Insertion Tips and Guidelines

Insertion of the catheter is done through the gasketed introducer port of a Cordis™ catheter. Full sterile technique is paramount in order to reduce infection rates. We commonly place the introducer catheter and the PAC sequentially with one sterile prep and setup, which prevents contamination at the introducer gasket as well as the necessity of reprepping of a previously placed introducer. At times, however, this cannot be avoided, or a PAC will be placed through a previously located introducer line. In this instance, the introducer catheter and surrounding skin should be prepped widely with chlorohexidine preparation. In all cases, wide sterile prep with chlorohexidine, full sterile precautions including gown, hat, mask, and sterile gloves are essential. Any break in sterile technique has been shown in many studies to significantly increase the infection rate and implementation of the above precautions has been reported to reduce the infection rate to near zero.^16–21 Another tip is to prep widely. We cannot stress this seemingly trivial point too strongly. Opening, preparing, and inserting a PAC results in an often-unwieldy octopus of catheter, tubing, and transducers that have to cross through a sterile to unsterile transition zone. Wide prep of all or most of the bed with a sterile half-sheet facilitates easy handling and reduced risk of contamination. Once an introducer catheter is in place, the PAC is removed from the packaging and the proximal end is passed to an assistant who will connect and flush the catheter ports. At this time, the transducer is connected, zeroed, and tested. The balloon is also tested. Placement of the catheter sheath allows future adjustments without additional sterile prep.

Constant verbal communication between the floater and the assistant is essential. Once the tip of the catheter is passed through the introducer and into the blood vessel, it must be advanced only when the balloon is inflated (up), and conversely it must never be withdrawn unless the balloon is deflated (down). The slang “floating” a swan refers to the fact that the catheter advances by floating along with the blood flow. Direct advancement can cause vascular injury or perforation. A constant communicative banter—“balloon up?,” “balloon up,” “balloon down?,” “balloon down”—can prevent injuries such as arterial, cardiac, and PA rupture.

Initially the catheter is inserted to the 10–15-cm mark, allowing the balloon to pass through the introducer. Once the catheter tip is safely past the introducer, the balloon is inflated and the catheter slowly advanced by feeding slack catheter and allowing the catheter to be pulled downstream by the blood flow. A combination of clinical experience, length of catheter inserted, and careful attention to the monitored transducer tracing allow successful placement. As the catheter tip enters the vessel, the initial pressure reading will be the CVP, which first transduced at ∼15–25 cm depending on patient size and insertion site. With another 5–10 cm of advancement, the catheter passes into the right atrium and the transducer will register a distinct right arterial waveform. Another few centimeters of advancement passes the catheter tip through the tricuspid valve (around 30 cm) and the easily recognizable spike of RV pressure will register on the monitor. Slow careful advancement of another 5–10 cm will pass the catheter into the PA (tracing), and will soon wedge the catheter to produce a flattening of the waveform and characteristic wedge tracing.

Often several attempts to ensure proper wedging are required. When a RV waveform is not evident after 35–45 cm or no wedge tracing occurs after the catheter has advanced 10–15 cm past the RV tracing, the balloon should be deflated, and the catheter pulled back to the RA or RV. Once a distinct tracing identifies the catheter location, another careful attempt at reinsertion and wedging can begin. When the catheter is inserted and wedged, the pulmonary capillary wedge pressure (PCWP) is noted and the balloon deflated. The catheter is locked into place and the sterile covering advanced to cover the full length of the catheter. Placement of the catheter is confirmed by a chest x-ray.

Initial Warnings and Potential Measurement Problems

While the PAC provides a lot of otherwise unobtainable data, only diligent attention and educated interpretation make PA catheterization worth the risk, time, and cost. Without proper placement, all obtained data are erroneous and will lead to false interpretation and incorrect (and possibly harmful) intervention. Careful attention to the PAC waveform helps confirm catheter placement. Entry into the pulmonary artery is evidenced by a decrease in mean pressure and characteristic wave form. Upon entering the PA, a triphasic waveform is seen, which reflects atrial and ventricular contraction. From the PA careful advancement of the catheter will “wedge” the catheter.²²

From the PAOP and CVP, we begin the pressure-volume measures. Once wedged or occluded, the inflated balloon prevents forward blood flow from the right ventricle or proximal pulmonary artery and allows only back-pressure to be transduced at the catheter tip. This transduced “wedge” pressure reflects a contiguous fluid column from the PA through the pulmonary vasculature to the left atrium (LA). The unobstructed column continues through the mitral valve into the (when open) left ventricle (LV), and hence,

where LVEDP is left ventricle end-diastolic pressure. It is crucial to note that this relationship holds true only when there is a contiguous fluid column between the PA and LV. A common reason for improper interpretation of PAOP occurs when this contiguous fluid column is interrupted or affected. Examples of interrupted fluid column include mitral valve disease (stenosis or insufficiency) and pulmonary venous obstruction. Also affecting the relationship between PAOP and left ventricular end-diastolic volume (LVEDV) is the ability to extrapolate volume from pressure measurements. Since PCWP is ultimately used as a measure of ventricular filling (volume), this necessitates the ability to predictably extrapolate volume from pressure, which is only possible with normal compliance. Indeed, LVEDP is analogous to left ventricular end-diastolic volume (LVEDV) only in a normally compliant heart and thorax. Many factors influence the ability to use PCWP as an accurate measure of left heart–filling pressures. As described previously, the analogous relationship between LVEDP and LVEDV is accurate only insofar as the ventricle is normally compliant. In cases where left heart compliance is reduced such as ventricular hypertrophy or ischemic myocardial damage, the PCWP will be elevated for a corresponding left ventricular pressure/volume and cannot be considered an accurate measure. While trend analysis can be used to monitor fluid resuscitation in these patients, injured or septic patients often manifest rapidly changing cardiac compliance making interpretation difficult and suspect. Indeed multiple animal models have shown a rapid flux of cardiac edema resulting from injury, resuscitation, and inflammation, all of which can rapidly change ventricular compliance and render the PAC an unreliable measure of cardiac filling.²²

Along with proper wedging in the pulmonary artery and a normally compliant heart, the pulmonary catheter must be placed in the proper lung zone. In order for PCWP to reflect a vascular rather than alveolar pressure, the catheter must be placed in West zone 3 of the lung. Here Pa > PV > PA, so the measured pulmonary venous pressure reflects vascular pressure instead of pulmonary alveolar pressure. If the catheter is in zone 1, PA > Pa > PV, or 2 Pa > PA > PV, the catheter will reflect alveolar pressure rather than the PA pressure.

Constantly changing thoracic pressures from ventilation (spontaneous or mechanical) are reflected in the PAC tracing. In a positive pressure–ventilated patient, the PCWP should be measured at expiration when the pleural pressures are closest to zero. This is reversed if the patient is spontaneously breathing. In a spontaneously breathing patient the pressure normalizes somewhere between expansion and inhalation and by convention, the pressure is measured at end inspiration. Direct monitoring and examination of the tracing rather than reliance on automatically collected single number display is always better. In an extreme situation, brief paralysis might be necessary to get a useful PAWP tracing in an agitated or ventilatory dyssynchronous patient.

Positive end-expiratory pressure (PEEP) also affects the PAC tracing, and interpretation of the PCWP in a patient affected by intrinsic or intrinsic PEEP is the subject of much conjecture and misinformation. PEEP (both intrinsic and extrinsic) causes transmural thoracic pressure, which positively effects PAWP measurements. To correctly determine the PAWP, the PEEP can be temporally disconnected (for a few seconds or breath cycles). Alternatively, the effect of intrinsic or extrinsic PEEP can be estimated by adding 2–3 cm H₂O to each 5 cm of PEEP >10.²²

Cardiac output (CO) is measured by thermodilution. Traditionally, a cooled saline bolus was injected proximally, and the temperature of the bolus was measured by a thermistor at the catheter tip. CO is calculated by determining the area under the curve and plotting the integral. An average of three to five injections were used to minimize variability. Newer catheters measure CO continuously by integrating a thermal coil, which gently heats the proximal blood, measures the temperature drop at the catheter tip and uses a similar method to calculate CO. In either method, intracardiac shunt and tricuspid regurgitation will lead to unpredictable error in cardiac output measurement. At least one set of authors suggest that greater reliability can be had by calculating CO through plotting of the Fick curve and extrapolating CO.

Pressure, Volume, and Work Measures

While modern PAC monitoring units or ICU nurses provide continuous data readouts of hemodynamic parameters, it remains essential to have a grounded understanding of the physiologic principles and calculations involved in PAC monitoring. Actual measured parameters begin with CVP. CVP is directly transduced from the proximal or CVP port of the catheter. Without obstruction between the right atrium and the vena cava, CVP is analogous to right atrial pressure (RAP), which is again analogous to right ventricle enddiastolic pressure (RVEDP) and right ventricle end-diastolic volume (RVEDV). Hence,

CVP is a direct measured extrapolation of RVEDV or cardiac preload. Just as CVP is a measure of right heart filling, PCWP measures left-sided pressure and volume. As described previously when the catheter is “wedged” in the pulmonary artery, there is a continuous fluid column between the PA and LV. Hence,

Cardiac output is also measured directly through the previously mentioned thermodilution. CO is converted to a more normalized cardiac index (CI) by dividing CO by body surface area (BSA). BSA is derived from height and weight normograms, but can also be easily calculated as follows:

Once CO is known, stroke volume (SV) can be calculated:

where HR is heart rate.

Right-sided ejection fraction (RVEF) is then calculated as follows:

Calculation of RVEDV requires a continuous thermistor, which is callable of dividing CI into systolic and diastolic components. Knowing RVEF allows RVEDV to be calculated:

Right-sided heart work estimates the work of the right heart as it pumps blood through the lungs back to the left ventricle. The right ventricle stroke work index (RVSWI) is estimated as follows:

where PAP is pulmonary artery pressure, and SVI is stroke volume index.

The left heart volume and work are calculated similarly to the right. The left ventricular work is the work of the blood circulating through the systemic circulation. This is estimated by

where LVSWI is left ventricle stroke work index, and MAP is mean arterial pressure.

Among the most useful calculated measures provided by the PAC is the pulmonary and systemic vascular resistance. Each estimates the resistance across the pulmonary or systemic circulation and is an extrapolation of vascular tone. Systemic vascular resistance (SVR) is the drop in pressure across the systemic circulation times the flow and is calculated as

where RAP is renal arterial pressure.

Pulmonary vascular resistance (PVR) is the drop in pressure across the pulmonary circulation multiplied by right-sided flow:

Goal-Directed Therapy Using Pulmonary Artery Catheter

There is little disagreement regarding the large amount of data provided by a properly placed and interpreted PA catheter. Whether the data provided result in better patient outcome is the source of considerable debate. Much clinical research has been done in an attempt to determine which PAC-derived physiologic parameters should be measured and which are predictive of outcome. Shoemaker and his group provided the initial work toward defining and grouping patient physiologic response to trauma.²³ As early as 1973 Shoemaker and colleagues showed in several studies that PAWP measured reduced cardiac output. They further showed that decreased oxygen delivery along with increased peripheral vascular resistance characterized non-survivors after trauma.²⁴ Based on these and other similar study results, many groups targeted resuscitation of severely injured trauma patients to achieve physiologic values retrospectively associated with survival. Ten years after his initial studies, Shoemaker and colleagues showed an overall survival benefit in patients who achieved normal physiologic values.^25,26 Finally, this group showed that there was less oxygen debt as measured by VO₂ in surviving patients.²⁷ Interestingly, there was no difference in oxygen consumption between the groups. In similar research, Bishop et al. examined physiologic parameters in 90 trauma patients and found that patients who achieved higher levels of CI, oxygen delivery, and oxygen consumption within the first 24 hours after admission had a lower incidence of ARDS and reduced overall mortality.²⁸ Based on these and similar findings, several groups then randomized patients to supernormal resuscitation versus traditional resuscitation with the intent of proving that invasive monitoring of aggressive resuscitation would allow achievement of supranormal physiology and better outcome. These studies showed mixed results. Bishop’s group randomized patients to be supranormally resuscitated to a cardiac index of more than 4.5, DO₂I greater than 670, and a VO₂I greater than 166, versus traditional resuscitation as defined by achievement of normal urine output and CVP.²⁹ The supranormally resuscitated group showed significantly lower mortality and decreased organ failure. Interestingly, optimal parameters were reached in 70% of the study group and 29% of the control group. Similarly, Flemming et al. randomized trauma patients with blood loss of more than 2000 cc to supernormal CI, DO₂, and VO₂, and found that attainment of these goals resulted in statistically significant decreases in organ failure, ICU stay, and ventilator days, and a trend toward decreased mortality.³⁰

Other studies, however, pointed out that attainment of resuscitation was more important than the means by which it was achieved and many more recent studies have been unable to show similar benefit to monitoring or goal-directed resuscitation. Velmahos et al. randomized severely injured patients to supranormal resuscitation versus traditional hemodynamic monitoring. While there was no difference in mortality, organ failure, sepsis, ICU days, or hospital stay between the two groups, they found that 70% of the supranormal group and 40% of the traditional group achieved supranormal physiologic parameters, and that attainment of these optimal values was associated with better outcome independent of treatment protocol. There were no outcome differences between the two groups. What was perhaps most interesting in this study, however, was the finding that nonresponders (those who failed to reach the set goals despite volume and ionotropic manipulation) fared significantly worse than the control group. This suggests the likelihood that aggressive resuscitation toward supranormal physiology will be harmful in those whose physiology cannot attain it, and is in fact exhausted such attempts. Other groups have reported similar results, which indicate that that the ability to achieve normal parameters is associated with the mortality benefit, and is ultimately more important than whatever treatment was designed to achieve them.³¹ Taken together, these results cast doubt regarding the usefulness of interventional monitoring and resuscitation. Indeed, many believe that aggressive resuscitation should be instituted in all patients without any need for invasive monitoring. Others argue for limited supraphysiologic resuscitation. Some data points to poorer outcome from PAC use. Indeed, deleterious consequences of aggressive resuscitation and the PAC have been reported. Hayes et al. reported that increasing oxygen delivery with ionotropic augmentation with dobutamine was associated with increased mortality.³² Others have reported similar increased mortality in supranormally resuscitated patients. At least one group has purported to show that the PAC is itself a predictor of morbidity independent of protocol-driven resuscitation. Rhodes et al. performed a prospective randomized trial of 200 patients where no protocol-driven resuscitation was instituted.³³ In this study, patients were randomized to PAC placement versus no PAC where patients were resuscitated based up the clinical judgment of the ICU staff. Data were obtained prospectively and examined retrospectively, and revealed that the PAC group received significantly more fluid in the first 24 hours and exhibited increased morbidity. Evidence for the deleterious effects of PAC protocol–driven resuscitation were seen in a study by Hayes et al. that showed an increased mortality in PAC-treated patients (54% vs. 34%).³² In this study there was a rigorous goal-directed attempt to achieve supranormal physiologic parameters. To achieve these goals, aggressive fluid resuscitation and ionotropic support were used.

In an attempt to make sense of the huge amount of data regarding PACs and endpoints of resuscitation, Kern and Shoemaker performed a meta-analysis on 21 randomized clinical trials and found that early hemodynamic optimization achieved significant mortality reduction only if optimized parameters were achieved prior to the onset of sepsis or organ dysfunction.³⁴ Heyland and associates analyzed seven studies and found no significant reductions in mortality achieved with optomization.³⁵ Lastly, an analysis by Poeze et al. showed that physiologic optimization resulted in decreased mortality. In this analysis the entire benefit was from patients who hemodynamically optimized perioperatively. Patients with established sepsis or organ failure prior to treatment were unlikely to benefit from “optimization.”³⁶

Others have questioned the accuracy of the actual PAC measurements. Epstein et al. studied the accuracy of VO₂ measurements by the calculated reverse Fick method versus direct measurement by indirect calorimetry in an attempt to determine the accuracy of PAC determined oxygen consumption measures. Their study showed a bias of 41 ml/min/m², indicative of an undermeasurement of VO₂ by PAC measures.³⁷ Finally, Luchette and colleagues studied the accuracy of continuous cardiac output measurements and found that there was higher signal-to-noise ratio and degraded accuracy when patient temperature was above 38.5° C, resulting in further questioning of the accuracy and usefulness of PAC measurements in severely injured or septic patients.³⁸

Mixed Venous Saturation: Monitoring Tissue Metabolism

While the results of supranormal resuscitation are mixed, many believe that the best use of a PAC is as a measure of tissue perfusion. PAC sampling of mixed venous saturation should serve as an adjunctive measure of tissue-level perfusion. Even if supranormal resuscitation remains controversial with mixed data, surely SVO₂ should provide useful data. Unfortunately, this measure of tissue perfusion, while seemingly useful as a measure of tissue oxygenation and subsequent outcome has also not ultimately been shown useful as the targeted endpoint of resuscitation. Gattinoini and associates published a randomized prospective trial involving 762 patients across 56 ICUs who were randomized to either traditional resuscitation or targeting of CI > 4.5 or SvO₂ > 70%. There was no difference in mortality or organ dysfunction among the three groups. Like the studies discussed above, subgroup analysis showed that patients who achieved hemodynamic targets had similar mortality rates independent of the group to which they were randomized.

Right Ventricle End-Diastolic Pressure as Measure of Cardiac Index and Cardiac Function

As predicted by the Starling curve, end-diastolic volume is the best predictor of preload. Unfortunately, PAC pressure measurements as discussed previously are confounded in the setting of changing intrinsic ventricular compliance and extrinsic PEEP. RVEDV monitoring provides the first actual measurement of preload. Rather than being a pressure surrogate for volume, RVEDV utilizes a PAC with a fast thermistor to give a directly measured volume measure. In an attempt to find an alternative to PAOP as a measure of cardiac function, Chetham et al. showed that CI correlated better with RVEDVI than PAOP.³⁹ The same group then studied 64 patients with respiratory failure and high PEEP requirements, and suggested that RVEDVI was a better measure of CI than PAOP at all levels of PEEP. CI was inversely correlated with PCWP at PEEP levels over 15. Taken together, these findings caused this group to conclude that RVEDVI should be used in lieu of PAOP as a measure of cardiac function.

Despite initial reservations that mathematical coupling was responsible for the close correlation between RVEDV and CI, several groups have shown that the directly measured volume measures are superior to extrapolations from pressure (CVP and PAWP) measures. Chang and colleagues showed that splanchnic malperfusion was better predicted by RVEDV than PCWP, which was in turn associated with MODS and mortality.⁴⁰ Other correlations among RVEDV, tissue perfusion, and outcome have been similarly reported by several groups.⁴² Lastly Kincaid et al.⁴¹ presented data to show that optimal RVEDV could be calculated for each individual patient. Taken together, these data support the use of RVEDV as a superior measure of cardiac function and goal-directed resuscitation.