influences must surely play some minor role, the primary reason for this heart rate fluctuation is an interaction between the medullary respiratory and cardiac centres. As one can clearly see from this tracing (Hayano et al, 1996), the heart rate slows immediately as the breath starts, before any of the other reflexes have had time to react: 

More interestingly, Pappano & Weir quote an even more illustrative study (), where dogs on cardiopulmonary bypass still had a sinus arrhythmia in spite of the fact that they had collapsed motionless lungs. In other words, this is a coordinated reflex, driven by the interaction of central control organs, and it does not rely on pressure data from the lungs. Bernston et al (1993) suggest that its main goal is to increase the efficiency of the pulmonary circulation during inspiration, by increasing the rate of blood flow across the exchange surface. Hayano et al (1996) confirmed this by reversing the normal respiratory sinus arrhythmia pattern (by atrially pacing their dogs), thereby demonstrating a deterioration of pulmonary gas exchange efficiency. To complete this discussion with a tedious formula, this reflex can be summarised as follows:

• Stimulus: presumably, the Pre-Bötzinger complex ("respiratory pacemaker")
• Sensors: none (unless you count respiratory control chemoreceptors)
• Afferent nerves: interneurons between Pre-Bötzinger complex and nucleus ambiguus
• Processor: nucleus ambiguus? 
• Efferent nerves: Vagus nerve, via the cardiac ganglion
• Effector:  SA node, AV node
• Effects: cyclical decrease of vagal output during inspiration