Determinants of preload

This chapter is relevant to Section G2(iii)  of the 2023 CICM Primary Syllabus, which asks the exam candidate to "define the components and determinants of cardiac output". Specifically, preload is the topic here. Question 15 from the second paper of 2015 and Question 9(p.2) from the first paper of 2010 both asked for a discussion of preload and its determinants.

In summary:

• Preload can be defined as:
  • Myocardial sarcomere length just prior to contraction, for which the best approximation is end-diastolic volume
  • Tension on the myocardial sarcomeres just prior to contraction, for which the best approximation is end-diastolic pressure
• The determinants of preload, if we choose to define it as a a volume, are:
  • Pressure filling the ventricle:
    • Intrathoracic pressure,
    • Atrial pressure
      • Atrial contractility and rhythm
      •  
      • Atrioventricular valve competence
      • Ventricular end-systolic volume
      • Ventricular compliance
    • Right atrial pressure
    •  
    • Mean systemic filling pressure
      • Total venous blood volume
      • Venous vascular compliance
    • Cardiac output, insofar as it supplies the total blood volume
  • Compliance of the ventricle:
    • Pericardial compliance:
      • Compliance of the pericardial walls
      • Compliance of the pericardial contents
    • Ventricular wall compliance:
      • Duration of ventricular diastole
      • Wall thickness
      • Relaxation (lusitropic) properties of the muscle
      • End-systolic volume of the ventricle (i.e. afterload)

Surely, the reader might point out, that is enough. For for the purposes of exam preparation it would have sufficed to define the components and determinants as asked, in the interest of brevity (it being the soul of wit, etc). Accordingly, the main players on this stage are introduced and briefly acknowledged in a separate chapter ("Definitions of preload afterload and contractility"), and the revising trainee is directed there instead.  What follows here represents a failure of the author to exercise even a minimum of self-restraint, a long-form exploration of preload the character of which makes it impossible to determine who the target audience are. 

Definition of preload

Preload, like obscenity, is hard to define. In the words of Carl Rothe (2003), this field of nomenclature is infected with "variability and inconsistency that confuse not only medical students but also clinicians and professors". Ergo, it seems unfair to ask the CICM primary candidate to synthesise in ten frantic minutes what the Illuminati of physiology have struggled with for the last two centuries. However, that has not stopped the CICM examiners. Specifically, they have asked for a definition of preload at least twice. Only in their comments to Question 9(p.2) from the first paper of 2010 did they ever develop the courage to formulate their own version of a definition:

"A definition based on stretch of the isolated myocyte prior to contraction, and extrapolation to the human heart, was expected"

This does not so much define preload as outline which sort of definitions might have been acceptable for a passing grade. That, perhaps, is better than nothing. Of the scholars who write on this topic, there is a clear division into two broad camps: people who prefer to use force or pressure, and people who prefer to define preload in terms of volume. 

Official college resources mainly seem to be in the force and tension camp. Pappano & Weir (p. 66-67) define preload as:

"...preload is the force (load) on the muscle prior (or pre-) to its being activated to contract."

So, this group defines preload as a tension or force, implying that preload should be measured and expressed in units of pressure.  They go on to explain that "preload applies tension to the muscle and stretches it passively to a new length. In the heart, preload is the stress exerted on the ventricle during diastole and can be represented by the Laplace equation for a thick-walled sphere". There is significant support for this Laplacian model.  The best exploration of it is probably Magder (2015), who produces some arguments to the net effect that preload should be "defined as the force, or pressure, that gives the initial length of the muscle before the onset of contraction".

However, it must be acknowledged that the contractile force produced in cardiac muscle by this abovementioned stretching is proportional to its length, not to the force which produced that length. This is the competing view of the other group. A representative definition is offered in Kam (2015):

"Preload is defined as the initial stretch (diastolic fibre length) of cardiac muscle before contraction"

This group of authors emphasises stretching and an increase in fibre length, inferring that preload should be expressed in terms of dimensions, eg. as a volume. This is a view supported by Rothe (2003) and highly respected textbooks like Berne & Levy. It is also "a definition based on stretch of the isolated myocyte", as mentioned in the college answer to  Question 9(p.2). 

So, the enraged trainee may shout, where the fuck is the point here? What should we memorise for the primary exam? After reviewing what is available, for the purpose of regurgitating a box-ticking soundbite in the CICM Part One vivas, one cannot help but recommend the Part One definition ahead of these others:

"Preload is defined as myocardial sarcomere length just prior to contraction"

This is a suitable (memorable, short, non-insane) definition of preload which is also repeated in other formidable online resources, as well as industrial monoliths such as the Encyclopedia of Intensive Care Medicine (Vincent & Hall, 2012). It is certainly not inferior to the published alternatives, and has the advantage of brevity. Plus, we online non-peer-reviewed resources have to stick together. It would probably be important to also satisfy the college examiners' need for "extrapolation to the human heart", which would consist of mentioning that this definition of preload is restricted to the isolated cardiac muscle fibre, and that surrogate parameters such as the end-diastolic volume are often substituted for preload in vivo.

Determinants of preload

If we agree that "preload is defined as myocardial sarcomere length just prior to contraction", then it follows logically that preload and end-diastolic volume are very closely related.  In fact myocardial sarcomere length must be defined by end-diastolic volume, because these sarcomeres comprise the walls of the myocardium, and a change in their length dimension must surely accompany any change in ventricular volume. Volume is generated when blood flows along a pressure gradient into an elastic vessel; the exact volume is therefore dependent on the compliance of that chamber, i.e. the volume which is generated per unit of applied pressure. Thus, end-diastolic volume can probably be described by some facile series of nesting dependencies, as follows:

• Pressure with which the ventricle fills,
  which is determined by:
  • Intrathoracic pressure,
    which will affect the right and left ventricle differently
  • Atrial contribution ("atrial kick"),
    the magnitude of which depends on:
    • Atrial contractility and synchrony (i.e. sinus rhythm or AF)
    • Valvular function 
    • Left ventricular end-systolic volume
    • Left ventricular compliance
  • Central venous pressure, or more accurately the right atrial pressure
  • Mean systemic filling pressure,
    which has its own determinants:
    • Total venous blood volume
    • Venous vascular compliance
  • Cardiac output, insofar as it supplies the total blood volume
• Compliance of the ventricle,
  which is determined by:
  • Pericardial compliance:
    • Compliance of the pericardial walls
    • Compliance of the pericardial contents (usually, an incompressible fluid)
  • Ventricular wall compliance,
    which depends on:
    • Duration of ventricular diastole
    • Wall thickness
    • Relaxation (lusitropic) properties of the muscle
    • End-systolic volume of the ventricle, 
      which is at the mercy of afterload.

Let us unpack these concepts in some detail.

Effect of intrathoracic pressure on end-diastolic pressure

The respiratory changes in intrathoracic pressure affect end-diastolic pressure (and therefore volume) in a number of ways. It is basically a brutally stupid hydraulic effect: an increased pressure inside the chest opposes the venous pressure of blood trying to enter the chest. The result is reduced right ventricular preload, which in turn gives rise to reduced left ventricular preload (as they are connected in a series). Left ventricular preload is also affected directly, as left ventricular afterload decreases, giving rise to a lower end-systolic volume. These matters are discussed in greater detail (and with stupid diagrams) in the chapter on the haemodynamic effects of positive pressure ventilation.

In summary:

• Effect of increased intrathoracic pressure on the right heart:
  • Increases RA and RV pressure;
    • Thus, decreases the pressure gradient for blood flow into the cardiac chambers
    • Thus, decreases the RV end-diastolic pressure (as relative to extrathoracic central venous pressure)
    • Thus, decreases RV end-diastolic volume
• Effect of increased intrathoracic pressure on the left heart:
  • Difference between intra-LV pressure and intrapleural pressure decreases the LV transmural pressure
  • This decreases the afterload
  • Decreases LV end-systolic pressure, because of reduced afterload
  • Decreases LA pressure, because of decreased RV cardiac output 
  • Decreases LV end-diastolic pressure, because  of the above

So, that's positive intrathoracic pressure. Negative pressure (eg. in a person spontaneously taking a deep breath, or somebody ventilated with NEEP) predictably has the opposite effects. This included the increase  in LV afterload.  Hausknecht et al (1987) explored this in a dog model, and determined that fundamentally the effect of negative intrathoracic pressure on left ventricular afterload was functionally indistinguishable from partial aortic occlusion.

In summary:

• Effect of decreased intrathoracic pressure on the right heart:
  • Decreases RA and RV pressure
  • Thus, increases blood flow into the cardiac chambers
  • Thus, increases the RV end-diastolic pressure (as relative to extrathoracic central venous pressure)
• Effect of decreased intrathoracic pressure on the left heart:
  • Difference between intra-LV pressure and intrapleural pressure increases the LV transmural pressure
  • This increases LV afterload, and therefore increases LV end-systolic pressure
  • Thus, increased LV end-diastolic pressure

The bottom line? Positive intrathoracic pressure decreases preload, with nontrivial haemodynamic consequences. This becomes clear from this data set by Jardin et al (1981) via Nunn (1984), where the cardiac index of ARDS patients is plotted against the PEEP. Observe the gradual decrease of the cardiac index and MAP with increasing intrathoracic pressure. Then watch how, at the higher pressure (PEEP of 30), the restoration of preload (fluid bolus of 10ml/kg) reverses all of the haemodynamic badness.

Effect of atrial "kick" on ventricular preload

The atrial "kick is a small contribution to the LV end-diastolic volume which is made by the atrium at the very end of the ventricular diastole. Without revising every element of the cardiac cycle, it will suffice to say that at the end of diastole, atrial systole increase the gradient of pressure between the atrium and the ventricle, which transiently increases the left ventricular end-diastolic filling and which contributes a small fraction of the final LV end-diastolic volume.

The volume contributed by the atrium is relatively small. It is said to be approximately 20% of the end-diastolic LV volume (Namana et al, 2018), or 24 ml for an average LVEDV of 120ml.  The left ventricular volume-time curve of an athletic young subject from Pedrizzetti et al (2003) used in the diagram above seems to give that sort of value. On average, normal LA volume ranges (indexed to BSA) are usually 19-41 ml/m², or 36-77ml for a normal male with a BSA of 1.9 m² (Aune et al, 2009), and the normal LA ejection fraction is around 55%. 

The bottom line: atrial contraction increases preload, and by logical extension, the loss of atrial contraction (eg. with AF) decreases preload. How much of an influence is this on the grand haemodynamic scheme of things?  One may occasionally hear older intensivists remark that the atrial kick can contribute 30% to the cardiac output, which probably does not come from any specific study, and which is probably nonetheless true. The readers of Deranged Physiology are grateful to Dr Vidyesh Wakade for capturing this monitor screen which demonstrates the contribution of the atrium beautifully (he captured the exact moment where it happened):

As you can see, initially the patient is in a junctional rhythm, and then develops some spontaneous atrial activity (P waves appear midway along the screen). The immediate haemodynamic benefit manifests as an increase in systolic blood pressure, essentially with the next beat.

Obviously, the atrial contribution to preload is going to be more significant in some situations as compared to others. Specific examples where it is essential are listed and discussed in the excellent chapter by Kurapati & Lowery (2019). They include:

• Tachycardia, where diastolic filling is otherwise inadequate: passive diastolic filling does not work half as well when there is insufficient time, and the atrial contraction at the end of diastole becomes essential. Its loss is acutely felt in atrial tachyarrhythmias, such as AF.
• Diastolic heart failure, or "Heart failure with preserved ejection fraction (HFpEF)" or where the compliance of the ventricle is poor, becomes atrium-dependent because the LV requires a much higher pressure to fill up to a normal volume. That pressure cannot be achieved by passive filling alone, and if the pulmonary venous pressure is left to do all the work, the LV ends up grossly underfilled, with haemodynamically disastrous consequences. 
• Aortic stenosis, where the left ventricle is still almost completely full at the end of systole, and is therefore in a poorly compliant part of its pressure-volume curve. In this scenario, the only way to squeeze any more blood into the LV is by atrial contraction. In many ways, this situation is similar to what happens in diastolic heart failure.
• Mitral stenosis,  where the mitral valve is the source of resistance to flow. In order to fill the LV, a significant pressure gradient between the atrium and the ventricle must be generated, and this can only happen if the atrium is contracting normally.

Central venous pressure (right atrial pressure) and preload

During most of the right ventricular diastole, as the tricuspid valve is open and the atrium is still, the right ventricle is being filled by central venous pressure. An increase in the central venous pressure will therefore result in an increase in the right ventricular volume, provided the RV compliance remains the same. In that sense, CVP is a determinant of preload.

More importantly, the rate of blood flow back to the heart is also determined by the CVP (or rather, the right atrial pressure), as this flow is driven by the pressure gradient between the mean systemic filling pressure (MSFP) and the right atrial pressure:

This relationship between CVP and preload was taken very seriously in the early dark ages of intensive care medicine, to the effect that CVP monitoring was commonly used to determine the response to fluid resuscitation, i.e. used as a surrogate for preload. It has now become apparent that CVP is actually very poor at this, mainly by virtue of being affected by numerous unrelated factors.

Mean systemic filling pressure as a determinant of venous return

"Mean systemic filling pressure" and "mean circulatory filling pressure" are concepts invented by Arthur Guyton to describe the relationship of venous blood flow to the regulation of cardiac output. They certainly catch the eye, for example because they require a state of complete cardiac arrest to be measured accurately. For an indepth discussion of these concepts, a whole separate page has been dedicated. In short:

• Mean circulatory filling pressure (MCFP) is the pressure in the entire circulatory system in the absence of flow. It is the pressure exerted by the walls of the circulation (including the heart and pulmonary vessels) on its fluid content, and so can be thought of as a measure of the elastic recoil potential stored in those walls.
• Mean systemic filling pressure (MSFP) is the pressure in only the systemic circuit, i.e. ignoring the heart and pulmonary circulation, also in the absence of flow.
• The two main factors that determine the MSFP are vascular tone and blood volume.
• Tone of the smooth muscle contributes to MSFP by altering the the elastic recoil pressure which produces the MSFP. With noradrenaline, MSFP is 14-15 mmHg, whereas in the absence of sympathetic tone it is probably closer to 5.6 mmHg.
• Volume in the circulatory system places a pressure on the walls of the vessels. Only about 15% of the total blood volume is contributing to generating the MSFP, as most of the volume is described as "unstressed". In other words, 85% of the blood volume is not exerting very much pressure on the vessel walls.

Effect of pericardial compliance and content on preload

The pericardial sack is often described as "inextensible", which basically means that a closed pericardium should place a finite upper boundary on the volume to which a heart can expand. It is also a constraint before that maximum is reached. Janicki & Weber (1980), plotting the pressure-volume relationships of dog ventricles pre and post pericardectomy, determined that the pericardial sack decreased the right and left ventricular compliance:

 In short, at any given end-diastolic pressure, the volume was greater when the pericardium was removed. This situation is obviously exacerbated where some sort of incompressible fluid (eg. blood and clot) fills the pericardial cavity. All sorts of misbehaviour ensue.

Effect of diastolic time on preload

Diastolic time is an important determinant of preload, particularly if one has agreed to describe it in terms of end-diastolic volume. Bristow et al (1963), in dog ventricles, demonstrated that heart rate had a significant (and almost linear) influence on end-diastolic volume and stroke volume:

From this, it follows that severe tachycardia should give rise to haemodynamic instability due to poor diastolic filling and resulting low cardiac output. That is indeed what happens when the heart rate is so high that there is absolutely no time to fill the LV between beats:

This excellent trace comes from Goldreyer et al (1976). The investigators selected a group of young people who all presented with complaints of syncope-inducing SVT. They then proceeded to torture their atria with electrodes. The inset diagram demonstrates stroke volumes which were recorded during the same study.

Effect of ventricular wall compliance on preload

Preload, whether you discuss it in terms of pressure or volume, will be influenced by the compliance of the LV wall. A poorly compliant ventricle means a lower volume for any given pressure, or a higher pressure required to generate the same level of sarcomere stretch. Either way, compliance is an important player.

Ventricular compliance has two major components. One is relatively fixed, i.e. not amenable to manipulation clinically - that is the stiffness of the internal cardiac fibrous skeleton, as well as the mechanical properties of the meaty myocardial muscle (for example, its weight). The other is the active relaxation which takes place in diastole. With the latter, even the relatively beefy left ventricle can relax enough for the pulmonary venous pressure to exceed ventricular diastolic pressure, allowing blood to move into the chamber (in fact it ends up being more compliant than the left atrium). Usually, when compliance is affected in a real in vivo heart, both of these components are affected; i.e the infarcted ventricle will overgrow with fibrous tissue and stop responding to normal lusitropic signals from friendly catecholamines. 

An excellent article by Gaasch et al (1975) goes into detail - too much detail - about the various ways in which a ventricle can refuse to accommodate blood volume. Fortunately for the reader, it is paywalled. And even this digression-prone author would hesitate to ambush his readers with overlong discussions of ventricular volume/mass ratios and the modulus of chamber stiffness. It would probably suffice to summarise that preload is decreased by either changes in compliance due to wall thickening, or wall fibrosis, or unreliable lusitropy, and often all three in combination. It would probably also be helpful to exploit the explanatory potential of an original diagram from Gaasch:

Effect of afterload on preload

Afterload and preload are interdependent, connected to each other by the ventricular end-systolic volume. Though the reader may object that the definition of afterload cannot be reduced to anything so glib as systemic vascular resistance, they would also have to agree that the properties of the extracardiac circulation are important in describing the forces acting against the ventricle when it is trying to eject blood. To make such people happy, we could refer to afterload as "arterial input impedance" or somesuch.  Using a ventricular pressure-volume relationship, we can see how this property can increase preload. Observe how an early end to systole increases the end-systolic volume and decreases the stroke volume:

To spell it out in point form:

• Increasing afterload (arterial input impedance) increases end-systolic pressure (i.e. the aortic valve closes earlier)
• This means that systole ends at a higher end-systolic pressure, and the stroke volume is decreased
• A decreased stroke volume, subtracted from the total ventricular volume, leaves behind a higher end-systolic volume.

From this, it follows that further increases in afterload would progressively reduce the stroke volume, and that, theoretically, there would be an afterload at which there would be no stroke volume because the ventricle would not be able to overcome the arterial input impedance, and would only be capable of isovolumic contraction. This was exactly the finding by Burns et al (1973), who measured the stroke volume of dog hearts:

Obviously this cannot be a normal state of affairs. The stroke volume has to be maintained somehow, otherwise the circulatory system would collapse with even minor changes in afterload. Indeed, in the dog heart experiments by Burns et al (1973), increasing afterload only led to this linear decrease in stroke volume while the investigators used a computerised pump to maintain a constant LV filling pressure in real-time. If they did not, the LV would compensate for the increase in afterload by increasing its preload. 

How was this was happening?  To be sure, it makes intuitive sense for the ventricle to increase its contractility when the end-systolic volume is increased. In fact this is the Anrep effect.  The higher end-systolic volume that is left behind after a systole with high afterload results in an increase in sarcomere stretch, which in turn leads to an increase in the force of contraction for the next beat, and this is fine and good. A reader with a passing understanding of the Frank-Starling relationship will not find this concept difficult to parse. But this affects contractility, not volume. What could produce the necessary increase in end-diastolic volume that is needed to maintain the stroke volume?

At this stage, reputable resources from all corners of the internet tend to repeat the same thing: that the end-systolic volume is added to the venous return. This statement can be triggering to the attentive reader, because they will readily identify a paradox that others might have passively accepted (we are grateful to Brenda for pointing this out and enriching this discussion thereby).  Consider: the stroke volume is a factor of cardiac output, and cardiac output is the same as venous return; but the stroke volume drops with high afterload, so surely so must cardiac output and venous return. Logically, it follows that with a fixed high end-systolic volume the falling venous return added to the decreasing stroke volume would produce a cycle of ever-decreasing stroke volume until the circulatory system collapses to wallow in some kind of mean systemic filling pressure. 

However, one must recall that the circulatory system is a loop.  The conditions of each beat do not propagate instantly through the circulatory system; the blood returning to the heart as preload was ejected as a stroke volume perhaps 15-20 seconds and 20-30 beats ago. As such, the small volume filling the ventricle following an abrupt increase in afterload is not the same as the volume that is filling it during the next beat. That volume was calmly making its way through the venous circulation and has no awareness of any afterload shenanigans.

This returning volume will be something like the last normal stroke volume (say, 70ml) and will attempt to enter the ventricle during diastole. But the ventricle is already occupied by the residual volume left behind by the systole that had a high afterload, and is therefore unexpectedly distended. As such, not all of the returning 70ml will be able to fit. How much additional filling will occur? That would actually be difficult to predict without confusing the discussion with talk of venous compliance and atrial contractility; all that can be said is that:

• The venous return will add to the end-systolic volume to some extent
• The added volume will be smaller than  the stroke volume that was present before the increase in afterload
• The end-diastolic pressure will be (slightly) higher because of the effect of trying to push the same volume of venous return into a less compliant ventricle, i.e. the poorer ventricular compliance at higher volumes will limit how much additional volume can be achieved in this diastole.
• The decreased stroke volume as the result of this interaction will produce a decrease in cardiac output, which will be exactly proportional to the change in stroke volume, which means from here on, the system will come to rest at an equilibrium where the volume of venous return is the "correct" volume (i.e. it exactly matches the volume required to fill the venticule to the new higher end-diastolic volume and pressure).

Again, we have cruel animal studies to support this. Bugge-Asperheim & Kiil (1973) observed the effects of progressively increasing LV afterload, achieved by the mechanical constriction of the aorta, and measured "myocardial chord length" - basically the distance between two piezoelectric elements embedded into the LV tissue, a suitable surrogate for sarcomere stretch. The vandalised results of this experiment are presented below:

Several things happened here. The stroke volume (at the bottom of the tracings) dropped by about 17%. At the same time, end-diastolic pressure increased modestly. The "chord length" also clearly increased. Tellingly, the minimal chord length (ie. the end-systolic length) increased the most, which is evidence of an increased end-systolic volume:

So, in summary, afterload increases preload because:

• An abrupt increase in afterload increases the end-systolic volume
• To this end-systolic volume, the previously stable venous return volume is (incompletely) added
• The effect of this is a higher end-diastolic volume and pressure (as permitted by the poorer LV compliance at higher volumes)