Myocardial oxygen supply and demand

 This chapter is relevant to Section G3(iii)  of the 2017 CICM Primary Syllabus, which asks the exam candidate to "describe myocardial oxygen demand and supply, and the conditions that may alter each".  It also has some exam relevance; Question 7 from the second paper of 2016 asked the trainees to compare the supply and demand of each ventricle separately, whereas Question 7(p2) from the second paper of 2007 asked about the effects of tachycardia.

In summary:

In terms of peer-reviewed resources, the best free article is Duncker & Bache (2008), which has a strong exercise physiology theme but is otherwise generic enough to work for this topic.  The 2013 chapter by Kern & Lim from the 8th edition of"Grossman & Baim's Cardiac Catheterization, Angiography, and Interventions" is also superb, and is somehow available for free from thoracickey.com. This reference was lurking at the bottom of the Part One summary for this topic, which remains the standard in brevity and precision - a trainee who is short on time should stop reading this right now, and go there instead. On the other hand, a trainee who has for some reason hyperfixated on this topic will find their disturbing preoccupation patiently indulged by Cardiac Energy Metabolism in Health and Disease by Lopaschuk & Dhalia (2014), and Jos Spaan's Coronary Blood Flow: Mechanics, Distribution and Control (1991). 

Myocardial blood flow

The myocardial blood supply consists of the coronary arteries. This specialised regional circulation terriotory is sufficiently interesting to merit its own chapter, for multiple reasons not the least of which being the larger number of CICM SAQs which it seems to attract. In the interst of brevity, the salient points can be summarised as follows:

• Coronary arterial anatomy: 
  • 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
• Coronary arterial blood flow:
  • The heart receives about 5% of its own output as blood supply,  of which the majority (~ 75%) occurs during ventricular diastole.
  • The flow is biphasic
    • For the LV, most of the flow (75%) occurs in diastole because during systole the intraventricular pressure is greater than coronary arterial pressure at the same time, which results in poorer perfusion. 25% of the flow still occurs in systole.
    • For the RV, which is under less pressure, the systolic/diastolic distribution of flow is 50:50
  • Total flow to the myocardium is about 80-160 ml/100g/min (5% of cardiac output) at rest

Myocardial oxygen delivery

The oxygen delivery to anything is determined by its blood flow and by the oxygen content of the blood. That means it is determined by this equation:

Coronary O₂ delivery  = Coronary blood flow × (sO₂ × ceHb × BO₂ ) + (PaO₂ × 0.003)

Where

• Coronary blood flow = 80-160ml/100g/min (wide range of reported values)
• ceHb = the effective haemoglobin concentration (Let's say 100g/L in ICU patients)
• PaO₂ = the partial pressure of oxygen in arterial gas (let's say about 75 in ICU patients)
• 0.003 = the content, in ml/L/mmHg, of dissolved oxygen in blood
• BO₂ =  the O₂ carrying capacity of blood (normally 1.39ml/ml)
• sO₂ = oxygen saturation (let's face it , its going to be 90-100%)

So, basically all of the variables mentioned here are under very tight control in the ICU. For the classical anaemic ICU patient whose haemoglobin is 100g/L, the total oxygen content of 1L of coronary arterial blood will be about 130ml at 100% oxygen saturation. So, if we take 100ml/min/100g as some sort of reasonable average value for coronary blood flow, and assume a 130ml/L oxygen content in the coronary blood, we come to the conclusion that every minute the heart of an average ICU patient receives about 13ml of oxygen per 100g of myocardial tissue.

But how much does it actually need?

Myocardial oxygen consumption and demand

A pedant would first make the point that myocardial oxygen consumption is not the same as myocardial oxygen demand, as logically there are situations where the demand is greater than consumption (i.e. where supply is limited), for example in ischaemic heart disease. However, among healthy laboratory animals and floating myocytes in Petri dishes, there is usually no shortage of oxygen supply, and so this objection can be put to rest for now. 

At a basic level, cardiac myocytes are far from lazy, even while doing what appears to be nothing. Left to their own devices in a warm bath of benign electrolytes, working ventricular myocytes will not contract, quietly waiting for an action potential to stimulate them ("mechanically quiescent" is the scientific term). However, even when doing nothing, they are still using a substantial amount of metabolic substrate. Spieckermann & Piper (1985) found the oxygen consumption of such quiescent myocytes was about 1 μmol per minute per gram of wet muscle tissue, which is also what one might expect in cardiac arrest, or during cardioplegia. To convert it into more familiar units, considering that 1 ml of an ideal gas contains 44.64 µmol of that gas, this oxygen consumption rate works out to be 0.0224 ml/g/min, or 2.24 ml/100g/min. Compare that to the total oxygen consumption of the human body at rest, which is generally said to be 3ml/kg/min, or about 0.3ml/100g/min.  Thus, even while totally motionless and paralysed, myocardial tissue is extremely energy-hungry. 

Obviously doing something has a higher metabolic cost than doing nothing, and the more something you do, the greater the oxygen cost. And the myocardium is never "mechanically quiescent", until your actual death. Most of the time it is said to be "resting", but by "resting" we mean "contracting about sixty times per minute". Ergo, the resting oxygen consumption is normally higher than the abovementioned theoretical minimum. Most sources (eg. Rooke & Feigl, 1982) tend to quote about 6-8ml/100g/min, and this is clearly not a very precise number, as it is an in vivo measurement which obviously depends on a lot of other variables.

When whipped along with catecholamines, the oxygen consumption of cardiac myocytes increases significantly. At maximum stimulation, Spieckermann & Piper were able to get their adrenaline-enraged myocytes to burn through 40 μmol per minute per gram, i.e. increasing their oxygen consumption rate by 40 times.  That'd be 89.6ml/100g/min. That probably represents some sort of functional maximum in an idealised setting; these cells were sitting in a petri dish with unrestricted oxygen supply and were contracting with zero loading conditions. For a summary and a comparison of these values with other tissues, a table of vague rounded figures is often seen in various textbooks:

Comparative Oxygen Consumption of Different Tissues

Tissue type	Oxygen consumption in ml/100g/min
Myocardium in cardiac arrest	2
Myocardium contracting at rest	6-8
Myocardium at maximum inotropy	90
Brain	3
Kidney	5
Skin	0.2
Resting skeletal muscle	1
Skeletal muscle during exercise	50

It is of course immediately clear to any sane reader that these numbers are some kind of completely irrelevant bullshit. A whole bunch of other weird figures also get thrown around, with no apparent relevance to clinical work, nor educational benefit. Presumably their position in the grand scheme of human knowledge is similar to that of the snackable factoids found in the margins of popular science books, eg "did you know that the heart pumps ten tons of blood per day?". This practice seems pervasive. Otherwise serious-sounding published resources start their sober discussion of myocardial energy metabolism with the facile statement that the heart "cycles about 6 kg of ATP every day - 20 to 30 times its own weight" (Neubauer, 2007, NEJM). It is of course impossible to rely on the assumption that exam question writers will see through these vulgar theatrics, and so conceivably at some stage somebody somewhere might make an MCQ out of this. 

Anyway. The bottom line is that the myocardium has a very high oxygen demand even when it appears to be under no stress whatsoever, which means it ends up removing much of the oxygen delivered to it via the coronary arteries. In other words, it has a high oxygen extraction ratio. 

Myocardial oxygen extraction ratio

Remember the abovementioned oxygen delivery estimate (13ml/min/100g, where every 100g of cardiac muscle receives 100ml of blood flow). With a resting oxygen consumption of 8ml/100g/min, the heart extracts 60% of the blood oxygen content, leaving only 40% behind. This is an oxygen extraction ratio of 60%.

In actual fact, when you look at in vivo measurements, this ends up being something of an underestimate. Most textbooks will quote an oxygen extraction ratio of about 75-80%. These figures seem to come from older studies such as Binak et al (1967). The investigators collected venous samples from the coronary sinuses of healthy volunteers. The mean myocardial oxygen extraction ratio was 68%, with a range of 51-80%. The lowest coronary sinus oxygen content recorded in that series was about 20%, corresponding to an oxygen saturation of 8%.

What are the implications of this clinically? Consider: the myocardium, in the course of doing nothing particularly stressful, already needs to remove most of the oxygen delivered to it. Obviously it is going to be rather intolerant of any situation where the amount of delivered oxygen fluctuates. If blood flow is insufficient, or if the oxygen content of blood is reduced, the margins are very narrow. Thus, various commonly encountered clinical scenarios are going to be poorly tolerated: anaemia, hypoxia, hypotension, etc etc... 

Myocardial metabolism

So, where does all this oxygen go? The heart is described as an omnivore, on the basis of the fact that it performs optimally while simultaneously using several different metabolic substrates, and able to switch from one substrate to another in the face of changing conditions. Without exceeding the readers' (presumably, ultralow) capacity to tolerate biochemical digressions, it will suffice to summarise myocardial energy metabolism as follows:

• The myocardium is capable of metabolising multiple substrates. Under normal conditions, the following distribution of metabolic sources can be expected:
  • Free fatty acids (65%)
  • Glucose (15%)
  • Lactate and pyruvate (12%)
  • Amino acids (3%)
  • Anaerobic glycolysis (5%)
• Availability of substrates changes these ratios slightly:
  • Wherever the serum fatty acids are high (eg. fasting or starvation),  glucose metabolism is suppressed and fatty acid use increases to 90% of the total
  • During prolonged fasting, the myocardium can metabolise ketones.
  • Wherever glucose is abundant (eg. after a meal), fatty acid use is decreased and glucose is used instead
  • During exercise, the heart increases its intake of lactate
  • Wherever oxygen is in deficit, glucose and fatty acid metabolism are turned off (as they demand oxygen), and anaerobic glycolysis occurs, where the heart becomes a net producer of lactate instead of a net consumer.
• Cardiac pathology and changing physiological requirements can change the myocardial substrate use:
  • The adult heart mainly uses fatty acids, whereas the foetal heart mainly uses glucose and lactate
  • Ischaemia, hypertrophy and increased afterload can also increase glucose use by the myocardium
• The myocardial ATP store is small, compared to its demands
  • About 4-8% of the total myocardial ATP are consumed with each beat
  • ATP is therefore depleted very rapidly in the absence of substrate supply

Pasqual & Coleman (2016) or Stanely et al (2005) are the most comprehensive references to cover this area, in the unlikely case anybody else is interested in the extra detail. If anybody is really interested in extra extra detail, 300 pages of de Jong's Myocardial Energy Metabolism (1988) will scratch their itch.  

Energy cost of different myocardial activities

Though it might seem like an organ with a fairly restricted range of behaviours, in truth the myocardium does several different things in the course of routinely performing its repetitive duties, and each of those things contributes differently to the total metabolic cost of doing business:

Contraction	60%
Relaxation	15%
Electrical activity	3-5%
Basal cell metabolism	20%

These numbers are from Gould's Coronary Artery Stenosis, a 1991 book the contents of which is impossible to find online, and therefore it is impossible to determine what sort of experiment they had come from. Moreover, whenever one encounters such a table in the literature, the values are always different, and occasionally by a factor of magnitudes. For example, when  Klocke et al (1966) looked at the energy cost of cardiac electrical activity, they found something completely different. The investigators perfused a dog heart with a calcium-free solution, thereby abolishing all mechanical and electrical activity. When the heart was subsequently paced, the oxygen consumption increased by only about 0.04ml/100g/min, this increase being purely due to consumption by the conducting system. That's only about 0.5% of the total cardiac oxygen consumption, contrary to the table from Gould. In short, caveat lector. 

Occasionally, one also finds a breakdown of the different oxygen costs for separate items of myocardial work, like some sort of weird invoice. Usually, textbooks will mention internal work and external work, where external work is the work done to eject the ventricular stroke volume, and internal work is all the other work, eg. what is done on the ventricle to change its shape and to change the pressure during isovolumetric phases. 

Given that work is the product of pressure and volume, these concepts can be easily represented on the pressure-volume loop:

Turns out, this is not a purely theoretical concept. PV loop area corresponds to ventricular oxygen consumption in a totally linear way. Takaoka et al (1992) were able to confirm this in human hearts. They even inflated balloons in their volunteer's IVCs to generate a series of loops with decreasing LV volumes, so that the ESPVR could be estimated. In short, this is a well-established relationship.

According to the otherwise excellent college answer to Question 7 from the second paper of 2016, about 85-90% of all cardiac oxygen consumption is spent on internal work; it is not clear exactly where they got this from and it does not seem entirely legitimate. For example, Suga et al (1983)  reported something very different. This assertion can be tested by looking at real pressure-volume loops and examining the relationship between their "stroke work" (i.e the pressure-volume loop area) and their "internal work". The areas are often similar. Certainly, the stroke work can be lower than internal work- it could theoretically even be zero (eg. in totally isovolumetric contraction where no stroke volume is ejected, the PV loop would have no area).  To quote Hiroyuki Suga,  "Stroke work, or external mechanical work, is a part of PVA and varies from as low as 0% of PVA in an entirely isovolumic contraction to as high as 70-80s of PVA in ejecting contractions with high ejection fractions". However, if you define external work purely as "work done to eject a volume", the relationship changes to something m

Determinants of myocardial oxygen demand

From the above, it follows that anything that changes the shape of the PV loop will change the myocardial oxygen demand. Thus, preload, afterload and contractility will all affect the myocardial oxygen demand for each beat. If one considers the demand in terms of oxygen required per unit time (eg. per minute), then heart rate also factors in, as each cardiac cycle has an oxygen cost. Additionally, the basal metabolic rate of the quiescent myocardium needs to be factored in, as well as the factors which influence it (eg. temperature, and drugs which modify the metabolic substrate utilisation of the heart). This exact same information looks much better when posed in the form of an unformatted list. The relationships of how much each factor changes the myocardial oxygen demand are also from Gould's Coronary Artery Stenosis (1991) and are therefore impossible to trace back to their origin.  

• Heart rate is the main determinant of myocardial oxygen consumption.
  • This is a linear relationship; i.e. increasing the heart rate by 50% increases the myocardial oxygen consumption by 50%
• Preload is a minor contributor:
  • Myocardial oxygen consumption increases by only 4% when preload is increased by 50% 
• Contractility is a major contributor;
  •  If dP/dT is increased by 50%, myocardial oxygen consumption increases by 45%
• Afterload is a major contributor:
  • Increasing the afterload (well, aortic pressure, to be accurate) by 50% increases the myocardial oxygen demand by 50%
• Coronary blood flow can also increase cardiac oxygen consumption (and contractility), which is a weird finding known as the Gregg effect (Downey, 1997), the mechanism of which does not appear to be determined at this point.
• Cost of electrical conduction:  thought to be minimal, 0.5-5% of the total cardiac oxygen demand
• Basal cost of cardiac metabolism, and the factors which affect it, which are:
  • Temperature: hypothermia decreases the basal metabolic rate of the heart, as well as indirectly affecting oxygen consumption by decreasing heart rate and contractility
  • Metabolic enzyme function modifiers, eg. perhexiline which inhibits carnitine O-palmitoyltransferase and increases the fractional proportion of carbohydrate metabolism by the heart, thereby decreasing the overall oxygen cost of myocardial metabolism.

Determinants of coronary arterial blood flow

To summarise everything written above, the myocardium is an energy-hungry organ and has a very high oxygen extraction ratio, which means that any increase in cardiac performance (and thus oxygen demand) must be met with a proportional increase in coronary artery blood flow, because more oxygen delivery cannot be achieved by increasing oxygen extraction any further.  This is in fact what happens. Coronary blood flow increases roughly in proportion to myocardial oxygen demand, and the oxygen extraction ratio of the myocardium remains remarkably stable over a wide range of cardiac performance variables. Kitamura et al (1972) forced a bunch of healthy your volunteers to exercise on upright cycles and found the oxygen extraction ratio barely changed, even as heart rate doubled and coronary blood flow tripled:

So, how does this happen? From a purely mechanical perspective, coronary blood flow is determined by only three factors: the coronary perfusion pressure (the difference between aortic and ventricular pressure) and coronary vascular resistance. 

• Coronary perfusion pressure is related to coronary blood flow in a fairly linear manner. Duncker et al (1994) were also able to demonstrate that exercise shifted this relationship to the right, i.e. there was less flow per unit pressure (which is probably due to the diastole-impairing effects of tachycardia):
  
• Coronary arterial resistance is mainly related to the resistance of the small endocardial arteries.  Coronary epicardial arteries (the big vessels known by their abbreviations, like the LAD) do not contribute very much to the resistance. In fact, De Bruyne et al (2001) found that in healthy people there was basically no pressure drop along these vessels even when the blood flow changed significantly. As one might expect, most of the resistance happens at the level of the precapillary arterioles, which are tiny vessels less than 0.1mm in diameter. This diameter is affected by a range of factors, which are discussed in great depth by Judy Muller Delp (2013). They are listed here to simplify exam cramming and regurgitation:
  • Metabolic activity: for example, in response to ischaemia and hypoxia, which produces coronary vasodilation. Adenosine is thought to be the main mediator (or at least one of them).
  • Autonomic control: the sympathetic and parasympathetic nervous systems regulate coronary vascular resistance in predictable ways: α-receptors vasoconstrict and both β₁ and β₂ receptors vasodilate.
  • Systolic compression: during systole, as the myocardium contracts, the coronary endocardial vessels are also compressed by this, and the resistance increases transiently (which is reflected in the flow-time curves discussed above).
  • Pharmacological agents: for some common examples, dipyridamole and nitrates such as glyceryl trinitrate come to mind.