Excitatory, conductive and contractile elements of the heart

This chapter is relevant to Section G1(i)  of the 2017 CICM Primary Syllabus, which asks the exam candidate to "describe the structure and functional significance of the excitatory, conductive and contractile elements of the heart".  Though that might seem like a straightforward request, in fact if you look closely you will notice that this topic is impossibly broad, and refers to most of what the heart is, or does. It is of course impossible to guess precisely what the syllabus-writers meant by this cryptic entry, but if one had to decipher it with education and assessment in mind (eg. "how would I have phrased this if I wanted to teach junior intensive care doctors"), one might scry the following meaning:

• Excitatory tissues are the SA and AV nodes, as well as random ectopic pacemakers. 
• Conductive tissues are the Bundle of His, Bachmann's Bundle, the left and right bundle branches,  the Purkinje fibres and cardiac myocytes themselves
• Contractile tissues are the atrial and ventricular wall myocytes
• For these tissues, the most important aspects of their "structure" are their anatomical locations and their "significance" is how they integrate together to coordinate the contractile function of the heart.

Reading over this, it sounds almost like they wanted a description of the propagation of the cardiac action potential. The discussion of this topic will therefore focus on describing this in a structure which follows its normal progression. Along the way, one encounters "contractile elements of the heart", which is a convenient way to describe them in a bit more detail anatomically, considering that there is no opportunity to do so elsewhere.

In summary:

SA node

The sinoatrial node is a tiny bundle of pacemaker cells which lives in the superior right atrium, roughly at the junction of the crista terminalis in the upper wall of the right atrium and the opening of the superior vena cava. This excellent diagram from Choudhury et al (2015) has been lightly modified to decrease its clarity:

\

This bundle of cells receives innervation from the cardiac plexus fibres, and is affected weirdly by their influence. A strange phenomenon known as pacemaker shift occurs. Basically, the SA node is a group of cells, each depolarising with slightly different timing to each other, and so each time the SA node fires, it is often not the same cell as last time. Then, as vagal or sympathetic stimuli are applied, different cell groups seem to start becoming the "dominant" pacing centre. This is attributed to different sensitivities of these cells to acetylcholine; it slows central pacemaker rate more than peripheral ones, causing the caudal cells that are less acetylcholine sensitive to assume control during periods of vagal stimulus.

Anyway. The CICM first part exam will probably expect the trainees to recall that the "native" rate of the SA node is said to be 60-100 beats per minute. In fact, that's the sinus nodal rate with the breaks on, as there is some resting vagal stimulus always applied. When it is denervated (eg. in cardiac transplant), the SA node usually runs at a rate of around 100bpm (Beck et al, 1969). In fact, in the course of the normal transplant surgery, the posterior bits of the original atria are apparently left in situ (Bexton et al, 1984, suggest that it is technically too fiddly to do otherwise), which means the recipient will have two SA nodes (one only directing the activity of the posteriormost atrial remnants). This makes the cardiac transplant recipient the ideal laboratory to investigate neural and humoural effects on the SA node.

Anyway. The SA node leads the P-wave. From there, the action potential is conducted to the AV node via the internodal tracts.

Bachmann's Bundle and the internodal tracts

This is a narrow band of aligned atrial muscle fibres which stretches from the right auricle to the base of the left auricular appendage. It was first described by Bachmann in 1916. "This band has appropriately been called the inter-auricular band", he wrote; but subsequent authors named after him anyway. It is in fact an extension of one of three major atrial circuits, which are described by Thomas N. James (1963):

• The anterior internodal tract passes from the sinus node to sweep anterior to the superior vena cava into Bachmann's bundle, where it divides to distribute to the left atrium and to curve back into the interatrial septum and descend to the A-V node.
• The middle internodal tract leaves the dorsal and posterior margins of the sinus node and courses behind the superior vena cava through the sinus intercavarum to the crest of the interatrial septum, and there descends into the A-V node, merging with fibers from the anterior tract as it approaches the node.
• The posterior internodal tract follows the crista terminalis from the sinus node to the Eustachian ridge and thence through the ridge to the posterior margin of the A-V node.

This sounds like something that would be better explained with a picture. The original illustration from Thomas' article is excellent, even though time and Xerox were unkind to it. With some adjustment, it has been reproduced below:

It is in fact not clear why Bachman's bundle is viewed separately, as it is essentially an extension of the anterior internodal tract. In case one cannot get enough of hagiological wankery, the middle internodal tract is occasionally referred to as Wenckebach's bundle, and the posterior internodal tract was once known as Thorel's bundle, although these eponyms have fallen into disuse with the epidemic loss of respect among young punks in medicine. 

These bundles are not "insulated", i.e. there does not seem to be any sort of fibrous sheath surrounding them (this also allows them to depolarise the surrounding atrium, as the signal propagates along them.) They are just muscle fibres from the atrium, which seem to be aligned in the same direction. 

Some of these are just normal atrial myocytes, and others are fibers with Purkinje-like histological characteristics (but these are not continuous tracts of them, there are just some random Purkinje-ish myocytes scattered around). Overall, there's clearly some attempt made to specialise these fibres for conduction: for instance, Sherf & James(1979) found that they are totally useless for contraction, as much of the contractile proteins are missing from them. The result of this is a speed of conduction which much more rapid than the surrounding atrial tissue: Wagner et al (1966) were able to record a conduction velocity of 1.7 m/sec along Bachman's bundle, whereas it was 0.4 m/sec everywhere else.

Anyway. It will suffice to summarise, for the purposes of exam revision, that: 

• The SA node conducts to the AV node via three discrete bundles of specialised tissue
• The conduction speed along these bundles is higher than the surrounding tissue, and is around 1.7 m/sec.

Fortunately, as one does not possess metres of atrial myocardium to traverse, sinoatrial node depolarisation is usually conducted to the atria very rapidly.  Erdem et al (2015) measured the interatrial conduction velocity in young healthy Turks (from the upper right atrium to the mitral ring), and determined that the signal arrived within about 70 msec. Interestingly, in other species which do have metres of atrium (eg. Megaptera), the timing of electrical conduction does not scale with body size, but remains virtually unchanged (barely double that of the human), suggesting that there is some mechanically optimal conduction rate which needs to be maintained for the cardiac cycle to work. A fascinating digression is offered by this comparative diagram:

Atrial myocytes and atrial walls

The atria are sacklike receptacles for blood, and participate in the mechanical events of the cardiac cycle mainly in a passive way, by filling with blood and increasing their internal pressure. Only for about 15% of the time do they actually do anything useful (i.e squeeze blood into the ventricles), in late diastole. Their functions, specifically, are:

• Reservoir during systole
• Passive conduit during early diastole
• Booster function during late diastole (that's the active role)

This is all discussed elsewhere, in the chapter on the cardiac cycle. For the time being, let us focus on their structure, contractile properties and excitatory potential. Anderson & Cook have an excellent paper ("The structure and components of the atrial chambers", 2007) which is freely available from Oxford Academic Press. 

Structure of the atrial chambers: The atria each have a venous portion, a vestibule, a body and an appendage. Each one can be considered a simple muscular tube, with the atrial appendage hanging out one the side like some kind of hernia. These images are from the Anderson & Cook paper, and are created from resin casts, where the right sided structures were cast in blue and the left in red and for some reason also yellow.

Note the corrugated topography of the atria. The atrial walls are grooved by the presence of multiple pectinate muscles, most so in the appendages. The appendages are tube-like and quite large; for the right atrium, the appendage forms pretty much the entire anterior portion of the chamber. Though Anderson & Cook digress extensively on structures such as the Eustachian valve and the Tendon of Todaro, this seems unnecessary even for the digression-prone author, as none of the readers are ever likely to be repairing tricuspid valves, and those involved in formal study of the subject will not be using Deranged Physiology as their main resource. 

Atrial wall muscle is probably more interesting to the intensivist, as it is a structure which they can inadvertently perforate in the course of their routine procedural work. In general, the left atrium is thicker than the right, as it is usually exposed to higher pressures, but both are pretty flimsy and are clearly not designed to do battle with the rigid tip of a central line. Parts of the right atrium might be thicker (its thickest region is at the top of the terminal groove, and is 5-8mm thick), but the areas most likely to be hit by stray guidewires (posterior and anterior walls) tend to be only 2mm in thickness. For the left atrium, the thickest region is the anterior wall (4-5 mm), and the thinnest part is the posterior wall (about 3 mm). For those who may want even more details about this sort of stuff, Wang et al (1995) have some excellent discussion; but for the rest, this image from Anderson et al (2000) will suffice to describe the relative thinness of this structure:

Atrial myocytes are arranged relatively haphazardly in the atrial walls, although in some areas (eg. the internodal bundles) they are mainly arranged along the axis of conduction in the walls of the atria, which allows those areas to conduct faster. Overall, they are smaller then ventricular myocytes, usually only have one nucleus (ventricular myocytes usually have two), and are distinguishable by the presence of electron-dense granules which contain ANP, the atrial natriuretic peptide (Gerdes, 2012).

These cells are usually a calm population of organised participators, and are not given to outbursts of wild individualism. They are not, for example, naturally prone to self-depolarise spontaneously - they usually wait for a signal, and then propagate it. However, under certain conditions, they can misbehave. When depolarised, they can develop something called depolarisation-induced automaticity (Antzelevitch et al, 2011), which tends to develop at a membrane potential of around -30 to -70 mV. Factors which promote the development of this automaticity include:

• Increased extracellular potassium (i.e. hyperkalemia)
• A reduced number of I_K1 channels, which are responsible for maintaining resting membrane potential (this occurs in chronic heart failure)
• A reduced ability of the I_K1 channel to conduct potassium ions (eg. by PKA-mediated phosphorylation of the channel., and therefore an effect of drugs such as isoprenaline and milrinone)
• Electrotonic influence of neighbouring cells

When this automaticity develops, rogue atrial pacemakers tend to go off with a much faster firing rate than standard pacemakers. Fortunately, the AV node acts as the gatekeeper, hopefully only transmitting a sane number of impulses per minute.

AV node

The AV node is a little (5-10mm) fusiform bundle of cells, usually found somewhere within the Triangle of Koch at the base of the right atrium. It sits on the right side of the interatrial septum, somewhere between the coronary sinus ostium, the tendon of Todaro and the septal leaflet of the tricuspid valve. It was originally separated from the Bundle of His somewhat arbitrarily, by being on different sides of the fibrous cardiac skeleton. It ultimately turned out that the two do differ on a cellular level, with AV nodal cells being short and fusiform, arranged without any clear orientation, whereas the cells in the Bundle of His are longer and arranged in parallel (Kurian et al, 2010). The AV node, like the SA node and internodal bundles, is not insulated from the rest of the "working" myocardium, and comes into contact with normal myocytes. The image below from Miyazaki (2014) shows a section which includes both the SA node and an AV node from a human infant:

Excitable and conductive properties of the AV node include:

• its ability to delay the conduction of atrial impulses to the Bundle of His
• its post-repolarisation refractory period, which restricts the rate of conducted impulses to a maximum heart rate of about 220.
• its ability to spontaneously depolarise and lead the myocardium along in a "nodal" rhythm

The AV conduction delay is a feature, not a bug. The cells here are intentionally slow; the total time for an action potential to get from the base of the atrium to the Bundle of His appears to be 90-110 msec (Meijler & Jance, 1988), which is long enough for the atrial contraction to finish before ventricular contraction begins. Consider what would happen if it did not: the atrium would still be contracting when the ventricular contraction occurs. The ventricle, with the force of their contraction, would slam the valves shut in the atria's face, and the atria would contract pointlessly against a closed bunch of valves. That would be an inelegant way to run a circulatory system. Thus, to introduce a built-in delay, AV nodal conduction velocity needs to be extremely slow when compared to the internodal bundles and atrial muscle. Scher et al (1959) gave 0.12 m/sec as the value, which has been subsequently quoted in every textbook.

Where is this delay coming from? The AV node is functionally divided into three regions (AN region, N region and NH region), where the AN region transitions from atria to node, and the NH region merges into the Bundle of His. The  N region (the central AV node) is where the delay mainly happens: conduction slows down to the most sluggish rate of all, about 0.05m/sec. Again, this number comes from the 1959 paper by Scher et al, who measured it in the heart of an 8kg dog; it appears to have propagated far and wide, and the latest edition of the CICM-endorsed "official" cardiovascular physiology textbook also quotes this value. Hoffmann et al (1959) measured a range of conduction velocities down to as low as 0.02m/sec, which suggests that a range of possible values could be mentioned in an exam answer.

What is the mechanism of this delay? It appears to be the absence of a fast current in the depolarisation of cells in the N region. Anderson et al (1974) measured this in the rabbit heart, and represented the different action potentials in a famous diagram:

Observe: the upstroke of the action potential in the "transitional, midnodal" cells was very slow. For normal cells, a rapid inrush of extracellular sodium produces the abrupt spike of depolarisation, but in these cells it appears to be absent. The absence of this inward sodium current was elegantly demonstrated when Zipes & Mendex (1973) injected the N-zone with pufferfish tetrodotoxin, which blocks voltage-gated sodium channels. The rest of the atrium was electrically silenced by a minuscule amount of tetrodotoxin, whereas the N cells had action potentials of exactly the same amplitude and with the exact same (slow) rate of rise. 

Postpolarisation refractoriness is another important feature of AV nodal function which offers a degree of control over the maximum ventricular heart rate. In short, the AV node will not conduct an action potential if it arrives within about 400 msec of the last potential. 

The functional AV node refractory period extends beyond the actual repolarisation of the cells. Linhart et al (1965) tested this by zapping some human atria with a 90 mV pacing current. Provided there was a respectful pause following the normal QRS ("driving stimulus"), the zap provoked an AV nodal current, and was conducted to the ventricles. "The time interval between the driving and the premature stimuli was then shortened in 10 msec steps until the premature impulse failed to evoke a QRS complex", the authors continued. Eventually, the interval was shortened to a point where the zap came too soon following the QRS, and did not propagate. This minimum interval was the functional refractory period of the AV node, and ranged between 300 and 500 msec (average of 405 msec).

An R-R interval of 400 msec gives a maximum conducted heart rate of around 150 beats per minute, which is in fact what you tend to see in rapid AF and atrial flutter. Obviously, we have all seen narrow complex tachycardias which run much faster than that. Various influences can affect this refractory period, of which probably the most important is the autonomic nervous system. Linhart et al made their subjects perform exercise and challenged them with atropine, managing to reduce the refractory period down to as low as about 240 msec, which would give a maximum heart rate of around 250 beats per minute. That probably represents some sort of maximum, and is consistent with what one might see in the ED and on the wards. Without offering any references, different textbooks quote different maximum heart rates, all of which are in the 200-240 bpm ballpark. 

Logically, from this it follows that vagal stimulus should slow AV nodal conduction, and that is in fact what we see when we perform those useless vagal manoeuvres on patients with SVT. This works by hyperpolarising the N cells,  mainly through increased potassium permeability. In fact, a high enough potassium concentration can induce AV block independently of the vagus (Greenspan et al, in 1965, used potassium chloride to produce impressive AV block in the hearts of dogs, at a serum range of 6.3-9.8 mmol/L).

AV nodal pacemaker capabilities are worth mentioning. This "junctional rhythm" originates from NH cells, below the slow part of the AV node, and spreads down into the ventricles. This can be useful as a means of maintaining normal ventricular coordination in the face of some sort of nodal or atrial dysfunction (eg. if somebody cryosurgically ablates your upper AV node). The rate of a junctional rhythm in humans varies (depending on autonomic control) and has been reported in the range of 72-40 beats per minute. It responds to vagal and sympathetic stimuli, which means that theoretically, a haemodynamically unstable patient with a narrow complex bradycardia may get away with just an isoprenaline infusion. 

At the same time, the action potential is transmitted to the atria in a retrograde fashion, which gives rise to a profoundly stupid A-V mechanical incoordination, where the atrial impulse meets a closed tricuspid valve and is instead transmitted back up into the central veins. This contraction gives rise to "cannon" waves, which can be seen by expert cardiologists in necks, or by non-expert intensivists in central venous pressure waveforms, as shown below:

You can also see this in complete heart block, and wherever a ventricular depolarisation is retrograde-conducted (eg. VVI pacing), but this is all a bit of a digression. A few thousand words into this chapter, we are still only up to the AV node. Thus, without further ado:

Bundle of His and the His-Purkinje system

The eponymous bundle, discovered by Wilhelm His in 1983, is basically a continuation of the lowermost NH part of the AV node, and there does not appear to be any specific histological reason to draw a line which separates the two structures. Electron microscopy of the region demonstrates " a gradation along the conduction system from typically nodal towards myocardium like cells", with no clear boundary (Mochet et al, 1975)

The bundle is short (maybe 10mm), and separates into two branches. The right continues down the subendocardial region of the right ventricle, and the left penetrates the interventricular septum, where it divides yet further into discrete anterior and posterior fascicles.

Here, the conducting system of the lamb's heart is reproduced from the original paper by Lydia DeWitt (1909), out of respect for the original author who painstakingly constructed this wax model from multiple sections (using the Born method of plate reconstruction). This paper is very early; in the image, k stands for Knoten, the name given to the AV node by Sunao Tawara only three years earlier, when he discovered it.

Anyway. Beyond the bundle branches, the conducting system fans out into a network of progressively thinner and thinner fibres. The thing that separates this circuitry from the Purkinje fibres is the fact that the bundle of His and all its subsequent branches are covered in a layer of fibrous tissue which insulates them from the ventricle.

In contrast, Purkinje fibres are naked. Purkinje cells are large, wide cells with a clearer cytoplasm and a diameter usually at least double that of surrounding myocytes. In this image (stolen from Washington State U.), they are pretty easy to identify:

Their size made them distinct enough for Jan Purkinje (Purkyně) to discover them in 1839, though his discovery went unrecognised until he published in German several years later (Sedmera & Gourdie, 2014).

Electrophysiological properties of the His-Purkinje system

Purkinje fibre cells all seem to share the same conduction velocity and electrical behaviour characteristics with Hisian cells and bundle cells; or at least nobody has really published anything easily findable that would conclusively discriminate between their electrophysiological properties. Ergo, it makes logical sense to discuss them all together.

Functionally, these cells have characteristics which are quite different to both nodal cells and to working myocytes. From Dun & Boyden, 2015, when compared to working myocardium cells, the Purkinje fibres have the following distinguishing features:

• The action potential duration is longer (~320 msec)
• Total amplitude of the action potential is larger (~120 mV)
• Rate of rise of action potential is higher (~450 mV/sec)
• The refractory period is longer
• The refractory period decreases with high heart rates; whereas, for the AV node, the refractory period increases when the atrial rate is high. This means:
  • At slow heart rates, the His bundle is what protects the ventricle from conducted atrial extrasystoles, and
  • At fast heart rates, the AV node is responsible for this filtering.

Much of these differences seem to stem from the ion channel expression in these cells, which is markedly different to that of normal cardiac myocytes. 

Conduction velocity in the Purkinje fibres is quoted as 1-4 m/sec in the official CICM textbook. Obviously this is another number which tends to vary among the published materials. For example, Durrer et al (1970) gives 2-3 m/sec as their value, which in all honesty was not directly measured. 

Automatocity of the Purkinje fibres is not usually seen, as their native rate is quite slow (~40 beats per minute) and they end up being overridden by faster pacemakers. However, left to their own devices (eg. in the abovementioned case of a cryoablated AV node), their automaticity can occasionally appear. Unless 

Activation of the myocardium by the distributed His-Purkinje system occurs in a systematic and organised fashion, as anything else would lead to some sort of comically inefficient pumping behaviour. With profound apologies to Durrer et al, the excellent isochronic region diagram from their paper is reproduced here in a minimally mangled state. These were hearts of dead men, who suffered cerebral catastrophes. They were removed within half an hour, reperfused, and then maintained for up to six hours in this reanimated state. Data collected from numerous tiny electrodes was printed out on an ink writer to be analysed, with the paper speed running at a grant-devouring 960cm/sec. Given the scale of the study, the investigators would have had to analyse hundreds of metres of output.

"Isochronic" meaning "at the same time". This  diagram illustrates that the left side of the septum activates first, within 5-10 msec of the action potential arriving. Apical subendocardial regions of the left and right ventricles are next, within 5 msec. By the next 5 msec pause, most of the subendocardium is activated and from there the impulse propagates outwards to the epicardium. The last to be activated is the posterobasal wall of the right ventricle, after about 60-70 msec. That's a pretty short QRS, and this is because the were the hearts of otherwise healthy dead young men. Durrer and colleagues were able to obtain the hearts within 30 minutes of death, and then maintained them extracorporeally in a warmed oxygenated broth of washed bovine erythrocytes. "The hearts continued beating in a spontaneous sinus rhythm for a period ranging from 4 to 6.5 hours", the authors boasted. 

Ventricular myocytes

The ventricular myocytes are the last and most boring part of the excitatory-conductive-contractile axis. "Contractile" is what they do best, and most people will agree it is better for them to stick to what they know- in fact for normal survival to continue they should remain conductive and excitatory only in the most passive sense.

Automaticity of the ventricular myocytes should be guarded against, and like the atrial muscle, the ventricle myocytes really lack any capacity for this outside of the most horrendous conditions. Most of the time, they coast at a stable resting membrane potential (around -90 mV) and do not have any self-depolarising currents. This, of course, can change if the cell undergoes some sort of metabolic stress. A good example is ischaemia and hypoxia: the cells can depolarise because the resting action potential starts drifting up into the -50 mV territory. The more positive the membrane potential, the higher the likelihood that the cell will act as an ectopic pacemaker, and the higher the rate of any such pacemaker (Katzung & Morgenstern, 1977)

Conduction velocity along this pile of muscle is relatively sluggish. For the LV, intramural conduction velocity was recorded as around 0.5 m/sec by Durrer et al (1970). For this reason, we have a specialised conduction system. 

Structure and mechanical function of the ventricles can basically be summarised as "thick muscular tube designed to propel oxygenated blood to the peripheral tissues", and one could seriously leave things there. However, when the college syllabus writers concocted their line about "describe the structure and functional significance", one must assume they meant all of the anatomical elements to be included. Thus:

• The muscular mass of the right ventricle is a conical muscular pouch with one corner formed by the tricuspid valve, another by the pulmonic valve, and the third by the apex. Its internal chamber shape is usually made concave by the bulge of the interventricular septum.
• The muscular mass of the left ventricle  is a prolate ellipsoid, circular in crossection, with significant nonuniformities in its internal wall geometry.    
• Functionally, the ventricles contract in order to expel blood through their efferent valve, then relax (actively) in order to entrain blood through their afferent valve.