Coronary blood flow

This chapter is relevant to Section G4(ii) of the 2017 CICM Primary Syllabus, which expects the exam candidate to "describe the distribution of blood volume and flow in the various regional circulations ... including autoregulation... These include, but not limited to, the cerebral and spinal cord, hepatic and splanchnic, coronary, renal and utero-placental circulations". The coronary circulation has come up several times in the past papers

• Question 19 from the first paper of 2021
• Question 11 from the first paper of 2018
• Question 10 from the second paper of 2014
• Question 11 from the second paper of 2008

It would be difficult to do this topic in detail without revisiting vast swaths of material from the chapter on myocardial oxygen supply and demand, because the autoregulation of coronary blood flow is closely linked to myocardial oxygen consumption, and so it would be hard to discuss one without the other.  

In summary:

The most comprehensive review of this topic is a 191-page textbook edited by Johnathan D. Tune (Coronary Circulation, 2014). This was the main reference from which virtually everything here has been derived, but it is not much of a recommendation, owing to its truly ridiculous size and density. A very brief overview, enough for some exam revision, can be found in Rehman et al (2019), but it is not enough to answer the CICM questions. Ramanathan & Skinner (2005) is detailed, short, and free - a winning combination for the CICM exam candidate.  

Anatomy of myocardial blood supply

In summary:

• Right coronary artery: lies in the groove between the right ventricle and the right atrium; runs anteriorly, and then posteriorly to encircle the heart
• Left main: short; divides into two branches:
  • Left anterior descending (descends anteriorly, as the name suggests)
  • Left circumflex (descends posteriorly in the atrioventricular groove)
• Anastomosis between the arteries exists at the arteriolar level

Because it defies logic to accurately represent anatomy for people who never get to look at it, the following schematic is offered in lieu of anatomical art or massive lists of distal branches:

However, it is not inconceivable that at some stage, somewhere, an intensivist will need to look at a coronary angiogram, even if only to nod thoughtfully while somebody else interprets it. For this reason,  some radiological anatomy is probably important. Raphael et al (1980) is excellent but paywalled, and so the next best resource is probably Thangavel et al (2014). That article is actually useful on a whole number of levels, and these excellent vector diagrams of coronary anatomy are only one of its useful contents:

Coronary blood flow

The heart receives about 5% of its own output as blood supply,  of which the majority (~ 75%) occurs during ventricular diastole. The coronary flow is biphasic: there is a small peak of flow during systole, then an interruption, and then another longer taller peak of greater flow during diastole. 

Hoffman & Buckberg (1976) are usually credited with this diagram, which plots the coronary blood flow on the same time axis as LV and aortic pressure, Wiggers style. Because their original paper is not available anywhere (it's trapped in an ancient issue of Prog Cardiol) and the only available copies are grainy scans, the original diagram is reproduced here only as an appendix in the corner:

These data were measured from human ventricles, which lends them an air of authenticity. However, when this topic is encountered in textbooks, often the diagram looks different. Specifically, there is often a stylised flow-time curve comparing the aorta, the left main and RCA, which looks nothing like the recording presented above. An example of this which is reproduced below comes from the old 6th edition of Barash, and this exact format is replicated everywhere, but it was not immediately obvious how these waveforms were determined because each textbook references another textbook with an infuriating circularity. Fortunately, with some patience, some original work was ultimately discovered, so that the early pioneers of physiology can receive their due credit:

Next to the sterilised textbook version of this diagram, the primitive woodcuts are presented with their original grain, labelled for clarity. The central panel is from Green Gregg and Wiggers (1935), collected from anaesthetised dogs with open chests. The authors were trying to settle the question of whether coronary flow "under normal conditions is reduced or even stopped during systole, or, on the other hand, undergoes a marked acceleration". To the right, RCA flow data from  Gregg (1937) is reproduced, also from dogs. These waveforms have subsequently propagated through medical textbooks and have become so ubiquitous that nobody feels the need to reference the original authors.

In summary, if anybody is ever asked to describe the distinct characteristics of coronary blood flow, the following features are usually expected:

• Left ventricular coronary flow is:
  • Sharply decreases during isometric contraction (and can in fact be negative)
  • Sharply increases during the early part of systole
  • Reaches its systolic maximum during the summit of aortic pressure
  • Decreases significantly with a decrease in aortic pressure (and, again, can become negative)
  • Increases sharply during isovolumetric relaxation
  • Is maximal at mid-diastole
  • Decreases gradually in late diastole, following the diastolic pressure gradient.
• Because the LV produces a higher internal pressure during systole, systolic coronary resistance is increased, and systolic coronary blood flow is lower. As a result, the left ventricle is better perfused in diastole.
• Right ventricular coronary flow is similar but is higher during ventricular systole because the right ventricular systolic chamber pressure is usually lower.
  • As a result, the right ventricle is equally well perfused in systole and diastole.
  • This means the right ventricle does not depend on diastolic filling time and is therefore less affected by tachycardia.

The total blood flow is usually reported as being about 50-120ml/100g of myocardial mass (Messer & Neill, 1962), or about 250ml/minute in total. That number, though often spotted in textbooks, is obviously going to differ depending on the circumstances in which it is measured, and is therefore rather meaningless. To demonstrate the range of reported values, the ancient study by Messer & Neill (1962) reported an LV blood flow of about 115ml/min for a normal 70kg person, whereas Mymin & Sharma (1974) reported 386 ml/min (±77). Goodwill et al (2017) give a value of 50-100 mL/min/100g for the left ventricle and 30-60ml/min/100g for the right ventricle, which gives a total of 80-160ml/min/100g. In short, pick any random value in that range and it will probably be accurate for some myocardium, somewhere.

Regulation of coronary blood flow

As discussed in the chapter on myocardial supply and demand, the energy demands of the heart are the main determinant of coronary blood flow, as the oxygen extraction ratio of this organ is so high that it cannot possibly meet demand by simply extracting more oxygen. Thus, as myocardial demand increases, coronary blood flow increases. This is probably the most important feature of this regional circulation. The coronary arteries accomplish this feat by means of changing their vascular resistance.

As with peripheral vascular resistance, coronary vascular resistance is generally the responsibility of small vessels, less than 0.1mm in diameter. The larger coronary arteries really don't contribute much to the overall regulatory mechanism - they can only alter coronary flow over decades, and only in one direction, by allowing their inner lumens to become encrusted in atheromatous filth. 

This autoregulation of flow, coupled to demand, means that the oxygen extraction ratio of the heart remains remarkably stable over a broad range of performance intensities. For example, consider the following study by Kitamura et al (1972). The investigators captured ten healthy young volunteers, catheterised their hearts,  and subjected them to nightmarish cycle ergometer exercises. As the subjects approached the maximum workload for this experiment, their coronary blood flow more than doubled (from 100ml/100g/min to around 260ml/100g/min), but the coronary sinus oxygen content barely budged, decreasing by about 25%. 

In summary, coronary blood is tightly coupled to myocardial oxygen demand, and this autoregulation is achieved by adjusting the myocardial arteriolar resistance. To borrow a turn of phrase from the otherwise useless college answer to Question 11 from the first paper of 2018, this autoregulation is the product of "metabolic, physical and neuro-humoral factors". For a satisfactory answer, a CICM trainee should probably be able to describe these, "and the relative importance of each". The excellent article by  Judy Muller Delp (2013) seems almost perfect to answer such questions, and is the main source for the discussion which follows. 

Regulation of coronary blood flow according to metabolic demand

As the metabolic demand of the heart increases, so the blood flow increases, which occurs mainly due to the vasodilation of small arterioles. Following from this, one might come to the conclusion that the heart muscle must produce some sort of vasodilating metabolite in the course of its normal function. More function means more vasodilation, means more blood flow. Certainly, that's the conclusion Starling and Markwalder came to, in 1913. Since then, we have been basically stuck at the same point, unable to elaborate further on which exact metabolite it is, or how exactly it exerts this effect. What has become clear over the ensuing century is that probably no single metabolite will ever be enough to explain this mechanism on its own, and that several regional humoural factors likely play a role. Of these, the most promising actors are discussed below:

Adenosine as a coronary autoregulatory mediator

It is certainly a vasodilator, and it is certainly released by ischaemic myocardium, and coronary arteries certainly do have adenosine receptors. However, the idea that it plays a dominant role in normal healthy autoregulation is somewhat sabotaged by the finding that the concentration of adenosine in normal exercising myocardium never actually reaches a vasoactive dose. Tune et al (2000) observed a four-fold increase in myocardial workload over which interstitial adenosine concentration remained essentially unchanged. They even blocked adenosine receptors and demonstrated that the responsiveness of the coronary circulation to exercise was perfectly intact.

Still, adenosine is listed as one of the regulatory factors in CICM examiner commends, and the trainees should probably mention it to score marks. And it probably does play some role, just not in the setting of routine day-to-day control of coronary blood flow. Most of the studies which find an autoregulatory effect associated with adenosine tend to find it in the depths of some sort of catastrophe. Its effects become more important when the myocardium is on its last legs. Delp (2013) concludes that "adenosine may contribute most to changes in coronary resistance under conditions in which extreme metabolic vasodilation predominates over other regulatory factors".

Oxygen and carbon dioxide as coronary autoregulatory mediators

They probably play some role: it would be logical to expect them to. Certainly, there is a predictable relationship between hypoxia, hypercapnia, and coronary blood flow. Specifically, both hypoxia and hypercapnia increase coronary blood flow.  Broten & Feigl (1992) produced this beautiful 3D graph to demonstrate this relationship, which is reproduced here with minor adjustments:

The authors demonstrated that, as coronary sinus blood became more hypoxic and hypercapnic, so did the coronary blood flow increase. In fact they were able to estimate the gas-dependent mechanism accounted for 20-30% of the total autoregulatory dilation. But how this happens? This remains unclear. Delp (2013)  does not think this is a direct effect, but rather something that happens as the result of other mediators being released.

Potassium as a coronary autoregulatory mediator

Potassium should be a vasodilator for the coronary arteries. It does in fact vasodilate arteries when present in the micromolar range. Murray & Harvey (1978) gave 40-μmol boluses of KCl directly into the coronary arteries of dogs and measured a 34-48% decrease in resistance. However, the change is rather short-lived, and the effect disappears over the timeframe of tens of seconds. This may play some role in transient coronary flow changes in response to immediate increases in cardiac metabolism, but it is not responsible for sustained changes in response to increased workload (for example, with continuing exercise).

Potassium-mediated vasodilation can also occur as the result of opening ATP-sensitive potassium channels. These channels are inhibited by intracellular ATP; i.e. wherever ATP is deficient, the channels open and hyperpolarise the membrane, resulting in smooth muscle vasodilation. It's hard to say how much of a role they play in the grand scheme of coronary vasodilator factors,  but it is also hard to deny that they have a definite role. Narishige et al (1993) used glibenclamide to block them, and demonstrated that the coronary circulation was no longer responsive to changes in pressure, i.e. that flow autoregulation was impaired. These channels are also thought to play a role in peripheral autoregulatory vasodilation, and are potentially one of the targets of vasodilator drugs such as hydralazine and minoxidil.

Hydrogen peroxide in coronary autoregulation

It's an attractive target for research - H₂O₂ is a highly reactive metabolic byproduct of oxygen metabolism and - if you were a regional circulatory bed - you would want to increase your own blood flow to move this metabolite out as fast as possible, as keeping it around could give rise to all sorts of unpleasant oxidative effects. H₂O₂ is definitely produced in proportion to myocardial metabolism (Saito et al, 2006), and it definitely has a vasodilatory effect, which appears to be exerted through basically oxidative damage. Apparently, vasodilation is mainly seen when the neutralising capacity of endothelial superoxide dismutase is exceeded, though H₂O₂ can also directly affect potassium channels on smooth muscle.

pH and lactate as coronary vasodilators

"Lactic acid or hydrogen ion" are mentioned in the list of coronary autoregulator mediators by college examiners in their comments to Question 11 from the second paper of 2008. They probably are, together and separately. Neutralised lactate (i.e. buffered to a normal pH) produced coronary vasodilation in concentrations as low as 3mmol/L, making is a plausible candidate for metabolic autoregulation (Mori et al, 1998). Calcium channels and potassium channels appear to be the effectors of this response, insofar as the investigators' blockade of them abolished the effects of lactate.  Regional acidaemia also tends to act as a coronary vasodilator. Ishizaka et al (1996) added HCl to the coronary perfusing solution until the pH was around 7.0, and demonstrated some significant vasodilatation, which was diminished if ATP-sensitive potassium channels were blocked.

Involvement of nitric oxide in coronary autoregulation

This endothelium-derived factor acts as a vasodilator all circulatory areas, and blocking its synthesis (eg. using L-NMMA) produces coronary vasoconstriction. However, over the normal range of flows and pressures, it does not seem to be responsible. Smith et al (1992) determined that it does most of its work when flow to the myocardium is markedly decreased, i.e. where ischaemia would take place.

Other vasoactive influences on the coronary circulation

This is a group of factors which can be loosely described as "non-metabolic" coronary arterial vasoconstrictors or vasodilators. In other words, they may have some effect on the coronary circulation, and this effect may even be associated with an increase or decrease in myocardial metabolic demand, but the two things are not directly related. This group includes intrinsic arterial regulatory mechanisms, the effects of exogenous drugs, and the effects of the autonomic nervous system.

Importance of "myogenic" mechanisms

All arterioles tend to have this property, where they constrict in response to increases in intraluminal pressure and dilate in response to decreases in intraluminal pressure. This is a totally mechanical effect which appears to be an intrinsic property of vascular smooth muscle; in the sense that removing the endothelium from the vessels does not seem to alter this response. It is also not something unique to the coronary circulation. Nor is it a "metabolic" autoregulatory response, as it does not respond directly to changes in myocardial metabolic demand. 

Autonomic control of coronary blood flow

The coronary microcirculation has sympathetic receptors, is well innervated by the autonomic nervous system, and responds to circulating sympathomimetics. Chilian (1990) discusses some of the experiments which were used to determine these facts. In short, the coronary circulation is well-supplied with α₁-adrenergic receptors, and the infusion of noradrenaline can cause vasoconstriction. Paradoxically, "autoregulatory escape" can also occur, with metabolic factors taking over and instead producing vasodilation (probably in response to increased demand produced by the systemic changes in afterload). 

The coronary circulation is also well-supplied with β-adrenoreceptors. These seem to be responsible for coronary vasodilation, even when you take into account the fact that activating them systemically would normally give rise to increased coronary workload and therefore increased metabolic demand. 

The parasympathetic nervous system obviously also plays a role in the overall balance of coronary vasomotor tone. Parasympathetic stimulation tends to produce coronary vasodilation, and this effect can be overcome by atropine (Feigl, 1969).

Pharmacological agents which influence coronary blood flow

All sorts of drugs can produce coronary vasodilation. The old article by Schwartz & Bache (1987) or the slightly less old article by Orlandi (1996) produce a decent list. For some weirdness, one can also look at the ancient manuscript by Charlier (1961) which lists such medieval tinctures as "spleen extract" and camphor. The modern ICU trainee would not be expected to produce an exhaustive list, and would be discouraged from mentioning anything too exotic. A sensible list would include:

• Vasodilators:
  • Dipyridamole
  • Adenosine
  • Nitrates, eg. glyceryl trinitrate
  • Theophylline
• Vasoconstrictors:
  • Vasopressin and terlipressin
  • Noradrenaline and phenylephrine
  • Indirectly, COX inhibitors