It is of the utmost importance to pay attention to the validation of cardiac output monitoring systems prior to their introduction into clinical practice. Validation studies, especially in preterm neonates, are scarce and generally include only small numbers of patients. A new technology for cardiac output measurement must be validated against a gold standard reference method that is known to be accurate and precise and does not affect the technology being tested. Ideally, this means validation against transit time flow probes, since this is considered the optimal in vivo reference method with a variability of less than 10%.4–6 The flow probe should be positioned around the pulmonary artery, since this will represent true systemic blood flow in the absence of shunts. Placing the flow probe around the ascending aorta will underestimate systemic blood flow, because coronary blood flow is not taken into account.⁷ Given the invasiveness of this reference method, its use is generally limited to animal studies. The Fick technology and the transpulmonary thermodilution (TPTD) cardiac output measurement are considered the clinical gold-standard methods in pediatric critical care.^8–10 However, these technologies are not feasible in newborn infants.

Bland-Altman analysis is the most appropriate statistical method for comparing cardiac output measurements using two different technologies.¹¹ Correlation and regression analysis are not sufficient for this purpose. With Bland-Altman analysis, the difference between the two methods (bias) is plotted against their mean. The accuracy is expressed as the mean bias, while precision is defined as the limits of agreement (LOA). The LOA can be calculated from the standard deviation (SD) of the mean bias (LOA = ±1.96 × SD). The LOA provides us with a range of the difference in cardiac output between two methods for 95% of the study population. It is recommended to express both accuracy and precision as a percentage of mean cardiac output instead of an absolute value.¹² Bias percentage (bias%) is defined as the mean bias divided by mean cardiac output multiplied by 100 (%), while the error percentage (error%) is calculated as 100% × LOA/mean cardiac output. The difference between accuracy and precision is further explained in Figure 12.1.¹³

For acceptance of a new technology, the accuracy (bias%) and precision (error%) should at least be comparable with the reference method. This stresses the importance of the use of a valid, and preferably gold-standard, technique for reference. A new technology is generally accepted when the error% is ±30% or less.¹² However, this cutoff value is based on the assumption that the precision of the reference method is ±10% to 20% with the acceptance of a new technology when the error% is no more than ±20%. When using a reference method with an error% of more than 20%, the cutoff value for acceptance of the tested technique should be adjusted.^12,13 The combined error% can be calculated using the following formula:

(12.3)errorCOMP+REF=(errorCOMP)2+(errorREF)2

The resultant error percentage as a function of the error percentage of the tested method (comparator) and the reference method, respectively, is displayed as an error-gram, as shown in Figure 12.2.¹²

In addition, the agreement in monitoring temporal changes in cardiac output (trending ability) should be studied by analysis of four-quadrant plots and polar plot methodology.^14–16 In a concordance plot the changes in cardiac output, assessed with the comparator (ΔCO_COMP), are plotted against the changes in cardiac output, assessed with the reference method (ΔCO_REF). When the changes in cardiac output are in the same direction within the two methods, these changes are concordant. Discordant measurements occur when the changes in one method are in the opposite direction of the other technology. The concordance rate is calculated as the number of concordant measurements divided by the total number of measurements (concordant plus discordant). A concordance rate of >95% is considered good, 90–95% marginal, and <90% poor trending ability.¹⁴ Polar plot analysis considers not only the direction but also the magnitude of temporal changes in cardiac output. In a polar plot each data point (ΔCO_REF vs ΔCO_COMP) is determined by an angle (the direction of change) and a radius (the magnitude of change). This enables the calculation of:

This chapter will provide an overview of all available technologies for the assessment of cardiac output, with emphasis on their applicability in (preterm) newborn infants. A summary of the characteristics, advantages, and limitations of each method is presented in Table 12.1. When available, the results of neonatal validation studies are shown in tables within the appropriate paragraphs.

Methods using the Fick principle might utilize the direct Fick method or one of its modifications to render the technique more clinically applicable. However, the modifications often come at the expense of accuracy. In 1870 German physiologist Adolf Eugen Fick stated that the volume of blood flow in a given period (cardiac output) equals the amount of a substance entering the bloodstream in the same period divided by the difference in concentrations of the substance upstream and downstream, respectively.¹⁷

Oxygen fick

Determination of cardiac output according to the original or direct Fick method requires the application of a face mask (or some means of assessing oxygen consumption) and consideration of arterial and venous oxygen concentration, which is usually obtained by taking blood samples for laboratory analysis. The direct Fick method employing a measurement of pulmonary oxygen uptake (discussed later) is considered the gold standard for assessing cardiac output, despite several disadvantages. It is of note though that recent advances in magnetic resonance imaging (MRI) technology have initiated a shift in our thinking; specifically, MRI-derived estimates of cardiac output measurement are now considered by many experts as the gold standard for the measurement of cardiac output. However, MRI-based cardiac output measurement is only feasible in stable patients who are fit enough to undergo MRI scanning (see Chapter 13). According to the direct Fick principle, cardiac output is calculated by dividing oxygen consumption (VO₂) by the difference in the oxygen content of the aortic blood (CaO₂) and the mixed venous blood (CmvO₂).

The applicability in its original form, measuring VO₂ instead of assuming it, is limited by the fact that in non-intubated patients a face mask must be used. With respect to the application in neonates, a further limitation is that multiple and multisite blood sampling is required. With oxygen being the substance for this method, the Fick principle states that during steady state, oxygen uptake in the pulmonary system equals the oxygen consumption in the tissues (Figure 12.3). Cardiac output (pulmonary blood flow) can be calculated by dividing the pulmonary oxygen uptake by the oxygen concentration gradient (difference) between arterial blood (CaO₂) and CmvO₂. Under steady-state condition, tissue oxygen consumption is equal to pulmonary oxygen uptake (VO₂). Hence

(12.4)CO = VO2CaO2−CmvO2

where CO is the cardiac output in L/min, VO₂ is pulmonary oxygen uptake in mL O₂/min, CaO₂ is oxygen concentration of arterial blood in mL O₂/L, and CmvO₂ is oxygen concentration of mixed venous blood (preferably determined in the pulmonary artery) in mL O₂/L.

Pediatric and adult patients differ in oxygen consumption. Cardiac index is 30–60% higher in neonates and infants to help meet their increased oxygen consumption. Fetal hemoglobin is present in decreasing concentration up to 3–6 months following birth, has higher oxygen affinity, and thus does not deliver oxygen to the tissues as effectively as does adult hemoglobin, when arterial oxygen saturation increases from the fetal levels of 75% in the ascending aorta to 98–100% after birth. In neonates the combination of a higher hemoglobin concentration (16–19 g/dL compared with 13.5–17.5 g/dL in men and 12–16 g/dL in women), higher blood volume per kilogram of body weight, and increased cardiac output compensate for the decreased release of oxygen from hemoglobin to the tissues.

Pulmonary oxygen uptake

Pulmonary oxygen uptake (VO₂), or oxygen consumption, can be measured via a Douglas bag, by mass spectrometry, spirometry, or metabolic monitors (indirect calorimetry).¹⁸ Table 12.2 depicts VO₂ measurements obtained in different patient populations and under different clinical conditions.^19–22 Instead of actually measuring VO₂, this can also be estimated with the use of different regression equations.^23,24 However, the estimation of VO₂, which is also referred to as the indirect Fick method, is subject to errors in the determination of cardiac output potentially exceeding 50%. Because of the potential errors and its questionable adaptability to neonates, the cardiac output obtained by the indirect Fick method may be used in neonates for orientation purposes only.

TABLE 12.2

Pulmonary Oxygen Uptake or Oxygen Consumption (VO₂) in Different Patient Populations

Group	VO2 (Mean ± SD)	Notes
Adults	125 mL O2/min/m2	Indexed for body surface area
Healthy newborns (Bauer, 2002)20	6.7 ± 0.6 to 7.1 ± 0.4 mL O2/min/kg	Indexed for body weight (n = 7)
Neonates with sepsis (Bauer, 2002)20	7.0 ± 0.3 to 8.2 ± 0.4 mL O2/min/kg	n = 10
Mechanically ventilated preterm infants (Shiao, 2006)21	8.0 ± 3.73 mL O2/min/kg	<8 h after blood drawn (n = 202)
	11.3 ± 5.65 mL O2/min/kg	≥8 h after blood drawn (n = 65)
Preterm and term infants before and 1 h after feeding (Stothers and Warner, 1979)19	4.8 mL O2/min/kg (estimated from Figure 6.1)	n = 9 preterm infants
		n = 9 term infants
Term infants during sleep22	5.97 mL O2/min/kg during REM sleep	n = 30
	5.72 mL O2/min/kg during non-REM sleep

Pulmonary oxygen uptake in adults, healthy term newborns, neonates with sepsis, mechanically ventilated preterm neonates, and preterm and term neonates before and after feeding.¹⁹–²¹ See text for details.

Oxygen concentration gradient

Oxygen concentration (cO₂) is calculated by determining hemoglobin concentration (Hb) and oxygen saturation (sO₂), which traditionally is obtained by blood gas analysis:

(12.5)cO2=(Hb × sO2 × 1.36)+(pO2 × 0.0032)

where Hb is hemoglobin in g/dL, sO₂ is oxygen saturation as gradient, 1.36 is the oxygen binding capacity of hemoglobin in mL O₂/g, 0.0032 is the solubility coefficient of oxygen in mL O₂/mmHg, and pO₂ is the partial pressure of oxygen in mmHg.

Note that the aforementioned equation obtains oxygen content (cO₂) in mL O₂/dL. To obtain oxygen concentration (cO₂) in mL O₂/L, one needs to multiply the result by 10 (1 L = 10 dL). Because dissolved oxygen contributes very little to the total oxygen-carrying capacity in the normal range of pO₂, oxygen content can be approximated by

(12.6)cO2=Hb × sO2 × 1.36

and the gradient by

(12.7)c(a−mv)O2=Hb × sO2 × (SaO2−SmvO2)

where Hb is hemoglobin in g/dL, SaO₂ is arterial (pulmonary vein) oxygen saturation as gradient, and SmvO₂ is mixed venous (pulmonary artery) oxygen saturation as gradient.

Alternative to blood gas analysis, SaO₂ and SvO₂ may be obtained via catheters (e.g., Opticath catheter in combination with Oximetric-3 monitors, Abbott Critical Care Systems, Abbott Laboratories, Abbott Park, Illinois). The limitation is, however, that not mixed venous but central venous blood is sampled, and these are not interchangeable regarding oxygen saturation or oxygen content. Arterial oxygen saturation (SaO₂) may be approximated noninvasively by SpO₂ obtained via pulse oximetry.

Cardiac output (calculation examples)

Assuming a neonate with VO₂ of 7 mL O₂/min/kg, Hb of 17 g/dL, SaO₂ of 99%, and SmvO₂ of 75%,

c(a−mv)O2=Hb × 1.36 × (SaO2−SmvO2)=17(gdl) × 1.36(mL O 2g) × (0.99−0.75)=5.55mL O2dL

With VO₂ normalized for weight, the cardiac output is

CI = VO2CaO2−CmvO2=7.0(mL O2min.kg)5.55(mLO2dL)=1.26(dLmin.kg)=0.126Lmin.kg

According to this calculation, cardiac output is 126 mL/kg/min in a neonate under physiologic circumstances. In a study by Tibby and colleagues, in five neonates with birth weights of 3.2 kg or less, the median CO following surgical correction of their congenital heart disease was 138 mL/kg/min.⁸ However, CO measured by echocardiography in neonates averages around 200 mL/kg/min. There are several reasons why our calculation suggests lower cardiac output compared with the CO assessed by echocardiography. While the Hb concentration assumed in the equation is normal for term neonates in the transitional period, it is likely too high for neonates who underwent cardiac catheterization for clinical reasons and had their CO determined using the Fick principle. In these patients oxygen consumption is also likely to be higher than that of a healthy term neonate. In addition, the methods using the direct Fick principle yield a lower CO compared with that assessed by echocardiography because of the physiologic shunting (approximately 20%) present in the lungs. Indeed, if we change the Hb concentration and the VO₂ to 15 g/dL and 8 mL O₂/min/kg, respectively, and add 20% to compensate for the physiologic shunting in the lungs, we get a CO of 196 mL/kg/min, which is the same as the estimated 200 mL/kg/min of average CO determined by echocardiography.

Carbon dioxide fick

With the CO₂-Fick method, instead of using oxygen as a marker, the exchange of carbon dioxide may be used. This principle is used in the modified carbon dioxide Fick method (mCO₂F) and the carbon dioxide rebreathing technology (CO₂R).

Modified carbon dioxide fick method

The mCO₂F is based on the principle that steady-state carbon dioxide production in the tissue is equal to pulmonary carbon dioxide exchange (VCO₂):

(12.8)CO = VCO2CvCO2−CaCO2

where CO is the cardiac output in L/min, VCO₂ is the pulmonary carbon dioxide exchange in mL CO₂/min, CaCO₂ is the carbon dioxide concentration in arterial blood in mL CO₂/min, and CvCO₂ is the carbon dioxide concentration of venous blood (preferably determined in the pulmonary artery), measured in mL CO₂/L.

Pulmonary carbon dioxide exchange (VCO₂) may be measured using volumetric capnography. In a ventilated patient VCO₂ can be determined by analysis of the expiratory airflow (Q_exp) and carbon dioxide fraction in the expiratory air (FeCO₂):

(12.9)VCO2={∫Qexp(t) · FeCO2(t) · dt} · T−1

where Q_exp is expiratory airflow in L/min, FeCO₂ is carbon dioxide fraction in expiratory air in mL CO₂/L, and T is time in minutes.

Carbon dioxide in an arterial or venous blood sample (CbCO₂) can be measured by the Douglas equation.²⁵ The amount of both pulmonary CO₂ exchange and systemic CO₂ production must be converted to standard temperature pressure, dry conditions. This method has been validated in an animal model, and the site of venous blood sampling has been shown to be of minor importance.²⁶ Interestingly, the presence of a significant left-to-right shunt through an artificial ductus arteriosus did not influence the accuracy of the mCO₂F method in another study using a juvenile animal model.²⁷ No studies in newborn infants have been published, probably related to the limitation of the need for frequent and multisite blood sampling.

Carbon dioxide rebreathing technology

In the rebreathing method mixed venous CO₂ concentration is estimated from exhaled gas and in this way obviates the need for direct measurement. The change in CO₂ exchange and resultant change in arterial CO₂ concentration secondary to an end-expiratory hold or addition of dead space is used in the CO₂-Fick equation:

(12.10)QPCBF=VCO2nCmvCO2n−CaCO2n=VCO2rCmvCO2r−CaCO2r

where Q_PCBF is the pulmonary capillary blood flow in L/min, VCO₂ is the pulmonary carbon dioxide exchange in mL CO₂/min, n denotes normal situation, r denotes rebreathing, CaCO₂ is the carbon dioxide concentration in arterial blood in mL CO₂/min, and CmvCO₂ is the carbon dioxide concentration of mixed venous blood in ml CO₂/L.

Pulmonary blood flow is considered constant during the measurements. Another assumption is that CmvCO₂ is not significantly changed throughout the period of rebreathing and nonrebreathing, implying that

(12.11)CmvCO2n≅CmvCO2r

Therefore

(12.12)QPCBF=VCO2n−VCO2r(CmvCO2n−CaCO2n)−(CmvCO2r−CaCO2r)=ΔVCO2ΔCaCO2

where ΔVCO₂ denotes the change in pulmonary CO₂ exchange and ΔCaCO₂ the change in arterial carbon dioxide concentration.

Pulmonary capillary blood flow only yields an estimate of the non-shunted blood flow participating in gas exchange. The blood bypassing the lung (shunted blood flow) may be estimated and added to Q_PCBF to determine overall cardiac output as follows²⁸:

(12.13)CO = QPCBF+QSHUNT · 

Moreover, it is not the change in CaCO₂ that is measured in clinical practice, but the change in the partial pressure of carbon dioxide (pCO₂) at the endotracheal tube, assuming that endotracheal ΔpCO₂, alveolar ΔpCO₂, and arterial ΔCaCO₂ are interchangeable. This may lead to errors in cardiac output calculation, especially in situations with large dead space ventilation.

The NICO system (Philips Respironics, Pittsburgh, Pennsylvania) is an example of a monitor incorporating the carbon dioxide rebreathing method. However, this device is not feasible in small children because of the large dead space of the rebreathing valve. In newborn infants, particularly in premature infants, partial rebreathing is contraindicated, since the technology is based on changes in arterial carbon dioxide concentrations causing significant changes in cerebral blood flow. Indeed, significant alterations in PaCO₂ have been shown to be associated with an increased risk of intraventricular hemorrhage, periventricular leukomalacia, and impaired neurodevelopmental outcome (see Chapters 1, 2, and 7).^29–31

In 1761 Haller reported a new methodology to measure pulmonary circulation time with the use of a colored dye in an animal model.³² This principle was adopted and modified initially by Stewart^33,34 and later by Hamilton,^35,36 and it’s now known as the indicator dilution technique. An indicator dilution curve can be obtained by measuring the change in time of the concentration of a known quantity of indicator that is injected proximal to the point of measurement. With the use of the Stewart-Hamilton equation, blood flow (cardiac output) can be derived from the dilution curve:

(12.14)CO = 60 · i∫C(t)dt’

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

No related posts.

Stay updated, free articles. Join our Telegram channel

Join