CARDIAC OUTPUT MONITORING

Aside from its pressure monitoring capabilities, undoubtedly the most important feature of the PAC is its ability to measure cardiac output by the thermodilution method. Cardiac output is the total blood flow generated by the heart, and in a normal adult at rest, it ranges from 4.0 to 6.5 L/min ( Table 32-14 ). To the extent that the body

TABLE 32-14 -- Normal hemodynamic values

	Average	Range
Cardiac output (L/min)	5.0	4.0–6.5
Stroke volume (mL)	75	60–90
Systemic vascular resistance (Wood units)	15	10–20
(Dynes · sec · cm-5 )	1200	800–1600
Pulmonary vascular resistance (Wood units)	1	0.5–3
(Dynes · sec · cm-5 )	80	40–180
Arterial oxygen content (mL/dL)	18	16–20
Mixed venous oxygen content (mL/dL)	14	13–15
Arteriovenous oxygen difference (mL/dL)	4	3–5
Oxygen consumption (mL/min)	225	200–250

1328

Three factors have driven efforts to measure cardiac output in clinical practice. The first is the recognition that in many critically ill patients, low cardiac output leads to significant morbidity and mortality.^{[622] [623]} Second, clinical assessment of cardiac output is often inaccurate, and in particular, many gravely ill patients with severely decreased cardiac output have normal systemic arterial blood pressure, which can be clinically misleading.^{[622] [624]} Third, newer techniques for measurement of cardiac output are less invasive than PAC monitoring and thus might provide benefit to many patients without the attendant risks of invasive monitoring.^{[624] [625] [626]}

Notwithstanding many of these considerations, the optimal cardiac output in a given clinical setting remains uncertain.^[627] As noted earlier, goal-directed therapies often target specific cardiac output values or hemodynamic values related to cardiac output, and these therapeutic approaches have not gained wide acceptance because of the inconsistent results of clinical trials. A wide variety of techniques are currently available for measuring cardiac output and other blood flow variables, and the advantages and disadvantages of each method must be appreciated for proper clinical application.

Thermodilution Cardiac Output Monitoring

The thermodilution technique has become the de facto clinical standard for measuring cardiac output because of its ease of implementation and the long clinical experience using it in various settings. It is a variant of the indicator dilution method, in which a known amount of a tracer substance is injected into the bloodstream and its concentration change measured over time at a downstream site. As its name implies, the thermodilution method uses a thermal indicator.

The fundamental physical basis for the indicator dilution methods is given by the Stewart-Hamilton equation, named after the two investigators who were instrumental in the development of this technique.^{[628] [629]} Its most generalized form is given in Equation 1.





where = cardiac output
I = amount of indicator


= integral of indicator concentration over time

Historically, various nontoxic dyes (Coomassie blue, Evans blue, indocyanine green) and radioactive tracers have been injected into both peripheral and central venous sites and their downstream concentrations analyzed, but these techniques were largely limited to cardiac catheterization laboratories and experimental studies.^{[630] [631] [632] [633]} The dye dilution methods have not maintained widespread clinical popularity because they require continuous withdrawal of arterial blood to plot the dye concentration curve, and the most common indicator, indocyanine green, can gradually build up with repeated injections. However, this method has been used to diagnose intracardiac shunts by observing the altered shape of the dye dilution curve that is produced by the abnormal recirculation of dye under these conditions.

Two developments have had a significant impact on the routine measurement of cardiac output in clinical practice. The first was the introduction of thermal indicators by Fegler,^[634] and the second was the incorporation of a temperature-measuring thermistor in the tip of a PAC. ^{[635] [636] [637]} Several unique features of the thermodilution cardiac output monitoring technique have led to its widespread popularity and rapid clinical acceptance. It can be performed quickly and repeatedly, the measurement does not require advanced diagnostic or technical skills, and it uses a nontoxic, nonaccumulating, and nonrecirculating indicator.

When a thermal indicator is used to measure cardiac output, the Stewart-Hamilton equation must be modified (Equation 2).





where V̇ = cardiac output (L/min)
T_B = blood temperature
T_I = injectate temperature
K = computational constant


= integral of temperature change over time

Blood temperature is measured continuously by the thermistor at the tip of the PAC, and injectate temperature is measured by a second thermistor at the injectate port.^[638] The computation constant (K) is provided in each PAC package insert, and it must be entered manually into the cardiac output computer before measurement of cardiac output. This constant is actually the product of three constants, V_I , K₁ , and K₂ , which are defined as follows.^{[639] [640] [641]}

V_I = volume of injectate

K₁ = density factor (product of the injectate's specific heat and specific gravity, divided by the product of the blood's specific heat and specific gravity)

K₂ = computation constant (catheter dead space, heat exchange in transit, injection time, conversion factor for units of L/min)

These constants adjust for the amount of thermal signal that will be introduced with each injection, as influenced by the physical characteristics of the catheter (i.e., size and composition) and the intended injectate that will be used. Typically, normal saline or 5% dextrose solutions are used, and the minor differences in their physical characteristics have minimal influence (<2%) on cardiac output measurement.^{[640] [642]} The last factor in the thermodilution equation,

1329

To perform a thermodilution cardiac output measurement, a fixed volume of iced or room-temperature fluid is injected as a bolus into the proximal CVP lumen of the PAC, and the resulting change in pulmonary artery blood temperature is recorded by the thermistor at the catheter tip. As in all other forms of cardiovascular monitoring, it is important to have a real-time display of the thermodilution curve resulting from each cardiac output measurement.^{[454] [643]} Such display allows the clinician to discern artifacts that would invalidate the cardiac output measurement, such as unstable blood temperature, recirculation, or incomplete indicator injection. Usually, a series of three cardiac output measurements performed in rapid succession are averaged to provide a more reliable result. Stetz and associates reviewed 14 clinical studies that compared thermodilution cardiac output with other methods, such as Fick or dye dilution, to determine the reproducibility of the thermodilution technique.^[644] These authors found that when single injections were used to determine cardiac output, a difference between sequential cardiac output measurements of 22% was required to suggest a clinically significant change. In contrast, when three injections were averaged to determine the thermodilution measurement, a change greater than 13% indicated a clinically significant change in cardiac output.

When carefully performed, thermodilution cardiac output measurements appear to provide results that are within 5% to 10% of other reference methods.^{[636] [640] [645] [646]} Few complications are directly attributable to the technique itself, although tachyarrhythmias and bradyarrhythmias related to injection of the indicator have been reported.^{[637] [647]}

Sources of Error in Thermodilution Cardiac Output Monitoring

Several important technical issues and potential sources of error must be considered for proper interpretation of thermodilution cardiac output measurements ( Table 32-15 ).^{[639] [640] [641]} The thermodilution technique measures right ventricular output and pulmonary artery blood flow. In the face of intracardiac or extracardiac shunts, right ventricular and left ventricular output will not be equal. In

TABLE 32-15 -- Factors influencing accuracy of thermodilution cardiac output measurement

Intracardiac or extracardiac shunts

Tricuspid or pulmonic valve regurgitation

Inadequate delivery of thermal indicator

Central venous injection site within the catheter introducer sheath

Warming of iced injectate

Thermistor malfunction from fibrin or clot

Pulmonary artery blood temperature fluctuations

After cardiopulmonary bypass

Rapid intravenous fluid administration

Respiratory cycle influences

patients with left-to-right shunts, early recirculation of the thermal indicator can be seen to distort the downslope of the thermodilution curve.^[641] Although mathematical solutions can be used to determine the shunt fraction based on further analysis of the distorted thermodilution curves caused by shunts, these solutions are not available on bedside cardiac output computers.^[641]

Patients with tricuspid or pulmonic valve regurgitation pose additional problems for measurement of cardiac output by the thermodilution technique because of recirculation of the indicator across the incompetent valve. In patients with severe tricuspid regurgitation, the thermodilution curves have an abnormally prolonged decay time, and the measured cardiac output may be inaccurate.^{[454] [639] [641] [648] [649]} Unfortunately, there is no simple method to correct for this problem. In the presence of significant tricuspid regurgitation, cardiac output is usually underestimated by the thermodilution method, but it may be overestimated, depending on the severity of valvular regurgitation and the magnitude of the cardiac output.^[650]

Other technical problems with thermodilution cardiac output measurement are caused by inadequate delivery of the thermal indicator, which can occur when the proximal CVP injectate port is not fully within the right atrium but rather contained within the introducer sheath.^[651] In this case, the thermal bolus will not be delivered properly, and an abnormal cardiac output curve will be produced.^{[652] [653]} This problem should be recognized when one has difficulty making the bolus injection in a patient whose PAC is inserted a relatively short distance, such as 40 cm from the insertion site. When an iced injectate is used, mishandling of syringes can warm the solution and reduce the thermal indicator administered. In addition, fibrin or blood clot on the PAC tip may lead to temperature-sensing thermistor malfunction and result in spurious cardiac output values.

Because this measurement method is a thermal-based technique, in many circumstances an unrecognized fluctuation in blood temperature may influence the cardiac output measurement. In most patients, pulmonary artery blood temperature falls rapidly in the initial minutes after cardiopulmonary bypass, when the rewarmed body core redistributes the heat gained at the end of bypass. As a result of this progressive decline in central core and pulmonary artery blood temperature, the thermal baseline is unstable. Thermodilution cardiac output measurements made in the minutes after bypass are notoriously unreliable and most often lead to marked underestimation of true cardiac output.^[654] Other inaccuracies in thermodilution cardiac output measurement occur when pulmonary artery blood temperature changes because of rapid fluid infusion through a peripheral intravenous site, through the PAC introducer sheath, or through one of the proximal PAC ports.^{[655] [656]} Either overestimation or underestimation of cardiac output occurs, depending on the timing of the additional fluid bolus.

The clinician can understand and predict all these measurement artifacts by recalling the Stewart-Hamilton equation and recognizing that cardiac output is inversely proportional to the area under the thermodilution curve. Any factor that causes a factitious reduction in this area will result in an erroneous increase in measured cardiac output. For example, if the constant entered in the

1330

The choice of iced (0°C) or room-temperature injectate has been examined in detail to determine the preferred method for accurate measurement of cardiac output. Although the size of the thermal bolus (and hence the signal-to-noise ratio) is increased when an iced injectate is used, this indicator is more time consuming, expensive, and cumbersome to prepare than a room-temperature injectate. In addition, if the injection port thermistor becomes disconnected during the use of iced injectate, the injectate temperature will be measured erroneously as room temperature and result in artifactually decreased cardiac output values because the T_B - T_I term will be too small. However, several groups have demonstrated equivalent accuracy in cardiac output determinations when iced and room-temperature injectates are compared, so it appears that room-temperature injectate is preferred for almost all clinical applications.^{[657] [658]} In adults, an injectate volume of 10 mL should be used, particularly with room-temperature solutions, because the measurements will be more accurate than those using smaller 3- or 5-mL volumes.^[658] In children, an injectate volume of 0.15 mL/kg is recommended.^[641]

One of the more controversial issues in standard bolus thermodilution cardiac output monitoring is the proper timing of measurement in relation to the respiratory cycle, particularly in patients receiving positive-pressure mechanical ventilation. By the nature of its measurement and mathematical expression, cardiac output is reported as flow per minute, even though it is measured by a thermodilution curve that is generated over several seconds and based on a small number of heartbeats. The cardiac output computer usually measures the first 60% to 70% of this curve and then extrapolates the downslope exponential to the baseline temperature.^{[639] [655]} Because the stroke output of the right ventricle varies as much as 50% during the respiratory cycle, the limited "sampling" of right ventricular stroke volumes measured with the bolus thermodilution technique may lead to widely variable measurements of cardiac output, depending on the point during the respiratory cycle when the measurements are performed.^{[659] [660]} Although the reproducibility of consecutive measurements improves markedly when the bolus injections are synchronized to the same phase of the respiratory cycle, ^{[661] [662]} accurate measurement of average cardiac output is achieved more reliably by making multiple injections during the different phases of the respiratory cycle and then averaging the results.^{[639] [641] [660] [663]} Because the pattern of variation in pulmonary blood flow during respiration depends on many factors, including the pattern of ventilation (spontaneous or positive pressure), respiratory rate, tidal volume, airway pressure, and so forth, there is no simple way to know whether timing measurements at end-expiration will always provide the highest, lowest, or an intermediate value for cardiac output.^{[639] [660] [661]} In the end, the clinician is faced with a tradeoff—accuracy versus reproducibility. Recent studies of continuous cardiac output (COO) monitoring techniques highlight these issues.