Methods of measurement of cardiac output and regional blood flow

This chapter explores the relationship of Section G7(iv) of the 2017 CICM Primary Syllabus, which expects the exam candidate to "describe the methods of measurement of cardiac output including calibration, sources of errors and limitations ​​​​​". It is also relevant to Section G7 (vi), "outline methods and principles used to measure regional blood flow", as the methods of measuring blood flow are the same irrespective of whether total or regional flow is being measured. This is a common feature of past CICM First Part Exams and should be a priority for the revising candidate, in spite of the fact that many of the techniques described here have suffered from a progressive loss of bedside popularity. Historical appearances have included:

• Question 10 from the second paper of 2017 (compare two methods)
• Question 19 from the first paper of 2014 (thermodilution alone)
• Question 12 from the first paper of 2011 (indicator dilution technique)

The most frightening potential variation of these questions would probably have to be something where the trainees have to create a table which compares and contrasts the advantages and limitations of each method. Hopefully, this tabulated summary will of use, if we ever see this happen again:

For a variety of reasons, not the least of which being their relevance to thermodilution measurements of cardiac output, the indicator dilution method and the Fick Principle are mainly discussed in the section concerning Swan-Ganz pulmonary artery catheters. Because these concepts got their own chapters, here they will only form a part of the dimly illuminated background.

As with every core topic, there is no shortage of quality peer-reviewed literature. Ehlers et al (1986) offer an excellent brief overview of the main techniques, which features a pragmatic "advantages/disadvantages" sort of breakdown. A freely available article by Lavdaniti (2008) is almost as good, just without the same structure.  Jhanji, Dawson & Pearce (2008) are another free alternative.

Cardiac output measurement by the Fick method

Put in the simplest way, the Fick method of measuring cardiac output relies on the observation that the total uptake of oxygen by the body is equal to the product of the cardiac output and the arterial-venous oxygen content difference. Logically, that principle is called the Fick principle. Rearranging the equation:

CO = VO₂ / (C_a - C_v)

where

• VO₂ is the total oxygen consumption, as a volume per unit time (eg. L/min)
• CO is the cardiac output, also as volume per unit time (L/min)
• C_a and C_v  are the arterial and venous oxygen content (eg. ml/L)

This whole thing is discussed in greater detail in a separate chapter dealing with the Fick Principle. It will suffice to say that the proper application of this method requires the cumbersome measurement of total inhaled and exhaled oxygen (usually using some sort of mask or collection bag), as well as the simultaneous measurements of arterial and mixed venous blood. Those would be the essential ingredients of the  "direct" Fick method. "Indirect" options also exist, where one of the more inconvenient measurements is replaced by some sort of estimated value, eg. where you use an age/weight/sex-based nomogram to estimate the VO₂. Obviously, the use of estimates introduces an element of error into a measurement which is already not particularly precise. The direct Fick method,  even when performed under perfect laboratory conditions, has an error range of around ±8%, according to an interesting animal study by Seely et al (1950).

Method:

• The oxygen consumption (VO₂) is measured by comparing the amount of inhaled oxygen and exhaled oxygen, usually by means of a collection bag and/or flowmeter
• The mixed venous oxygen content and the arterial oxygen content are measured directly, from the bloodstream
• Alternatively, for the indirect method, estimates can be made:
  • VO₂ can be estimated from nomograms
  • Mixed venous oxygen content can be assumed on the basis of normal values, or estimated from central venous samples
  • Arterial oxygen content can be estimated from pulse oximetry

Sources of error:

• The direct method becomes inaccurate if the cardiac output is erratic over the period during which the measurements are being collected
• The indirect method introduces a variety of inaccuracies, the magnitude and direction of which would be determined mainly by which measured value is being substituted with an estimate.

Advantages:

• This method is widely viewed as the "gold standard"
• The accuracy is acceptable for day-to-day haemodynamic management purposes
• The necessary data to calculate an indirect Fick cardiac output measurement is already available in many ICU patients (i.e. patient demographics and an arterial line)

Limitations:

• The measurement of VO₂ takes some minutes 
• The cardiac output must remain stable over the duration of measurement
• For the direct method, invasive measurements need to be taken, i.e. the patient will need arterial blood sampling and a pulmonary artery catheter
• The error range is around ±8%, as mentioned above (when compared to a flow rotameter measuring blood flow in the main pulmonary artery)

Cardiac output measurement by indicator dilution

For some reason, apparently completely unrelated to its exam importance, an entire chapter was dedicated to the indicator dilution method of cardiac output measurement. Fortunately, there is no point in reading it, as the basics are summarised here. In short, this methods rests on the premise that giving a known dose of a substance intravenously can be used to measure the cardiac output by measuring the rate of transit of that substance at some downstream detector. To be more precise, the area under the concentration/time curve can be used to determine the flow:

Cardiac output = indicator dose / area under the concentration-time curve

This is a simplification of the Stewart-Hamilton equation:

• V̇ = m/Ct, 
  
  where
  • V̇ = flow, or cardiac output
  • C = concentration
  • m = dose of the indicator, and 
  • t = time

Method

• An indicator substance is injected into the bloodstream, upstream of a detector
• The detector measures the 
• The concentration of the indicator over time is recorded as a curve
• The area under this curve is integrated to give the denominator for the cardiac output equation (V̇ = m/Ct)
• Multiple variations of this dilution method are available:
  • Thermodilution (by PA catheter or by PiCCO)
  • Lithium dilution (LiDCO)
  • Conductivity dilution using saline (the original Stewart method)
  • Indicator dye dilution (using indocyanine green or Evans Blue)

Sources of error:

• Injectate delivery technique (temperature, rate of injection, the volume of injectate, timing with respiratory cycle) plays a major role in the correct recording of measurements.
• Patient factors (eg. intracardiac shunts, valvular pathology) can disperse or dilute the injected indicator, resulting in an underestimate of the cardiac output
• The amount of injectate needs to be calibrated to the body size of the patient, i.e. a large injectate volume will overestimate the cardiac output of a small child
• Numerous correction factors are required for the thermodilution version of the equation, most of which are estimated rather than measured
• Computation of the (Ct) area can lose accuracy if the sampling rate of the detector is too low

Advantages:

• Access to mixed venous blood and arterial blood is not essential
• Numerous indicator options (cold or room temperature saline, dye, lithium, etc)
• It is convenient: with electronic calculations, thermodilution cardiac output measurement can be automated and continuous
• Good correlation with gold standard measurements of cardiac output

Limitations:

• Use of dye limits the frequency and repeatability of measurements, as it produces recirculation, and even the most rapidly cleared dyes are cleared after some minutes.
• Manual integration of the area under the concentration/time curve is laborious
• Automated calculation of cardiac output involves the use of correction factors and coefficients, which reduces its accuracy
• The method relies on uniform mixing of blood and unidirectional flow
• Thermodilution measurements have numerous potential sources of error
• Under laboratory conditions, the agreement between this method and the direct Fick method is within a margin of 25%.

Cardiac output measurement by pulse contour analysis

Cardiac output monitoring by pulse contour cardiac output monitor devices (PiCCO) is a method of continuously monitoring the cardiac output by using the shape of the arterial pressure waveform. It is also discussed in some detail by Jörn Grensemann (2018), if detail is what you are after. More than likely, it's not, in which case:

Method

• The arterial waveform is a pressure measurement, which can be converted into a volume measurement by means of a calibration factor.
• This calibration factor is derived from information about the pressure-volume relationship in the aorta, and incorporates arterial impedance, arterial compliance and systemic vascular resistance.
• These variables can be measured directly using indicator dilution measurements, or they can be estimated from nomograms based on patient demograhic data.
• The pressure./time arterial waveform can then be converted into a flow/time waveform, and the stroke volume can then be determined by integrating the area under the flow/time curve.

Sources of error:

• If the variables which are used to generate the calibration factor are measured directly, eg. by thermodilution, then they inherit all the sources of error inherent in that method of cardiac output measurement.
• If the calibration factor is estimated from nomograms, that obviously introduces an error because the nomograms may not represent the reality of any given patient.
• If the device is used for a prolonged period of time and the patient's condition has changed (specifically, the properties of the arterial vascular system), the calibration factor needs to be recalculated, otherwise the measurements will be inaccurate.

Advantages:

• Less invasive (usually, does not require mixed venous blood - only an arterial and central venous catheter )
• Convenient (you need an arterial catheter and CVC anyway)
• Continuous (the pulse contour analysis can be automated and continuous)

Limitations:

• Drifts from calibration between thermodilution measurements
• Becomes confused by atrial fibrillation, as the pulse contour becomes erratic
• Becomes confused by IABP
• Ineffective wherever flow is non-pulsatile (eg. ECMO)

Cardiac output measurement by Doppler velocity measurement

Again, for some reason LVOT Doppler measurement of cardiac output ended up getting its own (very brief) chapter, even though it has never been mentioned in any exam setting. A more detailed discussion of this technique and its limitations has been published by Huntsman et al ( 1983). In short, it rests on the assumption that the volume of blood, as it moves out of the heart during systole, can be represented mathematically as basically a cylindrical column.  The flat dimension of this column (i.e. cross-sectional area of the LV outflow tract) is obviously not perfectly circular, but it is close enough for the accuracy standards of cardiac output monitoring, and we tend to approximate it from two echo measurements of the LVOT. The column with this circular LVOT-shaped base moves in the direction of the systemic circulation with some sort of velocity. That velocity is obviously not constant, as the cardiac output is pulsatile, but that doesn't matter as long as you measure it and plot it as velocity over time. This gives you the area under the velocity-time curve, which is otherwise referred to as the velocity-time integral. Thus, the cross-sectional area of the aorta, multiplied by the distance travelled by the column of blood, gives you the volume ejected per beat; and once you have stroke volume and heart rate, you have cardiac output; or:

CO = HR × (VTI × CSA) 

where:

• CO is the cardiac output,
• HR is the heart rate,
• VTI is the velocity-time integral, i.e. the area under the velocity/time curve
• CSA is the cross-sectional area of the LVOT
  • Thus, VTI × CSA is the stroke volume

Method

• LVOT VTI is calculated by placing the pulsed Doppler sample volume in the outflow tract and recording the velocity over time.
• Usually, this is done using the "apical five chamber" view, with the sample volume placed below the aortic valve.
• The pulse wave Doppler velocity/time plot is recorded in this position, and the VTI is traced from the outermost edge of the modal velocity.

Sources of error:

• The probe must be pointing in the direction of blood flow; any angle away from this direction will result in an altered VTI and become a source of inaccuracy. Most authors (eg. Blanko, 2020) suggest that anything within 20° is good enough for government work.
• The stroke volume determined by the LVOT VTI method will vary across the respiratory cycle (by up to 10%), which means serial measurements (3-4 beats) need to be collected in order to accurately estimate the average cardiac output over a minute. Silver lining: this variability can itself be used to predict fluid responsiveness.
• Variability of the stroke volume in atrial fibrillation makes this method less accurate in AFing patients, and more beats (5-7) need to be traced and averaged to adjust for this.
• Tracing the VTI is subjective
• The method assumes laminar flow, which aortic flow is not.

Advantages:

• Non-invasive
• Easily available
• In the right hands, quite accurate (Villavicencio et al, 2019)

Limitations:

• Difficult to reproduce
• Interobserver variability
• Limited by ultrasound window availability (i.e. impossible if the patient has dressings or gas in the mediastinum)
• Accuracy dependent on beam angle 

Incidentally, there's a lot of different ways of measuring these parameters, and Doppler ultrasound is only one method, made more popular by its non-invasiveness. One could be a lot more intrusive. For example, Ehlers et al (1986) describe heated in-dwelling sensors which use heat transfer between a hot wire and the blood to determine the rate of flow, similar to the hot wire anemometers used to measure gas flow in mechanical ventilators.

Other cardiac output measurement methods

The rest of these are, for lack of a better word, niche options. One may go through their entire career in critical care without ever encountering even one of these methods, and to include them in any sort of exam would be the very peak of rudeness. They are really included here only because they are occasionally mentioned in textbooks, listed in order of most invasive to least invasive:

• Flow-measuring rotameter: this method, the most brutally stupid of all the methods mentioned thus far, is also the most accurate, and the most invasive. Invasive, in the sense that one basically has to destroy the organism while measuring its cardiac output. The method requires for the main pulmonary artery to be disconnected from the pulmonary circulation, so that all of the blood flow coming out of the right heart has to pass through the rotameter before being returned back to the systemic circulation. At least this way not a single drop of blood flow goes unmeasured, figured Seely et al (1950) who described this as the "gold standard" against which to measure the direct Fick method. Needless to say, it is only used in sacrificial lab animals.
• Electromagnetic flow probe: blood is a conductor, and when it moves through a magnetic field, a voltage is induced in it, which is proportional to its velocity. Thus, you can measure the velocity of blood by measuring this voltage, if the magnetic field strength is known. This requires an electromagnet to encircle the blood vessel, and electrodes to be in contact with the vessel wall. That's obviously going to be somewhat intrusive if the vessel of interest is the aorta. "Inappropriate in most clinical situations" is how Ehlers et al (1986) describe this method.
• Transthoracic impedance: The electrical conductivity of the chest is strongly related to the blood volume contained therein, and as the heart pumps this blood volume changes (by a volume approximately equal to the stroke volume). This is the basis of thoracic impedance measurement of cardiac output. The technique requires electrodes to be placed onto the patient, with a current of constant magnitude and high frequency to flow between them.  The impedance variation over time is recorded as a voltage signal. Apparently its shape is similar to the arterial pressure waveform. Unfortunately, this method is frustrated by pretty much everything that might also frustrate ECG monitoring (eg. artifact from patient movement). Moreover, differences in blood composition and variations in electrode positioning produce large and unpredictable errors. 
• Magnetic susceptibility plethysmography: this technique relies on the fact that magnetic field penetrates the cardiac muscle differently to cardiac blood, and so the changes in cardiac blood volume and cardiac position can be measured by a magnetometer. This device is placed on the chest, and the rest of the body is surrounded by a magnetic field. This is the least invasive technique (no sticky electrodes or high-frequency current are required) but it does involve keeping the patient inside a huge device which produces a uniform magnetic field. That's going to be difficult to explain to your unit director.