Maestrini, Frank, Starling, Guyton…….and Fluids at St.Emlyn’s

fluids stemlyns

Intravenous fluids are one of the most frequently implemented interventions that we, as doctors, prescribe. We give maintenance fluids (not so frequently in the ED) and we give resuscitation fluids. Recently the results of the SPLIT trial were published and the FOAM world came alive with debate and comment on which intravenous fluids we should be using. I want this post to be a recap of the physiology of the effects of administering fluids in the ED. Frequently we prescribe and deliver fluid boluses or fluid challenges to patients either to correct hypotension, tachycardia, reduced urine output and to treat sepsis. It was not until I began my ACCS (Acute Common Care Stem) training that I started to think more critically about why I was giving a fluid bolus and what response I was looking for. I would hate to tar Foundation doctors with same brush, but as a Foundation doctor I think I was learning so much about everything (and still am) that I perhaps didn’t ruminate over this as much as I should have.  I’m not going to discuss the nuances of the physiology of different intravenous fluids, but rather the physiological response of the cardiovascular system to fluid.

Let’s start with some basics.

CO = SV x HR

Straightforward, but this is the basis of our resuscitative fluid therapy.  We want to improve cardiac output by improving stroke volume. Alas, as usual it is not quite as straight forward as that.  First we must understand the work of Maestrini, Frank and Starling.  For those who don’t know Dario Maestrini, he was an Italian physiologist who in 1914 began the work that would establish the principles that became more commonly known as Starling’s Law.  Otto Frank did similar work in conjunction with Ernest Starling, who in 1918 formally described what we now know as the Frank-Starling or Starling’s Law.  As Starling was more famous he received more credit…he probably had more Twitter followers.  Irrespective of its origins, the law is incredibly important and useful.  The law, put simply, states that if all other factors are constant, an increase in the end-diastolic volume will result in an increased stroke volume.  This was a somewhat revolutionary theory, as prior to this it was thought that the heart, and the heart alone, determined cardiac output.


This is theorised to be a result of increased stretch of the ventricular wall (through increased venous return), which results in an increase in the force of contraction of the cardiac myocytes.  In addition to stretch of cardiac myocytes, stretch of the sinus node (in the right atrial wall) increases heart rate by up to 10%.  A stretched right atrium also triggers the Bainbridge reflex which by sympathetic and vagal stimulation also increases heart rate.  The graph above shows Starling curves in three different states.  As the graph shows, stroke volume increases as preload increases up to a point at which the myocytes are overstretched an stroke volume drops off.  Any ventricle and any heart operates on a number of different Starling curves.  In those patients whom we think that a fluid bolus would be useful, we are aiming to move the patient up the curve.  We should remind ourselves at this point of preload and afterload.  Preload is mainly determined by venous return and as such by the tone of the venous system and the circulating volume.  Veins can be considered as capacitance vessels and provide a buffering effect against the effects of a reduction in the circulating volume.  If, for whatever reason, the circulating volume drops the sympathetic nervous system increases venous tone, effectively squeezing fluid out of the veins to maintain an adequate end diastolic volume.  If we administer too much fluid too quickly, the autonomic nervous system does not have time to respond and we risk overloading the patient, causing pulmonary oedema.  This idea of administering fluid without causing pulmonary oedema is what is known as ‘Fluid Tolerance’.  Afterload is the resistance against which the ventricle contracts to eject blood. It is related to Laplace’s Law which effectively states that the tension on the wall of a sphere is the product of the pressure times the radius of the chamber, and the tension is inversely related to the thickness of the wall, or:

Wall tension = (intraventricular pressure x radius) / (Wall thickness x 2)

Hypertension and aortic stenosis, as examples, cause an increase in afterload.  Afterload is also analgous with Ohm’s Law which in cardiovascular is ΔP = Flow x Resistance, where ΔP is the difference in pressure that the heart generates between arteries and veins.  So we have preload (venous return) and afterload (resistance).  If this formula is reworked we have an equation that allows us to work out flow through a vessel and also a means to calculate resistance.  With regards to Ohm’s Law, we can consider the following formula: Cardiac Output = Arterial Pressure/Total Peripheral Resistance.  When we administer a fluid bolus, we look to see an increase in preload and a subsequent improvement in stroke volume.  The heart becomes the limiting factor in cardiac output when the venous return is a greater volume than the heart can pump.

Arthur Guyton, whose textbook you may have cried over at medical school, expanded the work of Maestrini, Frank and Starling and added his own theory of cardiac output.  In studies using dogs, Guyton created an arteriovenous fistula between the aorta and the inferior vena cava.  Initially the fistula was left closed, and Guyton paced the hearts of the dogs.  When the heart rate was increased, no observable difference in cardiac output was observed, however once the fistula was opened (causing increased preload and right atrial pressure) cardiac output increased in proportion to the heart rate.  This observation suggested that the heart is permissive in increasing cardiac output, however one must increase preload to increase cardiac output.  So right atrial pressure can be thought to represent preload, but Guyton also proposed the concept of mean circulatory filling pressure (MCFP).  MCFP is average pressure throughout the circulatory system and may be thought of as a measure of the elastic recoil potential stored in the walls of the entire circulatory system.  MCFP is a function of the circulating volume; increase the volume and you increase MCFP.  The driving pressure for venous return is proposed to be the pressure difference between MCFP and right atrial pressure.  Magder et al. made this concept easier to understand by using the bathtub analogy:

‘The rate at which a bathtub empties is a function of the height of water above the drain in the tub and the tub drain’s characteristics, which include the resistance to flow and the pressure downstream of the drain. In this analogy, the inflow of water from the tap may be thought of as arterial pressure and flow, the level of water in the tub as MCFP (the elastic recoil in the system), and the drain as the venous resistance to flow and right atrial pressure. The ‘force’ of flow from the tap (arterial pressure) into the tub does not directly affect drainage beyond increasing the level of water in the tub. It is the elastic recoil of the systemic vessels (both veins and arteries) that determines flow in the systemic circulation. If the downstream pressure is the same as the pressure in the tub, the tub will not empty. Analogously, when the pressure downstream (right atrial pressure) is equal to MCFP, there is no flow; flow can occur only when MCFP is greater than right atrial pressure.’

This description of the bathtub analogy comes from a fantastic paper, by Henderson, that explains Guyton’s model much better than I can, can be found here.

So that is a, by no means comprehensive, overview of some of the cardiac physiology around fluids, but how do we apply this clinically?

Glassford and colleagues conducted a systematic review of the physiological effects of a fluid bolus in septic shock and severe sepsis.  The paper is a really interesting read and I would recommend it highly.  They found 33 papers that had looked at fluid bolus therapy.  The median bolus was 500ml (100-1000ml) and the median duration of infusion was 30mins (10-60minutes), most commonly with 0.9% sodium chloride.  The physiological results they found are summarised in this diagram:


Only 4(!) of the studies looked at the effects for 60 minutes or longer.  What I find interesting about this, as an EM trainee, is that we do not routinely measure cardiac index or CVP in patients we administer a fluid bolus to.  We often look at the BP and HR as a surrogate marker of effectiveness of fluid bolus therapy, and as we can see, the reality is that these studies show no real benefit in reducing heart rate or improving blood pressure.  It is worth noting that none of these studies were randomised controlled trials, and as a disclaimer, I haven’t read all 33 papers included in the systematic review!  This isn’t quite practice changing evidence, but it does however raise the idea that perhaps we should be performing an RCT looking at the effects of fluid bolus therapy in such patients.

So what can we do in the ED to gauge responsiveness to a fluid bolus?

The passive leg raise (PLR) is commonly cited means of assessing fluid responsiveness, as depicted below:


The PLR is a fantastic way of assessing fluid responsiveness, however it really does require a measure of cardiac output or stroke volume.  As described in a systematic review by Cavallaro et al. PLR whilst monitoring cardiac output is much more sensitive and accurate than by measuring arterial pulse pressure.  A simple measure that has been proposed is pleth variability index (PVI), which aims to detect changes in stroke volume by using pulse oximetry waveform amplitude.  This sounds great for the ED! Everyone gets a sats probe and boom there we go.  However, the majority of studies at present find that PVI is poor predictor of fluid responsiveness.

SVSo heart rate, blood pressure, PVI and PLR may not give us any idea of fluid responsiveness in the ED?!  PLR seems like the best option but we must measure this effect in real time with some form of stroke volume or cardiac output monitoring.

Ultrasound assessment of stroke volume may be an answer, and emergency physicians are certainly becoming more adept with this technique.

What we do know is that a positive fluid balance in sepsis is an independent predictor of mortality in sepsis and is an important factor in 28-mortality in patients with AKI.

I am declaring myself as somewhat of a Marik-phile, and I recommend reading this paper for a more comprehensive review of fluid responsiveness.


What next?

We need better studies to determine the most accurate way to assess fluid responsiveness in the emergency department.  I am hoping to be involved in such a study in the coming year so watch this space!

For more information, check out these awesome links from EMCrit, LITFL and CrashingPatient.





Additional reading


Before you go please don’t forget to…

Infographic to download


1 Comment

  1. Derek Louey

    My observations

    1) Sepsis demonstrates a model of distributive shock
    2) Vasoplegia (decreased vasoconstrictive responsiveness) is a key component of this
    3) The impact of fluid on MCFP and venous return on circulatory parameters is best demonstrated in a low volume state (hypovolaemic shock)
    4) Would only 1000ml of fluid be expected to impact the above parameters with distributive shock?
    5) Reversing pathophysiology (vasodilatation) with vasoconstrictors/pressors is also a solution to increase MCFP, venous return and arterial resistance (at the expense of worsening microcirculatory flow in some tissue beds)


Thanks so much for following. Viva la #FOAMed

Translate »