Chapter 13. Clinical Use of the Pulmonary Artery Catheter

• Recent randomized trials found that use of a pulmonary artery catheter (PAC) did not influence the mortality of high-risk surgical patients or critically ill patients with shock or acute respiratory distress syndrome (ARDS).
• Insertion of a PAC is associated with a low incidence of serious complications. Of potentially greater risk to the patient than insertional complications are errors in recording and interpreting hemodynamic data that lead to faulty clinical decisions. Several studies have shown that there are serious deficiencies in the understanding of basic aspects of hemodynamic monitoring among physicians and nurses who use the PAC.
• Incomplete wedging can lead to marked overestimation of the actual pulmonary artery wedge pressure (Ppw) and always should be suspected when the measured Ppw exceeds the pulmonary artery diastolic pressure (Ppad). However, with pulmonary hypertension, incomplete wedging may be present despite a positive Ppad–Ppw gradient and should be suspected when the latter markedly narrows in comparison with previous values.
• Careful inspection of the PAC waveforms may be helpful in the diagnosis of underlying cardiac disorders: Acute mitral regurgitation results in prominent v waves in the Ppw tracing, pericardial tamponade is characterized by equalization of the Ppw and right atrial pressure (Pra) and by blunting of the y descent in the atrial waveform, tricuspid regurgitation often leads to a broad cv wave and a prominent y descent, and both constrictive pericarditis and restrictive cardiomyopathy result in prominence of both the x and y descents in the atrial waveform.
• Over the range of values most often seen in the ICU, neither the Pra nor the Ppw provides a reliable indicator of the adequacy of preload and of fluid responsiveness. However, the change in Pra with a spontaneous breath may be a useful indicator of fluid responsiveness in that failure of the Pra to fall with inspiration predicts that the patient is unlikely to benefit from a fluid challenge.
• Positive end-expiratory pressure (PEEP) causes the measured Ppw to overestimate transmural left atrial pressure; the effect of PEEP on transmural pressure can be quantified by calculating the percentage of alveolar pressure that is transmitted to the pleural space during a positive-pressure breath. Active expiration also causes the measured Ppw to overestimate transmural pressure and usually results in much greater errors than does applied PEEP.
• Thermodilution cardiac output (Q̇t) can be measured by the intermittent bolus method or continuously with a modified catheter. Several noninvasive and minimally invasive methods of measuring (Q̇t) are also available.
• Venous oxygen saturation in the pulmonary artery (SvO2) or superior vena cava (ScvO2) serves as a global indicator of the adequacy of O2 delivery relative to tissue O2 demands.

Although the pulmonary artery catheter (PAC) has been in use for more than 30 years,1 its value in management of critically ill patients remains controversial. Different studies have concluded that use of the PAC is associated with increased mortality,2,3 has no effect on mortality,4,5 or decreases major morbidity.6,7 Several prospective, randomized studies have been initiated within the last few years in an effort to better define the impact of the PAC on patient outcome.8 Three trials have been completed, and none found an impact of the PAC on mortality.9–11 The appropriateness of these trials has been questioned,8,12,13 and it has been suggested that efforts instead should be directed toward improvement of the standard of practice through intensive educational efforts, institution of more stringent accreditation policies, and evaluation of newer monitoring techniques before proceeding with expensive and time-consuming randomized clinical trials.8,12,13 The latter view is based in part on evidence that ICU nurses and physicians who use the PAC may have significant deficiencies in knowledge about some of the most fundamental aspects of hemodynamic monitoring,14,15 raising concern that the value of the PAC may be difficult to assess if data are not collected and interpreted optimally.16

We believe that carefully designed prospective clinical trials may provide some guidance regarding appropriate use of the PA catheter but are unlikely to clarify with certainty whether an individual patient who is critically ill and hemodynamically unstable will benefit from the information available from a PAC. At least for the time being, ICU physicians should continue to assess the benefits and risks of catheterization on a case-by-case basis, including in the decision analysis the applicability of alternative, less invasive methods of hemodynamic assessment.17,18 We share the view that the PAC is an “occasionally useful tool” that can be of value in guiding therapy of selected critically ill patients, especially when empirical therapeutic trials have proven unsuccessful or are considered hazardous.19 Implicit in this view, however, is the understanding that hemodynamic data must be collected accurately by ICU nurses and must be interpreted by physicians who are knowledgeable about cardiopulmonary pathophysiology. Faulty clinical decisions based on inaccurate or misinterpreted data may be a greater risk to the patient than the procedure per se. This chapter reviews clinical use of the PAC in the ICU, with particular emphasis on principles of data acquisition and interpretation and on the practical application of PAC-derived data in guiding therapy. Where appropriate, comparisons between the PAC and alternative methods of hemodynamic assessment will be discussed.

There are no absolute indications for PA catheterization. However, a PAC sometimes may aid in the diagnosis and management of a number of common clinical conditions (Table 13-1). These include cardiogenic and distributive shock, severe acute respiratory distress syndrome (ARDS), pulmonary edema of uncertain etiology, oliguric renal failure, perplexing lactic acidosis, and unexplained pulmonary hypertension. Preoperative insertion of a PAC also has been advocated for patients undergoing cardiac surgery and for high-risk patients undergoing major noncardiac operations. However, one large prospective study found no benefit to routine placement of a PAC before cardiac surgery and concluded that catheterization should be deferred until there was a clear indication for invasive monitoring.20 A large prospective, randomized trial involving high-risk surgical patients found no benefit to the use of a goal-oriented hemodynamic approach based on data obtained from a PAC.10 Anticipatory placement of a PAC because a patient might become unstable would not seem to be justified in most instances. Rather, placement of a PAC usually should be reserved for circumstances in which important questions about underlying pathophysiologic derangements cannot be answered confidently by less invasive means. In this regard, it is interesting to note that the controversial study linking PAC use to increased mortality found that the risk of the PAC appeared to be greatest in less critically ill patients who had the highest likelihood of survival on entry into the ICU.2

Patients for whom a PAC might be considered often can be managed effectively with alterative methods for assessing preload, cardiac output (Q̇t), and cardiac contractility.17,18,22 The decision to proceed with invasive hemodynamic monitoring should be influenced by a variety of factors, including an assessment of procedural risk, the level of confidence in clinical assessment, and especially by the availability and expertise with alternative noninvasive techniques. For example, the etiology of unexplained hypotension may become obvious when echocardiography reveals severe depression of left ventricular contraction, acute pulmonary hypertension with right ventricular failure, hypovolemia with inspiratory collapse of the vena cava, or pericardial tamponade, potentially obviating the need for insertion of a PAC to establish a diagnosis.22 On the other hand, patients with multiple and complex hemodynamic derangements that are likely to change over time, e.g., septic shock with multiorgan dysfunction, potentially may benefit from continuous hemodynamic monitoring with a PAC.

Complications of PAC insertion include those related to achieving vascular access and those resulting from the catheter itself (Table 13-2). Only catheter-related complications will be considered here.

Both atrial and ventricular tachyarrhythmias can develop as a result of catheter insertion.23–25 Ventricular ectopy during passage through the right ventricle (RV) is not uncommon but almost always ceases as soon as the catheter tip passes through the pulmonic valve. Sustained ventricular tachycardia is unusual, and ventricular fibrillation is rare; in two large series, only 1.3% and 1.5% of patients required antiarrhythmic therapy, chest thump, or cardioversion.23,24 Given the low incidence of sustained ventricular tachycardia, prophylactic administration of lidocaine is not recommended. Ongoing ischemia, shock, hypoxemia, electrolyte disturbances, acidosis, and/or high endogenous catecholamine levels may increase the risk of ventricular tachycardia. Arrhythmogenic factors should be eliminated before catheter insertion, when possible, and time in the RV should be kept to a minimum.

Transient right bundle branch block has been reported to occur in 0.05% to 5% of catheterizations.25 Although generally of little consequence, even transient right bundle branch block is obviously of major concern if the patient already has left bundle branch block. However, a study of 82 patients with left bundle branch block found no episodes of complete heart block during PAC insertion, and the two episodes of complete heart block that occurred while the catheters were in place were ascribed to the underlying disease rather than to the catheter.26 It is not necessary to place a prophylactic transvenous pacemaker when a PAC is inserted into patients with left bundle branch block, but an external pacemaker should be at the bedside.

Clinically silent thromboses occur commonly at the site of PAC insertion, and PAC-associated clots can form occasionally in the heart or pulmonary artery.23,27 In a recent prospective, randomized study, pulmonary emboli occurred more often in patients randomized to a PAC, but the incidence of pulmonary embolism in the PAC group was only 1.6%.10 In a second large study, none of the 335 patients randomized to a PAC developed clinically evident pulmonary embolism.11 Pulmonary infarction related to peripherally placed catheters is also infrequent, with an incidence of 0% to 1.4% in a large series.25 Even when infarction does occur, it is often evidenced only by a new radiographic abnormality beyond the catheter tip without apparent clinical deterioration. In brief, clinically significant thromboembolism and infarction attributable to the PAC appear to be quite uncommon.

Pulmonary artery rupture and pseudoaneurysms can occur as a direct result of PAC-induced vascular injury. Pulmonary artery rupture, the most serious complication of pulmonary artery catheterization, usually is heralded by the abrupt onset of frank hemoptysis and carries a mortality rate of up to 50%.25 Fortunately, this complication is rare, being observed in 0.06% and 0.2% of catheterizations in two large series.23,24 Pulmonary hypertension, cardiopulmonary bypass, and anticoagulation place the patient at increased risk for morbidity and mortality from pulmonary artery rupture.25 It has been suggested that pulmonary hypertension may favor distal migration of the catheter tip on balloon deflation, permitting vascular rupture when the balloon is reinflated.28 In some cases, rupture occurs with the first reinflation. Avoidance of distal catheter placement and of balloon overinflation may reduce the risk of pulmonary artery rupture. Pseuodaneurysms can develop as a result of nonlethal pulmonary artery rupture, posing a significant risk of subsequent hemorrhage that may prove fatal.29 The possibility of a false aneurysm should be considered whenever self-limited hemoptysis occurs after manipulation of the PAC. The diagnosis can be made by dynamic computed tomographic (CT) scanning or by angiography. Prompt treatment by coil embolization may prevent subsequent aneurysm rupture.29

A number of unusual complications of pulmonary artery catheterization have been reported (see Table 13-2). However, a review of nine large series found that major complications are quite uncommon.25 A different “complication” associated with use of the PAC—erroneous recording and interpretation of hemodynamic data that results in bad clinical decisions and adverse patient outcome—may be much more frequent than these procedural risks. The remainder of this chapter will focus on the principles of data acquisition and their physiologic relevance to caring for the critically ill.

Pressure Monitoring System

Essential system components required for pressure monitoring include a fluid-filled catheter and connecting tubing, a transducer to convert the mechanical energy from the pressure wave into an electrical signal, and a signal-processing unit that conditions and amplifies this electrical signal for display (Fig. 13-1). Two primary features of the pressure monitoring system determine its dynamic response properties: natural resonant frequency and damping coefficient.30,31 Once perturbed, each catheter-transducer system tends to oscillate at a unique (natural resonant) frequency determined by the elasticity and capacitance of its deformable elements. An undamped system responds well to the low-frequency components of a complex waveform, but it exaggerates the amplitude of components near the resonant value. Modest damping is desirable for optimal fidelity and for suppression of unwanted high-frequency vibration (noise); however, excessive damping smoothes the tracing unnaturally and eliminates important frequency components of the pressure waveform (see below).

For the hydraulic monitoring system to display accurate pressures, it is essential that the system be zeroed (balanced) at the phlebostatic axis (i.e., midaxillary line, fourth intercostal space). Body angle is not crucial, so one can zero the transducer with the orthopneic patient upright or semiupright. Once the transducer has been zeroed, however, movement of the transducer relative to the heart will cause the recorded pressure to underestimate or overestimate, respectively, the true value (Fig. 13-2). Because the pulmonary circuit is a low-pressure vascular bed, small errors in transducer position may be clinically significant. The transducer converts mechanical energy from the fluid-filled tubing into an electrical signal that is then amplified and displayed. The quality and cost of transducers vary considerably. The plastic disposable transducers used in the ICU are sufficiently accurate for routine clinical purposes.32

Catheter Insertion

The standard 7F PAC includes four essential elements: (1) a distal lumen for measurement of pressure in the pulmonary artery and for sampling of blood to determine mixed venous oxygen saturation (SvO2), (2) a proximal lumen whose orifice is 30 cm from the catheter tip for measuring right atrial pressure (Pra) and infusing fluids, (3) a lumen to introduce air for balloon inflation, and (4) a thermistor at the catheter tip to enable estimation of (Q̇t) by thermodilution (see Fig. 13-1). Various catheter modifications are available, including an additional lumen for infusion of fluids, a fiberoptic system to continuously assess SvO2, a heating coil and modified thermistor to allow continuous measurement of Q̇t a rapid-response thermistor that permits measurement of right ventricular ejection fraction, and a pacing electrode.

In the ICU, the PAC usually is inserted via the internal jugular or subclavian vein. A femoral approach may require fluoroscopy, although one study reported a surprisingly high success rate of 90% with nonfluoroscopic insertion via the femoral vein.33 Before insertion, the proximal and distal lumens are connected to the appropriate pressure tubing, balloon integrity is tested, and a rapid-flush test is performed to assess the dynamic responsiveness of the pressure monitoring system (Figs. 13-3 and 13-4). Although the proximal and distal ports can be connected to separate transducers, more often a single transducer is connected to the distal port, and the proximal port is connected to a separate infusion of intravenous fluid (see Fig. 13-1). Use of a “bridge” and stopcocks permits monitoring of Pra when desired. The stopcocks should be checked before insertion to be sure that the monitor displays pressure from the distal lumen. Inadvertent recording from the proximal lumen will result in an unusually long length of catheter being inserted without achieving an RV tracing. This pitfall also should be suspected if the displayed pressure is initially near zero and then suddenly increases (proximal lumen enters the introducer) or if ventricular ectopy (tip in RV) is noted while the monitor displays a Pra waveform.

The RV should be reached within 30 to 40 cm from the internal jugular or subclavian entry sites. After entering the RV, insertion of an additional 10 to 15 cm of catheter is usually sufficient to reach the pulmonary artery. Feeding excessive catheter while the tip is in the RV should be avoided to prevent coiling and possible knotting within the RV. Entry into the pulmonary artery is reflected by an abrupt rise in diastolic pressure (see Fig. 13-4). The catheter is then advanced gradually until a pulmonary artery occlusion (wedge) pressure (Ppw) is signaled by a transition to an atrial waveform and a fall in mean pressure (Figs. 13-4 and 13-5).

A number of factors may interfere with recognition of characteristic waveforms during catheter insertion. Severe hypovolemia or other causes of decreased stroke volume reduces the pulmonary artery pulse pressure and the difference between mean pulmonary artery pressure (Ppa) and Ppw, potentially creating difficulties in determining whether a valid Ppw tracing has been achieved. With pericardial tamponade or RV infarction, the right ventricular end-diastolic pressure (RVEDP) approaches pulmonary artery diastolic pressure (Ppad), making the transition from RV to pulmonary artery sufficiently difficult to appreciate that fluoroscopy may be required to confirm catheter position.34 Large swings in intrathoracic pressure may create major problems with waveform interpretation. If the patient is mechanically ventilated, elimination of large respiratory excursions with sedation (or temporary paralysis) may aid in delineation of the tracing and will enhance reliability of the measurements obtained.35 Another problem is excessive catheter “whip” caused by “shock transients” being transmitted to the catheter during RV contraction in hyperdynamic states (Fig. 13-6). Overdamping occurs when air bubbles, clots, fibrin, or kinks diminish transmission of the pulsatile pressure waveform to the transducer, resulting in a decrease in systolic pressure and an increase in diastolic pressure. An overdamped Ppa tracing may be mistaken for a Ppw, leading to unnecessary retraction of a properly positioned catheter. A simple bedside test for overdamping is the “rapid flush” test30 (see Fig. 13-3). Because of the length and small gauge of the catheter, very high pressures are generated near the transducer when the flush device is opened. An appropriately damped system will show a rapid fall in pressure with an “overshoot” and prompt return to a crisp pulmonary artery tracing on sudden closure of the flush device. In contrast, an overdamped system will generate a tracing that demonstrates a gradual return to the baseline pressure without an overshoot (see Fig. 13-3).

PAC-Derived Pressures

The properly positioned PAC provides access to pressures from three sites: right atrium (Pra), pulmonary artery (Ppa), and pulmonary vein (Ppw). Each of these pressures will be discussed sequentially in relationship to its determinants, waveform characteristics, and factors that commonly confound its interpretation. Subsequently, clinical use of PAC-derived pressure data will be discussed.

Right Atrial Pressure (Pra)

In sinus rhythm, the Pra waveform is characterized by two major positive deflections (a and v waves) and two negative deflections (x and y descents) (see Fig. 13-5). A third positive wave, the c wave, is sometimes seen. The a (atrial) wave is due to atrial systolic contraction. The a wave is followed by the x descent as the atria undergo postsystolic relaxation and the atrioventricular junction moves downward during early ventricular systole. When visible, a c wave due to closure of the atrioventricular valves interrupts the x descent. When a c wave is seen, standard nomenclature dictates that the initial descent is termed x and the second descent is termed x′34 (see Fig. 13-5). After the x descent, the v (ventricular) wave is generated by passive filling of the atria during ventricular systole. The y descent reflects the reduction in atrial pressure as the tricuspid valve opens (see Fig. 13-5).

In order to evaluate pressure waveforms adequately, it is essential to use a two-channel recorder that allows simultaneous recording of cardiac electrical activity and pressure. An electrocardiographic lead that clearly demonstrates atrial electrical activity should be chosen. Analysis of the atrial pressure tracing begins with identification of the electrical P wave. The first positive-pressure wave to follow the P wave is the a wave. The right atrial a wave usually is seen at the beginning of the QRS complex, provided that atrioventricular conduction is normal (see Fig. 13-5). When visible, the c wave follows the a wave by an interval equal to the electrocardiographic PR interval (see Fig. 13-5). The peak of the right atrial v wave normally occurs simultaneously with the T wave of the electrocardiogram, provided that the QT interval is normal (see Fig. 13-5).

Normal Pra is approximately 2 to 8 mm Hg.34 In the absence of left ventricular (LV) dysfunction, the Pra is typically 2 to 5 mm Hg lower than the Ppw.36 However, the Ppw may be markedly higher than the Pra in patients who have either systolic or diastolic LV dysfunction.37 Conversely, the Pra may exceed the Ppw in patients with RV failure due to increased pulmonary vascular resistance (PVR) or RV infarction. In the absence of tricuspid stenosis or regurgitation, mean Pra approximates RVEDP. However, there is only a modest correlation between Pra and right ventricular end-diastolic volume (RVEDV), and the Pra required for optimal filling varies among patients.38,39 Besides being an indicator of RV filling pressure, Pra also represents the downstream pressure for venous return. Normally, the decrease in intrathoracic pressure during a spontaneous breath produces a reduction in Pra, increasing the gradient for venous return from extrathoracic veins. When the right atrium is at its limits of distensibility, however, Pra will not fall with inspiration and may even rise (Kussmaul's sign). The response of Pra to a spontaneous breath may provide insight into the adequacy of volume expansion (see below).39

Pulmonary Artery Pressure (Ppa)

The pulmonary artery waveform has a systolic pressure wave and a diastolic trough (see Fig. 13-5). A dicrotic notch due to closure of the pulmonic valve may be seen on the terminal portion of the systolic pressure wave, and the pressure at the dicrotic notch closely approximates mean Ppa.40 Like the right atrial v wave, the pulmonary artery systolic wave typically coincides with the electrical T wave (see Fig. 13-5). The pulmonary artery diastolic pressure (Ppad) is recorded as the pressure just before the beginning of the systolic pressure wave.

Ppa is determined by the volume of blood ejected into the pulmonary artery during systole, the resistance within the pulmonary vascular bed, and the downstream (left atrial) pressure. Normal values for Ppa are as follows: systolic, 15 to 30 mm Hg; diastolic, 4 to 12 mm Hg; and mean, 9 to 18 mm Hg.34 The normal pulmonary vascular network is a low-resistance circuit with enormous reserve, so large increases in cardiac output (Q̇t) do not cause pressure to rise significantly. This large capillary reserve normally offers such slight resistance to runoff during diastole that the difference between the Ppad and the Ppw (the Ppad–Ppw gradient) is 5 mm Hg or less. Increased pulmonary vascular resistance (PVR) causes the Ppad–Ppw gradient to widen, whereas an increase in left atrial pressure results in a proportional rise in the Ppad and Ppw.41,42 Therefore, the Ppad–Ppw gradient is used to differentiate pulmonary hypertension due to increased PVR from pulmonary venous hypertension.

An increased Q̇t alone will not cause pulmonary hypertension. However, in the setting of increased vascular resistance, the degree to which Ppa increases will be influenced by the Q̇t. Pulmonary hypertension may result from the combination of a modest increase in vascular resistance and a major increase in Q̇t due to sepsis, cirrhosis, agitation, fever, or other factors. The relative contributions of blood flow and the pulmonary vasculature to the increase in Ppa can be assessed by measuring Q̇t by thermodilution and calculating PVR (PVR = [Ppa − Ppw]/Q̇t). It should be appreciated, however, that interpretation of the PVR is confounded by the fact that the pulmonary vascular bed behaves like a Starling (variable) resistor; PVR increases as flow (Q̇t) decreases, and the calculated PVR must be interpreted with respect to the Q̇t at the time the measurement is made.43 For example, a fall in Q̇t owing to hemorrhage may produce a rise in calculated PVR even though the pulmonary vascular bed has not been affected directly; the PVR then may normalize once the Q̇t returns to its baseline value. Conversely, calculated PVR may decrease solely due to an increase in Q̇t. The latter may be particularly relevant when assessing the response to vasodilators that affect both the pulmonary and systemic vascular beds. With decreased Q̇t, a rise in calculated PVR that is accompanied by an increase in driving pressure within the pulmonary circuit (Ppa–Ppw) would clearly indicate active pulmonary vasoconstriction, whereas a reduction in driving pressure at increased Q̇t would provide unequivocal evidence of vasodilation.44

Pulmonary Artery Wedge Pressure (Ppw)

The Ppw tracing contains the same sequence of waves and descents as the right atrial tracing. However, when the atrial waveform is referenced to the electrocardiogram, the mechanical events arising in the left atrium will be seen later than those in the right atrium because the left atrial pressure waves must travel back through the pulmonary vasculature and a longer length of catheter. Therefore, in the Ppw tracing, the a wave usually appears after the QRS complex, and the v wave is seen after the T wave (see Fig. 13-5). Thus the systolic pressure wave in the pulmonary artery tracing precedes the v wave of the Ppw tracing when referenced to the electrocardiogram. An appreciation of the latter relationship is critical when tracings are being analyzed to ensure that balloon inflation has resulted in a transition from an arterial (Ppa) to an atrial (Ppw) waveform and to detect the presence of a “giant” v wave in the Ppw tracing (see below).

The Ppw is obtained when the inflated catheter obstructs forward flow within a branch of the pulmonary artery, creating a static column of blood between the tip of the catheter and the point (junction, or j point) in the pulmonary venous bed where it intersects with flowing blood (Fig. 13-7). Since the fully inflated catheter impacts in segmental or lobar pulmonary arteries, the j point is usually located in medium to large pulmonary veins. If the catheter were to be advanced with the balloon only partially inflated (or uninflated), obstruction to flow would occur in a much smaller artery, and the j point accordingly would move upstream to the smaller pulmonary veins. Since there is normally a resistive pressure drop across the small pulmonary veins, the Ppw recorded with a distal uninflated catheter will be slightly higher than the Ppw obtained with a fully inflated catheter.45 Owing to resistance in the small pulmonary veins, the Ppw will underestimate the pressure in the pulmonary capillaries (see below), but the absence of any appreciable resistive pressure drop across the larger pulmonary veins dictates that the Ppw will reliably reflect left atrial pressure (Pla).

For the Ppw to accurately represent Pla, it is essential that the tip of the inflated catheter lie free within the vessel lumen, and the inflated balloon must completely interrupt forward flow within the obstructed artery. Obstruction to flow at the catheter tip can lead to overwedging, whereas failure of the balloon to seal the vessel lumen results in a partial or incomplete Ppw. Overwedging is recognized by a progressive rise in pressure during balloon inflation and usually results from the balloon trapping the tip against the vessel wall. In such cases, the continuous flow from the flush system results in a steady buildup of pressure at the catheter tip, or at least as high as required to cause compensatory leakage from the trapped pocket (Fig. 13-8). If overwedging occurs, the deflated catheter should be retracted before reinflating the balloon to achieve a more suitable Ppw tracing and to prevent possible vessel injury.

An incomplete Ppw tracing occurs when the inflated balloon does not completely interrupt forward flow, resulting in a recorded pressure that is intermediate between mean Ppa and Ppw. As a result, the measured Ppw will overestimate the true value, potentially leading to serious errors in patient management. In the absence of pulmonic valve insufficiency or prominent a or v waves that increase its mean value, the Ppw should be equal to or less than the Ppad. Therefore, incomplete wedging always should be suspected if the Ppw exceeds the Ppad.46 However, incomplete wedging also can occur despite the presence of a positive Ppad–Ppw gradient. Incomplete wedging occurs often in patients with pulmonary hypertension whose increased pulmonary vascular resistance results in a marked increase in the Ppad–Ppw gradient. In this setting, the measured Ppw can increase significantly above the true value owing to incomplete wedging yet still remain less than the Ppad, giving the impression that a reliable Ppw has been obtained47 (Fig. 13-9). When this occurs, the measured Ppad–Ppw gradient will decrease in comparison with previous values. With pulmonary hypertension, incomplete wedging should be suspected when the Ppad–Ppw narrows unexpectedly or when at the time of insertion a normal Ppad–Ppw gradient is found when a widened gradient would be suspected (e.g., severe ARDS)47 (Fig. 13-10). Another clue to incomplete wedging is provided by a pressure waveform that is more consistent with Ppa than Ppw (see Fig. 13-9). Incomplete wedging can result from a catheter that is too proximal, in which case advancement of the inflated catheter may be corrective. Alternatively, a catheter that is too distal, perhaps with its tip at a vascular branch point, also can lead to incomplete wedging. This circumstance is suggested by measuring a lower (more accurate) Ppw when the balloon is only partially inflated.47 In this case, retraction of the deflated catheter before full balloon inflation may yield a more accurate Ppw and potentially reduce the risk of vessel injury due to distal catheter placement.

One method that has been suggested to confirm the reliability of the Ppw is aspiration of highly oxygenated blood from the distal lumen of the inflated catheter.48 However, there are several important considerations when using the PaO2 of aspirated blood to confirm a wedge position. First, failure to obtain highly oxygenated blood in the Ppw position could occur if the catheter tip is located in a vessel whose capillary bed supplies an area of markedly reduced alveolar ventilation.30,49 Second, it is recommended that an initial 15 to 20 mL of “dead space” blood be withdrawn and discarded before the sample for analysis is obtained so as to reduce the likelihood of obtaining a false-negative result when the inflated catheter has truly wedged.49 Finally, a false-positive result (i.e., high O2 saturation in aspirated blood when the catheter is not wedged) can occur if the sample is aspirated too quickly. It is recommended that the sample be aspirated at a rate no faster than 3 mL/min.49

Respiratory Influences on the Ppw

The Ppw is an intravascular pressure, but it is the transmural pressure (intravascular minus pleural, Ppw − Ppl) that represents the distending pressure for cardiac filling and the hydrostatic component of the Starling forces that govern transcapillary fluid movement. During normal breathing, the lung returns to its relaxed volume at end expiration, with alveolar pressure being atmospheric and Ppl being slightly negative. The Ppw therefore should be measured at end expiration, the point in the respiratory cycle when juxtacardiac (pleural) pressure can be estimated most reliably (Fig. 13-11). Either a strip recording or the cursor method should be used to record the end-expiratory Ppw because digital readouts that average over the respiratory cycle may overestimate or underestimate Ppw.

Even when strip recordings are used, there may be inaccuracies in identifying the end-expiratory Ppw.14 In one study, agreement among a group of critical care physicians who interpreted the same pressure recordings was only moderate.50 A recent study found considerable interobserver variability among ICU nurses and physicians who were asked to record the Ppw from the same strip recordings. The most important factor contributing to interobserver variability was the presence of a large amount of respiratory variation.51 Furthermore, a brief educational program did not improve the agreement among observers.52 These data highlight some of the ongoing educational deficiencies among ICU physicians and nurses regarding some of the most basic aspects of hemodynamic monitoring and support the need for more intensive instruction, as proposed in the recent report by the Pulmonary Artery Consensus Conference Organization (PACCO).53

A frequent error in identifying end expiration is the false assumption that during mechanical ventilation the lowest point in the Ppw tracing reflects end expiration. While this may be true during controlled ventilation, inspiratory efforts that trigger mechanical breaths will produce a nadir in the Ppw tracing immediately after end expiration54 (see Fig. 13-11). Identification of end expiration in the Ppw tracing is aided by the knowledge that expiration is usually longer than inspiration, two exceptions being marked tachypnea and inverse-ratio ventilation. In reality, identification of end expiration in the Ppw tracing should not be difficult so long as the respiratory influences on the Ppw are interpreted in relationship to the patient's ventilatory pattern. When confusion occurs, a simultaneous airway pressure tracing may help to identify the end of expiration.

Even when the Ppw is reliably recorded at end expiration, the measured value will overestimate the transmural pressure if intrathoracic pressure is positive at that point in the respiratory cycle. Positive juxtacardiac pressure at end expiration may result from the increase in lung volume that results from applied positive end-expiratory pressure (PEEP) or auto-PEEP or from increased intraabdominal pressure due to active expiration or the abdominal compartment syndrome. In the latter two circumstances, lung volume is not increased. Whatever the cause, increased juxtacardiac (pleural) pressure at end expiration should be taken into account when Ppw is used in clinical decision making.

Applied PEEP and Auto-PEEP

PEEP may influence the Ppw in one of two ways. First, the positive alveolar pressure conceivably could compress the pulmonary microvasculature sufficiently that there is no longer a continuous column of blood between the catheter tip and left atrium, resulting in a Ppw that reflects alveolar rather than pulmonary venous pressure. Second, even when PEEP does not interfere with the reliability of the Ppw as an indicator of pulmonary venous pressure, it nonetheless may result in increased juxtacardiac pressure so that the measured Ppw overestimates transmural pressure (Pla – Ppl).

In theory, the lung can be divided into three physiologic zones based on the relationship of pressures in the pulmonary artery (Ppa), alveolus (Palv), and pulmonary vein (Ppv)30 (Fig. 13-12). This model would predict that the catheter tip must be in zone 3 (Ppv > Palv) at end expiration for the Ppw to provide a reliable estimate of left atrial pressure (Pla), and this would not be the case if PEEP was greater than Pla. However, in the great majority of instances the Ppw will reflect Pla reliably even when PEEP exceeds the latter.55–58 Several factors may help to explain this apparent paradox. First, regardless of the values of PEEP and Pla, as long as there is a patent vascular channel between the catheter tip and the left atrium, the Ppw should reflect Pla. Since flow-directed catheters often place themselves below the level of the left atrium, local Ppv will be higher than Pla, encouraging vascular patency.59 Even when the catheter tip lies at or above the atrial level, there may still be a branch of the occluded artery extending below the left atrium that prevents the wedged catheter from recording Palv60 (Fig. 13-13). Finally, damaged lungs may not transmit Palv as fully to the capillary bed as normal lungs. In a dog model of unilateral lung injury, agreement between Ppw and Pla in the injured lung was excellent up to a PEEP of 20 cm H2O, whereas Ppw overestimated Pla in the uninjured lung at a PEEP above 10 cm H2O55 (Fig. 13-14). A clinical study involving patients with ARDS found good agreement between the Ppw and left ventricular end-diastolic pressure (LVEDP) even at a PEEP of 20 cm H2O.56 Since high levels of applied PEEP generally are restricted to patients with ARDS, the Ppw is likely to reflect Pla reliably even when high-level PEEP is required.

Although uncommon, a non-zone 3 condition should be considered when the Ppw tracing does not possess characteristics of an atrial waveform and the end-expiratory Ppw approximates PEEP. In this circumstance, a simple bedside method of ensuring a zone 3 condition may be useful.57 This technique involves a comparison of the change in pulmonary artery systolic pressure (ΔPpas) and change in Ppw (ΔPpw) during a controlled ventilator breath (Fig. 13-15). Since the former reflects the increment in Ppl during a positive-pressure breath, a ratio of ΔPpas/ΔPpw close to unity would indicate that the ventilator-induced rise in Ppw also results from change in Ppl, thereby ensuring a zone 3 condition. In contrast, ΔPpw will exceed ΔPpas if the Ppw tracing tracks the larger pressure change within the alveoli, in which case a zone 3 condition may not be present.57 Over 90% of patients with ARDS have a ΔPpas/ΔPpw close to unity (0.7 to 1.2), even at a PEEP of 20 cm H2O.57 In those few instances when Ppw tracks Palv during inspiration, a zone 3 condition could still be present at end expiration, when alveolar pressures are lowest. In brief, data from several sources strongly indicate that the end-expiratory Ppw nearly always will represent the downstream vascular pressure (Ppv, Pla) in ARDS, even when high levels of PEEP are required. Concern that the Ppw instead may be representing Palv should be limited to those rare instances in which the Ppw tracing has an unnaturally smooth appearance that is uncharacteristic of an atrial waveform, the end-expiratory Ppw is close to 80% of the applied PEEP (because mm Hg ∼ 0.8 × cm H2O), and the ΔPpw is significantly greater than ΔPpas during a controlled ventilator breath.

Even though PEEP seldom interferes with the reliability of the Ppw as a measure of Pla, it does cause the Ppw to overestimate the actual transmural pressure by increasing Ppl at end expiration (Fig. 13-16). The effect of PEEP on Ppl is determined by two factors: the PEEP-induced increase in lung volume and chest wall compliance. The degree to which lung volume increases in response to PEEP is inversely related to lung compliance.30,61 Decreased chest wall compliance (e.g., increased intraabdominal pressure or morbid obesity) enhances the fraction of PEEP transmitted to the pleural space, whereas reduced lung compliance (e.g., ARDS) may blunt PEEP transmission. One study found that the percentage of PEEP transmitted to the pleural space (as estimated with an esophageal balloon) in ARDS varied from 24% to 37%.62 However, changes in esophageal pressure may underestimate the actual changes in juxtacardiac pressure when the heart and lungs are both expanded.63,64 Thus in individual patients it may be difficult to estimate the actual juxtacardiac pressure reliably and hence the transmural pressure (Ppw − Ppl) with PEEP.

Two methods for measuring transmural pressure on PEEP have been described. The first employs a brief ventilator disconnection, during which time the Ppw falls rapidly to a nadir and then subsequently rises due to altered ventricular loading conditions. It has been shown that the nadir Ppw within 2 to 3 seconds of PEEP removal closely approximates the transmural pressure while on PEEP.65 Although this technique potentially yields a more accurate estimate of transmural pressure, it may encourage alveolar derecruitment and hypoxemia in patients with severe ARDS. In addition, the nadir method will not be reliable in patients with auto-PEEP owing to airflow obstruction because their intrathoracic pressure falls very slowly after ventilator disconnection. The second technique, which does not require ventilator disconnection, uses the transmission ratio of ΔPpw/ΔPalv during a controlled ventilator breath to calculate the percent of alveolar pressure that is transmitted to the pleural space.66 (In zone 3, ΔPpw should reflect the change in pleural pressure, and ΔPalv is defined as plateau pressure-PEEP.) The Ppl on PEEP is then estimated by multiplying PEEP by the transmission ratio, allowing calculation of transmural pressure (Ppw − Ppl)66 (Fig. 13-17A). In one study, estimates of transmural pressure using the latter technique were virtually identical to those obtained by the nadir method in patients on PEEP who did not have dynamic hyperinflation66 (see Fig. 13-17B). As noted earlier, the nadir method is unreliable for estimating transmural pressure in patients with airflow obstruction and auto-PEEP, whereas the technique involving calculation of the transmission ratio retains its validity in this patient population.66 Even though these techniques may provide valid estimates of transmural pressure, it is unclear whether they contribute positively to patient management. In clinical decision making, use of the Ppw should not focus excessively on its absolute value. Rather, change in the Ppw with therapeutic interventions and their correlation with relevant end points (e.g., blood pressure, PaO2, Q̇t, and urine output) are of greater importance, and such changes can be assessed without correcting the measured Ppw for the effects of PEEP.

Auto-PEEP may create greater difficulties in use of the Ppw than applied PEEP for several reasons. First, hemodynamically significant auto-PEEP may be occult. Second, because auto-PEEP usually occurs in the setting of chronic obstructive pulmonary disease (COPD) with normal or increased lung compliance, a larger component of the alveolar pressure may be transmitted to the juxtacardiac space. Third, the absence of parenchymal lung injury may promote non-zone 3 conditions. As noted earlier, estimates of transmural pressure based on the ΔPpw/ΔPalv ratio are more reliable than the nadir Ppw technique in patients with auto-PEEP owing to airflow obstruction.66 From a practical standpoint, the potential hemodynamic significance of auto-PEEP can be determined easily by assessing whether a 30- to 45-second interruption of positive-pressure ventilation leads to an increase in blood pressure and Q̇t67 Although this maneuver usually also results in a lower Ppw, an unchanged Ppw does not exclude the presence of hemodynamically significant auto-PEEP because a large increase in venous return could offset the reduction in juxtacardiac pressure.

Active Expiration

When the abdominal expiratory muscles remain active throughout expiration, the resulting increase in juxtacardiac pressure causes the end-expiratory Ppw to overestimate transmural pressure35,68–70 (Fig. 13-18). Although initially described in spontaneously breathing patients with COPD, this problem also occurs in the absence of obstructive lung disease and in mechanically ventilated patients.35,70 Since the pressure generated by the abdominal expiratory muscles is transmitted directly to the pleural space and is not “buffered” by the lungs, active exhalation typically causes the end-expiratory Ppw to overestimate transmural pressure to a much greater extent than does the application of PEEP.70 With active exhalation, it is common for the end-expiratory Ppw to overestimate transmural pressure by more than 10 mm Hg.35,70 Failure to appreciate the effect of active exhalation on the measured Ppw may result in inappropriate treatment of hypovolemic patients with diuretics or vasopressors on the basis of a misleadingly elevated Ppw.

When respiratory excursions in the Ppw tracing are due entirely to inspiratory muscle activity, the end-expiratory Ppw will remain unaffected. However, respiratory excursions that exceed 10 to 15 mm Hg increase the likelihood of active expiration.35 Inspection of the Ppw tracing may provide a clue to active expiration if pressure rises progressively during exhalation. However, an end-expiratory plateau in the Ppw tracing does not exclude positive intrathoracic pressure due to tonic expiratory muscle activity.69,70 Abdominal palpation may be useful in detecting muscle activity that persists throughout expiration. In mechanically ventilated patients, sedation (or even paralysis) may be used to reduce or eliminate expiratory muscle activity.35,70 In the nonintubated patient, recording the Ppw while the patient sips water through a straw sometimes helps to eliminate large respiratory fluctuations. An esophageal balloon also can be used to provide a better estimate of transmural pressure.68 In circumstances where prominent respiratory muscle activity cannot be eliminated and esophageal pressure is unavailable, transmural pressure often is better estimated by the Ppw measured midway between end inspiration and end expiration.70 However, the latter is not true in all instances,70 and it may be most appropriate simply to recognize that an accurate estimate of transmural pressure may not be possible in this situation.

Clinical Use of Pressure Measurements

There are three principal uses of PAC-derived pressures in the ICU: (1) diagnosis of cardiovascular disorders by waveform analysis, (2) diagnosis and management of pulmonary edema, and (3) evaluation of preload.

Abnormal Waveforms in Cardiac Disorders

Unfortunately, physicians and nurses sometimes focus solely on the numbers generated by the pressure monitoring system without carefully assessing the pressure waveforms. Analysis of pressure waveforms may prove valuable in the diagnosis of certain cardiovascular disorders, including mitral regurgitation, tricuspid regurgitation, RV infarction, pericardial tamponade, and limitation of cardiac filling due to constrictive pericarditis or restrictive cardiomyopathy.

Acute mitral regurgitation most often is due to papillary muscle ischemia or rupture. When the mitral valve suddenly becomes incompetent, an unaccommodated left atrium accepts blood from the left ventricle during systole, producing a prominent v wave (Fig. 13-19). A large v wave gives the Ppa tracing a bifid appearance owing to the presence of both a pulmonary artery systolic wave and the v wave (see Fig. 13-19). When the balloon is inflated, the tracing becomes monophasic as the pulmonary artery systolic wave disappears (see Fig. 13-19). A giant v wave is confirmed most reliably with the aid of a simultaneous recording of the electrocardiogram during balloon inflation. Although the pulmonary artery systolic wave and the left atrial v wave are generated simultaneously, the latter must travel back through the pulmonary vasculature to the catheter tip. Therefore, when the pressure tracing is referenced to the electrocardiogram, the v wave will be seen later in the cardiac cycle than the pulmonary artery systolic wave (see Fig. 13-19). In the presence of a giant v wave, the Ppad is lower than the mean Ppw, and the mean pressure may change only minimally on transition from Ppa to Ppw, giving the impression that the catheter has failed to wedge. This impression may lead to insertion of excess catheter, favoring distal placement that could lead to pulmonary infarction or to rupture of the artery on balloon inflation.

A large v wave leads to an increase in pulmonary capillary pressure, often with resulting pulmonary edema. When mitral insufficiency is due to intermittent ischemia of the papillary muscle, giant v waves may be quite transient (see Fig. 13-19). Failure to appreciate these intermittent giant v waves may lead to a mistaken diagnosis of noncardiogenic pulmonary edema because the Ppw will be normal between periods of ischemia.

Large v waves are not always indicative of mitral insufficiency. The size of the v wave depends on both the volume of blood entering the atrium during ventricular systole and left atrial compliance.71,72 Decreased left atrial compliance may result in a prominent v wave in the absence of mitral regurgitation. Conversely, when the left atrium is markedly dilated, severe valvular regurgitation may give rise to a trivial v wave, especially when there is coexisting hypovolemia.72 The important effect of left atrial compliance on the size of the v wave was demonstrated by a study that simultaneously evaluated the height of the v wave and the degree of regurgitation, as determined by ventriculography.72 One-third of patients who had large v waves (>10 mm Hg) had no valvular regurgitation, and a similar percentage of patients with severe valvular regurgitation had trivial v waves.72

Hypervolemia is a common cause of a prominent v wave. When the left atrium is overdistended, it operates on the steep portion of its compliance curve; i.e., small changes in volume produce large changes in pressure (Fig. 13-20). As a result, passive filling from the pulmonary veins can lead to a prominent v wave, and the latter may be quite large if Q̇t is increased. With hypervolemia or intrinsic reduction in left atrial compliance, the a wave also may be prominent, provided that the underlying rhythm is not atrial fibrillaton. Following diuresis, the a and v waves become less pronounced. Another cause of a large v wave is an acute ventricular septal defect because the increased pulmonary blood flow enhances filling of the left atrium during ventricular systole. Thus both papillary muscle rupture (or dysfunction) and acute ventricular septal defect (VSD) can be associated with prominent v waves, and these two complications of myocardial infarction usually must be differentiated by other means (see below).

Tricuspid regurgitation most often is due to chronic pulmonary hypertension with dilation of the right ventricle. With tricuspid regurgitation, there is often a characteristically broad v (or cv) wave in the central venous (right atrial) tracing (Fig. 13-21). The v wave of tricuspid regurgitation generally is less prominent than the v wave of mitral regurgitation, probably because the systemic veins have a much greater capacitance than do the pulmonary veins. One of the most consistent findings in the Pra tracing of patients with tricuspid regurgitation is a steep y descent. The latter often becomes more pronounced with inspiration (see Fig. 13-21). Kussmaul's sign, an increase in Pra with inspiration, also is observed commonly in patients with severe tricuspid regurgitation.

RV infarction may complicate an inferoposterior myocardial infarction. Common findings include hypotension with clear lung fields, Kussmaul's sign, positive hepatojugular reflux, and a Pra that equals or even exceeds the Ppw. The Pra tracing in RV infarction often reveals prominent x and y descents, and these deepen with inspiration or volume loading.34 With RV infarction, the RV and pulmonary artery pulse pressures narrow, and with RV failure, the RVEDP may approximate the Ppad (Fig. 13-22). This, together with the frequent presence of tricuspid regurgitation, may lead to difficulties in bedside insertion of the PAC, and fluoroscopy may be required.34 In the setting of a patent foramen ovale, patients with RV infarction may develop significant hypoxemia due to a right-to-left atrial shunt. Profound hypoxemia with a clear chest radiograph and refractory hypotension also would be consistent with major pulmonary embolus. The hemodynamic profiles of these two disorders are different, however, in that massive pulmonary embolism is characterized by a significant increase in the Ppad–Ppw gradient, whereas the latter is unaffected by RV infarction.41

Pericardial tamponade is characterized by an increase of intrapericardial pressure that limits cardiac filling in diastole. With advanced tamponade, intrapericardial pressure becomes the key determinant of cardiac diastolic pressures, resulting in the characteristic equalization of the Pra and Ppw. Intrapericardial pressure is a function of the amount of pericardial fluid, pericardial compliance, and total cardiac volume. The x descent is preserved in tamponade because it occurs in early systole when blood is being ejected from the heart, thereby permitting a fall in pericardial fluid pressure. In contrast, the y descent occurs during diastole when blood is being transferred from the atria to the ventricles, during which time total cardiac volume and intrapericardial pressure are unchanged. As a result, there is little (if any) change in Pra during diastole, accounting for the characteristically blunted y descent of pericardial tamponade73 (Fig. 13-23). Attention to the y descent may prove to be quite useful in the differential diagnosis of a low Q̇t with near equalizaton of pressures. An absent y descent dictates that echocardiography be performed to evaluate for possible pericardial tamponade, whereas a well-preserved y descent argues against this diagnosis.

Constrictive pericarditis and restrictive cardiomyopathy have similar hemodynamic findings. Both disorders may be associated with striking increases in Pra and Ppw due to limitation of cardiac filling. In restrictive cardiomyopathy the Ppw usually is greater than the Pra, whereas in constrictive pericarditis the right and left atria exhibit similar pressures. In contrast to pericardial tamponade, the y descent is prominent and often is deeper than the x descent. The prominent y descent is due to rapid ventricular filling during early diastole, with sharp curtailment of further filling during the later portion of diastole (Fig. 13-24).

Diagnosis and Management of Pulmonary Edema

The Ppw is used commonly to aid in the differentiation of cardiogenic and noncardiogenic pulmonary edema. For uninjured lungs, the expected Ppw threshold for hydrostatic pulmonary edema is approximately 22 to 25 mm Hg. (A higher threshold is common if the Ppw has been chronically elevated.) When capillary permeability is increased, however, pulmonary edema occurs at a much lower Ppw. Indeed, one generally accepted criterion for ARDS has been a Ppw <18 mm Hg. It is important to appreciate, however, that an isolated Ppw reading does not reliably predict whether pulmonary edema occurred on the basis of increased capillary pressure (Pcap) alone or on the basis of altered permeability, especially when recorded after a therapeutic intervention. Acute hydrostatic pulmonary edema occurs frequently despite normal intravascular volume on the basis of an acute decrease in LV compliance resulting from ischemia or accelerated hypertension. By the time a PAC is placed, the acute process often has resolved, resulting in a normal or even reduced Ppw, depending in part on what type of therapy (e.g., diuretics or vasodilators) has been given. In this circumstance, maintaining the Ppw ≤18 mm Hg over the next 24 hours should lead to marked clinical and roentgenographic improvement if pulmonary edema had been due to elevated Pcap prior to catheter insertion. Conversely, lack of improvement or worsening would suggest altered permeability as the etiology of pulmonary edema. One must be careful, however, when hydrostatic pulmonary edema is due to intermittent elevations in Ppw due to myocardial ischemia. Transient ischemia-related elevations in Ppw may be missed by intermittently recording Ppw (see Fig. 13-19), potentially leading to an erroneous diagnosis of ARDS. Some bedside monitors store pressure data from the previous 12 to 24 hours, and inspection of a graphic display of the stored data may be useful in detecting transient elevations in Ppa that occur during periods of intermittent ischemia. Just as patients with hydostatic pulmonary edema may have a normal Ppw, patients whose pulmonary edema is due primarily to increased permeabililty may have an increased Ppw due to volume expansion.74 In brief, the pathogenesis of pulmonary edema formation should not be based solely on Ppw, and clarification of the underlying mechanism may require a period of careful clinical and radiologic observation.

Ppw, the pressure in a large pulmonary vein, represents a very low-end estimate of the average pressure across the fluid-permeable vascular bed. Normally, about 40% of the resistance across the pulmonary vascular bed resides in the small veins75 (Fig. 13-25). When pulmonary arterial and venous resistances are normally distributed, the Gaar equation predicts Pcap by the formula Pcap = Ppw + 0.4(Ppa − Ppw).75 Since the driving pressure (Ppa − Ppw) across the vascular bed is normally very low, Pcap will be only a few millimeters of mercury above Ppw. However, a significant pressure drop from Pcap to Ppw could occur under conditions of increased venous resistance, increased Q̇t, or both. For example, the markedly increased venous resistance of pulmonary veno-occlusive disease results in clinical evidence of increased Pcap (e.g., pulmonary edema, Kerley B lines) despite a normal Ppw.76 Other clinical conditions that selectively increase venous resistance are not well defined.

A number of techniques for determining Pcap have been described.77–79 The transition from a Ppa to a Ppw waveform after balloon occlusion includes an initial rapid decay and a subsequent slower decay (see Fig. 13-25). In experimental animals, the inflection point between the rapid and slow components has been shown to represent Pcap, as measured by isogravimetric or simultaneous double-occlusion (arterial and venous) techniques.77 Estimates of Pcap from visual inspection of pressure tracings after balloon occlusion has been used in humans.78,79 One study concluded that Pcap was on average 7 mm Hg higher than the measured Ppw in patients with ARDS.78 In this study, the estimated Pcap and the calculated Pcap by the Gaar equation were highly correlated, implying that arterial and venous resistances are increased equally in ARDS.78 It should be appreciated, however, that it may be difficult to determine Pcap confidently by inspection of the pressure tracing following balloon occlusion.80 Furthermore, even if an accurate estimate of Pcap can be obtained, it is unclear how this would have any practical advantage over the Ppw in guiding fluid management. The important point is that Ppw is a low-range estimate of Pcap; the true value of the latter lies somewhere between Ppa and Ppw. It follows that increases in the driving pressure across the microvasculature caused by increases in Q̇t have the potential to exacerbate edema formation.

Downward manipulation of Ppw by diuresis or ultrafiltration will reduce Pcap and may benefit gas exchange markedly in patients with ARDS. There is no minimum value for Ppw below which removal of intravascular volume is contraindicated, provided that Q̇t is adequate. If the clinical problem is severely impaired gas exchange requiring high FiO2 or high PEEP, then a trial of Ppw reduction is reasonable as long as Q̇t and blood pressure remain within acceptable limits. As with all therapeutic manipulations, clinically relevant end points (e.g., PaO2, blood pressure, and Q̇t should be assessed before and after Ppw reduction.

Assessment of Preload

When afterload and intrinsic contractility are held constant, the forcefulness of ventricular contraction is determined by end-diastolic fiber length (preload).81 The most commonly used indicators of preload are Ppw and Pra.82 Indeed, one of the principal reasons for developing the PAC was to have a bedside method of assessing LV preload.1 In order to assess preload reliably, the Ppw must accurately reflect LVEDP, and LVEDP must correlate well with left ventricular end-diastolic volume (LVEDV). Under most circumstances, the Ppw provides a close approximation of LVEDP. Exceptions include an overestimation of LVEDP by the mean Ppw with mitral stenosis or mitral regurgitation with a very large v wave and underestimation of LVEDP by the mean Ppw when diastolic dysfunction or hypervolemia causes the LVEDP to increase markedly with atrial systole (“atrial kick”).30 (With a large v wave, LVEDP is best estimated by the pressure just before the onset of the v wave; with a prominent a wave, LVEDP is best estimated by the pressure at the z point, just after the peak of the a wave.)81 Unfortunately, even though the mean Ppw is usually equivalent to LVEDP, factors that alter LV compliance (e.g., hypertrophy or ischemia) or change juxtacardiac pressure (e.g., PEEP or active exhalation) may profoundly influence the relationship between LVEDP and LVEDV (Fig. 13-26A). It is not surprising, therefore, that among different patients, an equivalent LVEDV may be associated with widely varying Ppw83 (see Fig. 13-26B).

The optimal Ppw (for preload) can be defined as the Ppw above which there is minimal increase in stroke volume. In normal individuals, optimal Ppw is often 10 to 12 mm Hg.84 During resuscitation from hypovolemic or septic shock, the optimal Ppw is usually ≤14 mm Hg,85 whereas in acute myocardial infarction it is often between 14 and 18 mm Hg.86 However, these target values certainly are not valid in all cases. By measuring stroke volume at different Ppw values, a cardiac function curve can be constructed, thereby defining optimal Ppw for an individual patient. This may be particularly useful in patients who also have established or incipient ARDS because a Ppw above the optimal value will increase the risk of worsening oxygenation without offering any benefit with regard to Q̇t. It should be appreciated, however, that the relationship between Ppw and Q̇t may change as a consequence of alterations in LV compliance, myocardial contractility, or juxtacardiac pressure and therefore may need to be redefined if clinical status changes.

A clinically relevant test of the utility of the Ppw is its ability to predict the hemodynamic response to a fluid challenge when hypotension, oliguria, or tachycardia leads to uncertainty about the adequacy of preload. Studies that have examined the utility of the Ppw in predicting fluid responsiveness have been reviewed recently.18 Seven of nine studies found that the Ppw was no different in fluid responders and nonresponders, and analysis by receiver operating characteristic (ROC) curves in two studies indicated that the Ppw was not particularly useful as a predictor of fluid responsiveness.87,88 One study did find a significant inverse relationship between Ppw and fluid-induced change in stroke volume, but the degree of correlation was only moderate.89 Although these data suggest a major limitation of the Ppw as an indicator of preload, it is clear that there must be a Ppw above which volume expansion almost always would be futile and a Ppw below which a positive response to fluid virtually is certain. To define these cutoff Ppw values confidently, however, would require a large study in which individual values for Ppw and fluid-induced change in stroke volume are reported for each patient, with a wide range of Ppw values being examined. Individual values for Ppw and fluid-induced change in stroke volume were reported in two studies.89,90 Although no patient with a Ppw >18 mm Hg had a positive response to fluid and all patients with a Ppw <8 mm Hg were fluid responders, the number of patients within these Ppw domains was too small to draw any firm conclusions about the usefulness of these cutoff values.89,90 Furthermore, in the great majority of patients who had a Ppw between 8 and 18 mm Hg there was no apparent relationship between Ppw and fluid response. In brief, while the Ppw is used often in clinical practice as an indicator of the adequacy of preload, most of the evidence to date would suggest that it has limited utility in predicting fluid responsiveness, at least over the range of values encountered most often in the ICU.

Overall, the data for Pra were quite similar to those described for the Ppw, in that three of five studies found no difference between the Pra values of responders and nonresponders,18 and in one of the studies that did find a difference, there was only a modest inverse correlation between Pra and the fluid-induced change in stroke volume.89 Although there does not appear to be an appreciable difference in the predictive value of the Pra and the Ppw with regard to fluid response for most patients, it might be anticipated that the Pra would be inferior to the Ppw in a population of patients with severe isolated LV dysfunction related to acute myocardial infarction or other causes.37 Conversely, in patients with severe RV dysfunction and preserved LV function, the Pra may be more relevant than the Ppw.91

Although Ppw and Pra are used most widely in guiding fluid therapy in the ICU,18,82 measurements of cardiac volumes also have been used to predict fluid responsiveness.89,92–94 A modified PAC with a rapid-response thermistor permits simultaneous measurement of RV ejection fraction (RVEF) and stroke volume, from which RV end-diastolic volume (RVEDV) can be calculated. Several studies have compared the Ppw and RVEDV as predictors of fluid responsiveness.89,92,93 One study found the RVEDV to be superior to the Ppw and suggested that either a positive or negative response to fluid could be predicted reliably when the RVEDV index was less than 90 or greater than 138 mL/m2, respectively.93 However, a subsequent study found that these threshold values for RVEDV index were unreliable and that both the Ppw and Pra predicted the response to fluid better than RVEDV.89 It is not clear that the RVEDV is any better than the Ppw (or Pra) at predicting fluid responsiveness.

Left ventricular end-diastolic area (LVEDA), measured by transesophageal echocardiography, is another potential indicator of fluid responsiveness. Two studies found that LVEDA was significantly lower in responders than in nonresponders and that LVEDA was superior to Ppw as a predictor of fluid responsiveness.87,90 However, there was considerable overlap in the LVEDA values between groups, and analysis using ROC curves demonstrated minimal utility of LVEDA in predicting the fluid response of individual patients.87 Another study found that the Ppw was a better predictor of fluid response than LVEDA during cardiac surgery.95 In addition, a recent study of patients in septic shock also found the LVEDA to be of minimal value in predicting the response to volume expansion.96

In contrast to the static indicators of preload mentioned earlier, methods that rely on the dynamic response to respiratory changes in intrathoracic pressure have performed somewhat better at predicting fluid response. One method is based on the presence or absence of a reduction in Pra during a spontaneous breath.39 When Pra decreased with inspiration, a positive response to fluid was likely (although not inevitable). Conversely, when a spontaneous breath did not produce a fall in Pra, Q̇t remained unchanged or decreased after a fluid challenge.39 In this study, neither the Ppw nor the Pra discriminated responders from nonresponders.39 A follow-up study confirmed that patients with an inspiratory decrease in Pra had a much greater probability of responding to fluid than did patients whose Pra did not change with inspiration.97 A second method is based on the change in arterial pressure from inspiration to expiration during controlled positive-pressure ventilation.88 In a study involving septic patients with circulatory failure, the respiratory variation in arterial pulse pressure was much greater in responders than in nonresponders, and a threshold value was found that discriminated these two groups with a high degree of accuracy.88 A second study that used a modification of this technique also found the respiratory variation in arterial pressure to predict the response to fluid much better than the Ppw or LVEDA in patients with sepsis.87 Although more investigations are needed to confirm these studies, methods that rely on the dynamic response to respiratory changes in intrathoracic pressure ultimately may prove to be better indicators of the adequacy of preload than static indicators such as the Ppw, Pra, RVEDV, and LVEDA.18

Principles of Measurement

The thermodilution technique for measuring Q̇t is an indicator-dilution method in which the indicator is thermal depression (cold). Cold fluid is injected through the proximal lumen of the PAC and, after mixing thoroughly in the RV with venous blood returning from the periphery, passes into the pulmonary artery, where a thermistor at the tip of the catheter senses dynamic changes in temperature. The Stewart-Hamilton formula relates Q̇t to temperature change over time:

where V is injected volume, TB is blood temperature, TI is injectate temperature, ∫TB(t)dt is change in blood temperature as a function of time, and K1 and K2 are computational constants. The numerator contains known constants or measured values, whereas the denominator is the area beneath the time-temperature curve derived by computer integration of the thermistor signal. Although it would seem that increasing the strength of the signal by cooling the injectate should enhance the accuracy and reproducibility of thermodilution (Q̇t) (Q̇ttd), studies have shown that iced and roomtemperature injectates usually are similar in effect as long as careful attention to technique is employed.98

Errors may occur if the wrong computational constant is entered into the computer or if either the volume or the recorded temperature of the injectate is incorrect. In addition, the temperature-time curve should be inspected for irregularities. A valid curve shows a rapid early ascent to a peak value and then smoothly returns to the thermal baseline within 10 to 15 seconds of injection (Fig. 13-27). Inadequate blending of injectate with blood, prolonged (>4 s) or uneven injection, thermistor contact with the vessel wall, or abrupt changes in heart rate may result in a grossly irregular curve, and the data should be discarded. The measured Q̇ttd may vary significantly depending on the phase of the respiratory cycle at which the injection is begun.99 Timing of injection may be less important when the respiratory rate is high because the duration of injection will span the entire respiratory cycle. The average of three to five randomly spaced injections probably gives the best overall measure of (Q̇t).

Certain conditions can render Q̇ttd an unreliable estimate of net systemic Q̇t. Since individual measurements are made over a few cardiac cycles, major irregularities in cardiac rhythm during the period of injection could affect the extent to which Q̇ttd reflects average Q̇t over a longer period. Left-to-right shunts (e.g., atrial septal defect or ventricular septal defect) increase the ratio of pulmonary to systemic Q̇t causing Q̇ttd to seriously overestimate systemic Q̇t. The effect of tricuspid regurgitation on the reliability of Q̇ttd is controversial. In two studies, Q̇ttd significantly underestimated Q̇t measured by the Fick method or continuous-flow Doppler in patients with severe tricuspid regurgitation.100,101 In contrast, other investigators found that tricuspid regurgitation had no discernible impact on the agreement between simultaneous measurements of Q̇ttd and Fick-derived Q̇t.102,103 Given the uncertainty regarding accuracy of Q̇ttd in individual patients with significant tricuspid regurgitation, it may be advisable to measure Q̇t with an alternative method (see below) when significant tricuspid regurgitation is diagnosed by inspection of the Pra tracing (see above), by echocardiography, or by a characteristically attenuated time-temperature thermodilution curve (see Fig. 13-27).

Continuous monitoring of Q̇ttd is possible with a modified PAC that contains a 10-cm-long thermal filament that transfers heat directly to the blood using a pseudorandom binary sequence.104 The amount of heat applied is safe, and the upper limit of temperature attained within the filament is 44°C. Temperature change is detected at the catheter tip and is cross-correlated with the input sequence to produce a thermodilution washout curve. The displayed Q̇t value is an average of the last 10 separate determinations over the previous 6 to 8 minutes.104 Several studies have compared measurements of Q̇ttd by the continuous and bolus injection methods.104–106 Reported agreement between methods appears good, with observed differences of little clinical significance. An in vitro study in which flow was controlled found both the bolus and continuous thermodilution methods to be reliable, but the continuous method was more accurate when flow was very low (<3 L/min).106 A recent study of patients with a left ventricular assistance device that allowed an independent intravascular measurement of Q̇t also found the continuous cardiac output to be reasonably accurate.107 Although the clinical impact of continuous Q̇ttd monitoring is uncertain, it could have potential utility when Q̇t is likely to undergo marked fluctuations either spontaneously or in response to therapy. From a practical standpoint, it may be more convenient to use the continuous thermodilution method when frequent assessments of Q̇t are desired during titration of inotropic or vasodilator agents or with incremental application of PEEP.

Alternative Methods of Measuring Q̇T

Several noninvasive or minimally invasive alternatives to the thermodilution technique for measuring Q̇t have been introduced to clinical practice recently. These methods include pulse contour analysis of the arterial pressure waveform, esophageal Doppler analysis of aortic blood flow, thoracic electrical bioimpedance, the indirect Fick method employing partial carbon dioxide rebreathing, and a transpulmonary thermodilution technique.17 Because all methods have limitations, there is no “ gold standard” for measuring Q̇t. However, for logistic reasons, assessment of these less invasive methods generally has been done by comparison with the PAC-based bolus thermodilution technique as the reference technique. The partial CO2 rebreathing technique, a method applicable to intubated patients, is entirely noninvasive and has shown good agreement with the bolus thermodilution method in some studies. However, although the issue currently remains unsettled, there has been some concern regarding accuracy of this technique in the setting of severe lung injury (where CO2 shunting occurs) or when there is variability in tidal volume due to spontaneous breaths, potentially limiting its general applicability in the ICU.17,108 Esophageal Doppler measurements of Q̇t generally have shown good concordance with those obtained by bolus thermodilution, including investigations performed in critically ill patients with sepsis.109,110 To achieve reliable results with this method, it is important that nurses who perform the procedure receive sufficient training to achieve proficiency.111 Thoracic electrical bioimpedance has been evaluated in numerous studies. A meta-analysis of over 150 studies concluded that while it may be useful for trend analysis, this method is probably not reliable enough for diagnostic purposes.112 Bioimpedance methodology may be less accurate when there is significant pulmonary or peripheral edema or in the setting of large pleural effusions, possibly limiting its utility in the ICU.17 A recent study suggested that advances in technology have made the current systems more reliable than earlier versions.113 In contrast to the standard bolus thermodilution technique using a PAC, which measures right-sided Q̇t, the transpulmonary thermodilution technique measures left-sided Q̇t by injection of a cold injectate intravenously and detection of the change in temperature in the arterial system.17,114,115 The agreement between these two methods has been reasonably good, with the transpulmonary technique values being somewhat higher, possibly due to loss of indicator (cold) in the lung.17 Pulse contour analysis of the arterial pressure waveform to measure stroke volume was developed 20 years ago.17 Most studies comparing the pulse contour and bolus thermodilution methods of measuring Q̇t have shown excellent correlation.17,116 The standard pulse contour device must be calibrated with a Q̇t value measured independently, such as by the transpulmonary thrermodilution technique. An alternative method of calibration is an indicator dilution technique in which lithium chloride is the indicator.117 The lithium pulse contour method has been shown to correlate very well with the bolus thermodilution technique, and in one study the coefficient of variation for multiple measurements was significantly less for the former method.118

Clinical Use of Q̇Ttd

Combining Q̇ttd with measurement of systemic and pulmonary pressures allows calculation of vascular resistances: systemic vascular resistance (SVR) = (Pa − Pra)/Q̇t and pulmonary vascular resistance (PVR) = (Ppa − Ppw)/Q̇t. To assess the appropriateness of Q̇t to body mass, Q̇ttd is divided by body surface area (BSA) to calculate cardiac index. Unlike flow (Q̇t), pulmonary and systemic pressures are not dependent on body size. Therefore, to avoid misinterpretation due to variation in body mass, it is also appropriate to compute indices of systemic and pulmonary vascular resistance by using cardiac index rather than Q̇t in the resistance calculations.

Assessment of Q̇ttd and SVR may be particularly helpful in the assessment of hypotension, oliguria, and unexplained lactic acidosis. Measurement of Q̇ttd is of particular value in the management of septic shock and ARDS. Septic hypotension often is due primarily to excessive arterial vasodilation, but decreased venous return and myocardial dysfunction may contribute to the hemodynamic derangement. By assessing the relative contribution of Q̇t and SVR to systemic hypotension, the clinician may be able to apply fluid, vasopressor, and inotropic therapy more rationally. In ARDS, a fall in blood pressure could result from hypovolemia, LV dysfunction, pneumothorax, pulmonary emboli, increasing PVR from ARDS, a change in applied PEEP, or dynamic hyperinflation, all of which produce hypotension through a decrease in Q̇t. Evaluation would consist of measurement of intravascular and airway pressures, perhaps along with obtaining a chest x-ray, electrocardiogram, or echocardiogram. On the other hand, these conditions would be of little concern if it were known that Q̇t remained unchanged, indicating that the fall in blood pressure was due entirely to a reduction in SVR.

Sepsis is the most common cause of hypotension with a low SVR, but other conditions may produce a similar hemodynamic profile. Severe hypotension resulting from excessive arterial vasodilation can occur as a result of acute adrenal insufficiency, thiamine deficiency (beriberi), severe pancreatitis, or poisoning with a variety of drugs or toxins (e.g., nitrates, calcium channel blockers, aspirin, and tricyclic antidepressants). Systemic arteriovenous fistulas, as occur with cirrhosis, also lead to a high Q̇t–low SVR hemodynamic profile, albeit usually with normal blood pressure. Failure to include noninfectious processes in the differential diagnosis of a high Q̇t–low SVR state may result in misdiagnosis of life-threatening, treatable conditions.119

The hemodynamic response to a fluid challenge, inotropes, vasodilators, or alterations in PEEP is often assessed in part by measuring changes in Q̇ttd. Critically ill patients may undergo significant spontaneous fluctuations in tissue O2 demand and Q̇t throughout the day. Therefore, determinations of Q̇ttd ideally should be made immediately before and after a specific intervention. Even when assessment is optimal, the normal variability in measured Q̇ttd may confound the response to an intervention. A study involving hemodynamically stable patients found that measured Q̇ttd varied by an average of 7.7% over a period of 1 hour.120 Therefore, the clinician must observe a change in measured Q̇ttd of at least 10% to 15% before concluding that the observed response was actually due to a specific intervention.120,121

Mixed venous oxygen saturation (SvO2) is the O2 saturation of blood in the pulmonary artery after the venous effluent from various organs has been mixed thoroughly in the RV. The SvO2 can be measured intermittently by slowly withdrawing a sample of blood from the distal port of the unwedged PAC or continuously with a fiberoptic PAC that measures O2 saturation by reflectance oximetry.122 Intermittent sampling of SvO2 is accomplished by discarding the initial 3 mL of blood and then withdrawing a sample very slowly so as to avoid contamination with capillary blood. The SvO2 should be measured by co-oximetry because the steep slope of the O2 dissociation curve in the venous range means that small errors in measurement of PvO2 may result in substantial errors in calculation of SvO2.

Ordinarily, the O2 saturation is lower in the superior vena cava than in the inferior vena cava, but the reverse is true in shock due to redistribution of blood flow away from splanchnic, renal, and mesenteric beds.123,124 Thus the oxygen saturation of blood obtained from a central venous catheter in the superior vena cava (ScvO2) will overestimate SvO2 in shock. Although the absolute values of SvO2 and ScvO2 may differ, they tend to track together with changing hemodynamic conditions.124 Furthermore, it has been argued that a low ScvO2 in the setting of shock may be clinically relevant because it implies an even lower SvO2.124

Oxygen delivery is the product of Q̇t and arterial O2 content, the latter being determined by the hemoglobin and arterial O2 saturation (SaO2). The body's O2 consumption (VO2) is determined by the underlying metabolic activity of tissues and normally is independent of O2 delivery.125 Thus, as O2 delivery falls, so too will SvO2 (Fig. 13-28). Therefore, measurement of SvO2 helps to assess the adequacy of O2 delivery in relationship to tissue O2 requirements. Under normal conditions, the SaO2 − SvO2 difference is 20% to 25%, yielding an SvO2 of 65% to 75% when arterial blood is well oxygenated. An SvO2 in the normal range may be associated with very different levels of Q̇t depending on underlying aerobic metabolism. For example, a “low” Q̇t of 3 L/min may be entirely appropriate if metabolic activity is reduced, in which case the SvO2 would be normal. However, the same Q̇t in association with a decrease in SvO2 would suggest inadequacy of either intravascular volume or cardiac function. Similarly, a low SvO2 despite a normal (or increased) Q̇t indicates that the latter is inadequate to meet increased metabolic demands for O2. In essence, the SvO2 helps to define the appropriateness of measured Q̇t.

Increased SvO2 may be seen in a variety of conditions. In sepsis, SvO2 is often normal, but in some cases there is extreme peripheral vasodilation, and Q̇t increases disproportionately to metabolic demands, resulting in an increase in SvO2. Cirrhosis is one of the more common causes of a markedly increased SvO2, with values occasionally in excess of 85%. A number of vasodilating agents and left-to-right cardiac shunts also result in an increase in SvO2. Lastly, agents that interfere with mitochondial cytochrome activity (e.g., cyanide) may produce striking elevations in SvO2 owing to the inability of tissues to use O2.

Continuous measurement of SvO2 with a fiberoptic PAC has been advocated as a useful technique for early detection of inadequate Q̇t relative to tissue O2 demand. If hemoglobin and SaO2 remain unchanged, then a reduction in SvO2 would predict either a primary reduction in Q̇t or a rise in tissue O2 requirement that is unmet by a compensatory increase in Q̇t. Theoretically, continuous SvO2 monitoring could lead to more rapid detection of hemodynamic instability than simple monitoring of blood pressure and heart rate, but it is unclear whether it has a significant impact on patient care.126–128 Perhaps the most rational use of the fiberoptic PAC is in the management of patients with cardiac dysfunction, in whom rapid changes in Q̇t may be expected to occur on the basis of the underlying disease process or in response to pharmacologic interventions. This would be particularly true when measurement of Q̇ttd has been documented to be erroneous in the setting of significant tricuspid regurgitation or as a result of markedly variable heart rhythm. In the latter circumstances, SvO2 assumes primary importance as an indicator of Q̇t, and continuous monitoring might prove more convenient than intermittent sampling from the catheter's distal lumen.

Clinical Use of Sv o2

Use of the SvO2 may be a pivotal indicator of the adequacy of hemodynamic support. A low SvO2 in the absence of severe hypoxemia often indicates a deficiency of oxygen delivery relative to tissue O2 needs and the need for augmentation of Q̇t with volume loading or inotropic support or with red blood cell transfusion if anemia is present.129 In contrast, a normal or high SvO2 does not necessarily indicate adequacy of oxygen delivery (as in sepsis), but the need to further augment Q̇t in this setting is controversial.129 A number of studies have evaluated the use of SvO2 in critical illness, but few have addressed important end points such as mortality or major morbidity. One randomized study reported decreased morbidity after cardiac surgery when therapy was targeted to achieve an SvO2 greater than 70% and a lactate level of less than 2 mg/dL.130 In contrast, another study involving a more heterogeneous group of critically ill patients in the ICU found no clear benefit to goal-directed hemodynamic therapy based on the SvO2.131 However, it has been argued that the potential benefits of SvO2-based hemodynamic resuscitation may be greater during the early management of critically ill patients.124 While placement of a PAC to measure SvO2 is impractical during initial resuscitation, a recent study evaluated the use of vena caval O2 saturation (ScvO2) obtained from central venous catheters placed in the emergency department in the early management of patients with severe sepsis.132 Patients whose resuscitation was targeted to achieve an ScvO2 of greater than 70% had significantly lower mortality than those randomized to conventional therapy based on mean arterial pressure and central venous pressure.132

Although the PAC is used primarily to aid in the management of fluids and vasoactive drug therapy during critical illness, it has additional applications that may be useful occasionally. These include microvascular cytology,133 evaluation of arrhythmias,34 detection of intracardiac shunts,34 and determination of caloric requirements.134,135

Microvascular cytology of blood drawn through the distal port of the wedged PAC has been used to diagnose lymphangitic carcinoma.133 It is unlikely that this technique will replace transbronchial biopsy as the initial diagnostic approach for this condition. However, very breathless patients with suspected lymphangitic carcinoma may tolerate insertion of a PAC better than bronchoscopy, and those with bleeding tendencies cannot undergo transbronchial biopsy safely.

Narrow-complex, regular tachyarrhythmias at a rate of 140 to 180 beats per minute are common in the ICU. It is sometimes difficult to differentiate atrial flutter from sinus tachycardia and paroxysmal supraventricular tachycardia, even with the aid of a 12-lead electrocardiogram. Although the response to adenosine is often the best way to define the underlying atrial rhythm, in some cases atrial flutter also may be diagnosed by inspection of the Pra tracing34 (Fig. 13-29A). The Pra tracing also may be of value in defining wide-complex premature beats as ventricular in origin if clear-cut cannon a waves are seen34 (see Fig. 13-29B).

Intracardiac left-to-right shunts can occur between atria or ventricles. One of the more serious complications of myocardial infarction is acute perforation of the ventricular septum. Postinfarction ventricular septal defect is not diagnosed reliably by clinical examination. A ventricular septal defect is associated with a step-up in O2 saturation from the right atrium to the pulmonary artery. In addition, the patient may be in shock despite a normal (or elevated) Q̇ttd because the latter measures right-sided rather than systemic flow. Both a ventricular septal defect and acute mitral regurgitation may present as postinfarction pulmonary edema with a new systolic murmur. A step-up in O2 saturation from the right atrium to the pulmonary artery occasionally occurs in acute mitral regurgitation due to reflux of pulmonary venous blood into the pulmonary artery, and a ventricular septal defect occasionally produces a large v wave.34 Differentiation of ventricular septal defect from acute mitral regurgitation by pressure waveform and O2 saturation criteria could prove difficult, but Doppler echocardiography should clarify the diagnosis.

Fick-derived measurement of the volume of oxygen consumed (VO2) may be used to help guide nutritional support. Caloric needs are a function of overall metabolic activity, which in turn can be assessed by measuring VO2. Resting energy expenditure (REE) is the caloric intake required to meet ongoing metabolic demands. The Fick-derived estimate of REE is as follows: REE (kcal/d) = VO2× 7.0, or REE = Q̇t × Hb × (SaO2 − SvO2) × 95.134 Prediction of REE by the Fick method, using PAC-derived data while assuming a respiratory quotient of 0.85, has been shown to correlate reasonably well with measurement of REE by indirect calorimetry.135 While placement of a PAC solely for the purpose of assessing caloric needs would be inappropriate, it may be reasonable to use the available data to help guide nutritional support when a catheter is already in place.