Complications of Pulmonary Artery Catheter Monitoring

The use of PACs has been associated with various adverse effects that may be divided into complications that occur during initial attempts to gain venous access, those occurring during pulmonary artery catheterization, late complications associated with catheter residence in the body, and adverse consequences associated with misuse of equipment and misinterpretation of data. For the most part, problems during attempts to gain central venous access are similar for PAP or CVP monitoring (see Table 32-6 ). However, catheterization of the right heart chambers beyond the cavoatrial junction may cause complications uniquely associated with PACs and rarely, if ever seen with CVP monitoring alone ( Table 32-10 ).

Although many PAC-related complications have been described over the past 25 years, their incidence is not clearly known because they are described, for the most part, in isolated case reports or small case series.^[348] When all adverse effects are considered, including self-limited arrhythmias observed during catheter insertion, it appears that minor complications occur in more than 50% of catheterized patients.^[318] However, of greater clinical relevance, major morbidity specifically attributable to PAC use is uncommon.^{[349] [350]} Based on an exhaustive review of the literature that included 860 publications in

TABLE 32-10 -- Complications of pulmonary artery catheter monitoring

Catheterization

Arrhythmias, ventricular fibrillation

Right bundle branch block, complete heart block

Catheter residence

Mechanical, catheter knots

Thromboembolism

Pulmonary infarction

Infection, endocarditis

Endocardial damage, cardiac valve injury

Pulmonary artery rupture

Pulmonary artery pseudoaneurysm

Misinterpretation of data

Misuse of equipment

its 1993 report and another 90 newer studies in its 2003 update, the American Society of Anesthesiologists Task Force on Pulmonary Artery Catheterization has emphasized that the reported incidence of complications from PAC monitoring varies widely in the medical literature.^[318] The task force suggested that serious complications caused specifically by pulmonary artery catheterization occur in 0.1% to 0.5% of PAC-monitored surgical patients. Two large prospective observational studies of more than 5000 patients each provide good support for this contention. In 1984, Shah and coauthors reported the use of PACs in 6245 patients undergoing cardiac and noncardiac operations.^[249] Quite remarkably, only 10 patients (0.16%) had serious complications resulting in morbidity and only 1 patient (0.016%) died as a result of pulmonary artery catheterization. A 1998 European report of PAC use in 5306 patients undergoing cardiac operations confirmed a similar low incidence of major morbidity, with injury to the right ventricle or pulmonary artery occurring in only 4 patients (0.07%).^[351] Finally, only 1 of the 2000 adverse events reported in the Australian Incident Monitoring Study of 1993 involved the use of a PAC, in contrast to 64 other adverse events involving access to the arterial or venous systems.^[106] Although these large studies suggest that the incidence of serious complications attributable to use of the PAC is low, the frequency of complications in a particular clinical setting or patient group remains unknown. However, clinicians who use PACs must be familiar with all the potential complications associated with catheter use in an attempt to avoid these problems or recognize them early and treat them successfully.

Arrhythmias (also see Chapter 34 ) are the primary complication observed during pulmonary artery catheterization. In fact, self-limited atrial or ventricular arrhythmias are so common during PAC passage through the heart that most clinicians do not consider this a complication, but rather a confirmation that the PAC is traversing the cardiac chambers en route to the pulmonary artery. Shah and colleagues observed transient premature ventricular contractions in 68% and atrial dysrhythmias in 1.3% of their catheterized patients. ^[249] Of greater clinical significance, persistent ventricular arrhythmias requiring treatment occurred in only 3.1% of patients, none of whom suffered prolonged hemodynamic instability. Although the balloon-tipped PAC is less arrhythmogenic when it strikes the endocardium than a standard intravenous catheter or transvenous pacing wire is,^[352] PACs have been reported to induce sustained atrial fibrillation, ventricular tachycardia, and ventricular fibrillation.^{[344] [351] [352] [353]}

As uncommon as these malignant arrhythmias may be, the physician must always consider this risk when placing a PAC and recognize that certain patients may be at increased risk. For example, in a patient with critical aortic stenosis scheduled for valve replacement, one may choose to avoid use of a PAC or delay flotation of the catheter until the heart is exposed and cardiopulmonary bypass support is immediately available. In general, the risk of ventricular fibrillation is greatest in patients with myocardial ischemia, particularly those with acute right ventricular ischemia.^[354] In a study of 2821 patients who underwent pulmonary artery catheterization in the coronary care unit, ventricular fibrillation developed in 25 of 2327 (1.1%)

1305

Prophylactic use of intravenous lidocaine before pulmonary artery catheterization has not been shown to consistently reduce the incidence of ventricular arrhythmias.^{[355] [356]} When ventricular tachycardia or frequent ventricular premature beats occur during the catheterization procedure, the balloon should be deflated and the catheter withdrawn to the right atrium. This terminates the arrhythmia in nearly all cases, and the pulmonary artery may be catheterized by standard procedures once the arrhythmia has ceased. A more difficult clinical question arises when hemodynamically significant arrhythmias develop later in the monitoring period, after the PAC has been in place for hours or days. Although it is unlikely that the PAC is responsible for these new arrhythmias, the position of the catheter tip should always be checked by observation of the pressure waveform and chest radiograph. A catheter tip that is malpositioned within the right ventricle is significantly more arrhythmogenic than a catheter tip, cushioned by the inflated balloon, that is floating through the heart during initial catheterization. Once proper PAC position is confirmed and other causes of ventricular arrhythmia such as myocardial ischemia are excluded, several clues may point toward the PAC as being the cause of the rhythm disturbance, including left bundle branch block morphology of the ventricular premature beats, their resistance to medical therapy, and their response to withdrawal of the catheter into the right atrium. ^[352]

As the PAC passes through the right ventricle and strikes the interventricular septum, transient right bundle branch block occurs in up to 5% of patients.^{[357] [358] [359]} This complication presumably results from the minor trauma of catheter passage through the heart. It occurs less frequently with balloon-tipped catheters than with standard stiff cardiac catheters, which are known to cause right or left bundle branch block, depending on which ventricular chamber is catheterized.^{[358] [360]} Right bundle branch block is of no real clinical consequence except in patients with preexisting left bundle branch block, in whom complete heart block may be precipitated. Despite this theoretical concern, complete heart block from pulmonary artery catheterization is uncommon. Shah and coworkers catheterized 113 patients with preexisting left bundle branch block, and complete heart block developed in only 1 patient (0.9%).^[249] In a different population of 47 high-risk patients with left bundle branch block, many of whom had acute myocardial infarction or heart failure, Morris and associates inserted 82 PACs without a single episode of complete heart block during the catheterization or for 24 hours thereafter.^[357] Prophylactic placement of a transvenous ventricular pacing wire is not necessary in patients with preexisting left bundle branch block who require PAC monitoring. Instead, transcutaneous pacing equipment, an external pulse generator, and a temporary transvenous pacing wire or pacing PAC should be readily available to treat heart block and severe bradycardia. It is important for the anesthesiologist to recognize that the onset of heart block from pulmonary artery catheterization may be delayed and only fully manifested after the administration of anesthetic drugs. For example, incomplete right bundle branch block after PAC insertion has been reported to progress to complete right bundle branch block and bradycardia after induction of anesthesia.^[359]

Various mechanical problems with PACs or introducer sheaths have been reported, and a physician who is familiar with these potential problems should be able to avoid most of them. In the setting of cardiac surgery, PACs may be damaged by surgical instruments or become entrapped by sutures or the cannulas used for these operations.^{[249] [361] [362] [363] [364] [365] [366] [367]} Whenever the surgical procedure involves the right heart structures, it is prudent to ensure free movement of the PAC before chest closure.

Sternal retraction during cardiac surgery poses another mechanical problem with PACs, particularly when they are inserted through the external jugular or subclavian routes. The PAC may kink as it exits the introducer sheath and makes an acute angle between the sheath and vessel wall.^{[335] [336] [337]} Withdrawing the introducer sheath slightly or limiting the amount of surgical retraction may resolve this problem. These mechanical difficulties should be suspected whenever there is a damped appearance of the monitored pressure traces or difficulty infusing fluid or withdrawing blood through one of the catheter lumens.

Although gross structural defects in the PAC should be recognized by inspection before catheter insertion, more subtle manufacturing problems may escape detection. Some of these, such as interluminal communications between the PAP, CVP, and balloon inflation lumens, may be suspected only when information derived from the catheters is inconsistent with other clinical data.^{[326] [327] [368] [369]}

PACs may dislodge temporary transvenous pacing wires, become entangled with other cardiac catheters, or form knots within the heart.^{[370] [371] [372]} Arnaout and coauthors reported a PAC knot around the tricuspid valve chordae tendineae that resulted in severe tricuspid valve regurgitation after catheter removal.^[370] Catheter knots should be suspected when difficulty is encountered during withdrawal of a PAC, and the diagnosis may be confirmed by chest radiography. Knots may be untied in vivo by radiologists with the use of intravascular snares and fluoroscopic guidance. ^[373] If the knot has already been drawn tight, removal under direct surgical exposure of the venous cannulation site may be required. ^[374] These problems are more likely when the PAC is inserted an excessive distance and the catheter loops or coils within the cardiac chambers, thus emphasizing the importance of prevention.

The use of PACs has been associated with thrombocytopenia,^[375] and the catheter may serve as a nidus for intravascular thrombosis.^[376] Although the incidence of thromboembolic complications is increased in patients who require PAC monitoring for longer periods, thrombi have been noted to be attached to a PAC within 1 to 2 hours of their insertion.^{[377] [378]} When drugs such as aprotinin and epsilon-aminocaproic acid are used to reduce perioperative bleeding, the risk of thrombus formation on the PAC may be even greater.^{[379] [380]} Although heparin bonding of the external surface has unquestionably reduced the thrombogenicity of PACs, it does not entirely eliminate

1306

Although minor venous thrombi and pulmonary ischemic lesions are fairly common in patients with PAC monitoring, major pulmonary embolism arising from the use of a PAC is a rare occurrence.^{[376] [383] [384]} Rather than thromboembolism, the most common cause of pulmonary infarction in patients with PACs is occlusion of pulmonary blood flow by a malpositioned catheter that occludes a branch of the pulmonary artery.^[383] The radiographic hallmark of this condition is a wedge-shaped infiltrate radiating from the tip of a peripherally located PAC. This iatrogenic complication may be prevented by diligent monitoring of the PAP waveform recorded from the catheter tip to allow early recognition of a catheter that has migrated distally, wedged within a small pulmonary artery, and obstructed distal blood flow.

PAC-related infection occurs more commonly in patients who require monitoring for more than 3 days and those with preexisting sepsis.^[385] Using sophisticated microbiologic techniques in 297 critically ill medical and surgical patients, Mermel and associates demonstrated a 22% incidence of local infection of the introducer sheath, but a low 0.7% incidence of bacteremia caused by the PAC.^[386] Although PACs may be colonized from multiple sources, the patient's skin is the most important source of organisms, and it is the introducer sheath rather than the PAC itself that is usually the infected device. In view of these findings, it is not surprising that scheduled changes of PACs do not reduce the risk of bloodstream infection, particularly when the catheter is changed over a guidewire through the original venipuncture site.^[387] If clinical factors mandate changing the PAC, placement through a new vascular access site will reduce the risk of infection in comparison to guidewire-assisted exchange. However, cannulation at a new puncture site carries a greater risk of mechanical complications, and these risks and benefits must be balanced in each individual patient.^[387]

The most serious infection related to PACs is endocarditis. The pulmonic and tricuspid valves are the most likely sites, but right atrial endocarditis has also been reported.^{[388] [389]} Although these catastrophic infections appear to be rare, they represent part of the continuum of endocardial damage, thrombosis, and infection that has been documented to be associated with the use of PACs.^[389] Immunocompromised, debilitated, and bacteremic patients may be at particular risk for these complications.

In addition to predisposing the patient to endocarditis, PACs may damage the right heart endocardium and the tricuspid and pulmonic valves.^{[382] [389] [390] [391]} Rowley and coworkers performed autopsies on 55 previously catheterized patients and found that 53% had right-sided endocardial lesions. ^[389] Although both severe tricuspid regurgitation and severe pulmonic regurgitation have been reported,^{[370] [390] [391]} these conditions are rare complications of PAC use. In general, the anatomic redundancy in valvular tissue appears to prevent clinically important valvular regurgitation in patients who have PACs in place. Using color flow Doppler echocardiography, Sherman and colleagues demonstrated that placement of a PAC caused a slight increase in the magnitude of tricuspid regurgitation, but in no instance did the PAC lead to severe right-sided valvular insufficiency. ^[392]

Perhaps the most life-threatening complication of pulmonary artery catheterization is one that is uncommon and often avoidable through meticulous attention to insertion and monitoring techniques. Pulmonary artery rupture occurs in approximately 0.02% to 0.2% of catheterized patients and appears to be a fatal complication in nearly 50% of cases.^{[249] [393] [394] [395] [396]} Although it may occur in a variety of monitoring situations, many reported cases have involved patients undergoing heart surgery requiring cardiopulmonary bypass.^{[395] [396] [397] [398]} Several patient factors common to this clinical setting have been proposed to increase the risk of this complication, including hypothermia, anticoagulation, advanced age, and pulmonary hypertension.^{[398] [399] [400]} The latter may predispose patients to arterial injury during balloon inflation as a result of the increased gradient between the proximal arterial and distal wedge pressures or because pulmonary hypertension distends the pulmonary vasculature and causes the PAC to wedge in a distal, less compliant vessel. Both factors may result in eccentric balloon inflation that could drive the exposed catheter tip into the vessel wall.^{[401] [402]} However, whether preexisting pulmonary hypertension increases the risk of pulmonary artery rupture during PAC monitoring remains unproven.

Several mechanisms for pulmonary artery injury have been proposed and investigated. Hardy and associates, studying cadaveric specimens, determined that the pressure required to rupture the pulmonary artery in patients older than 60 years was well within the range of pressures normally exerted during balloon inflation, thus suggesting that a balloon forcefully inflated in a pulmonary artery smaller than the balloon diameter may rupture the vessel wall.^[400] Other proposed mechanisms for pulmonary artery rupture include chronic erosion by a catheter tip abutting the vessel wall or eccentric balloon inflation that forces the uncushioned catheter tip through the vessel wall.^{[399] [403]} Regardless of the precise mechanism by which pulmonary arterial injury occurs in an individual patient, case reports highlight that this complication often results from suboptimal catheter insertion and monitoring techniques. These procedural errors include unnecessary catheter manipulation, excessive insertion distances predisposing to catheter tips becoming lodged in distal small pulmonary arteries, unrecognized persistent wedge pressure, prolonged balloon inflation, or improper balloon inflation with liquid rather than air.^{[329] [395] [399] [402] [403]} It is critical that the clinician recognize artifactual "overwedged" pressure recordings that indicate a peripheral location of the PAC tip or impaction against the vessel wall and correct this problem immediately by withdrawing the catheter into the proximal part of the pulmonary artery (see later).

The hallmark of catheter-induced pulmonary artery rupture is hemoptysis, which may cause life-threatening

1307

Treatment focuses on stabilizing cardiorespiratory performance, maintaining adequate gas exchange, and arresting the hemorrhage. Specific therapeutic steps are highly individualized; they depend on the setting in which the complication occurs and may be quite different when pulmonary artery rupture is suspected in a debilitated patient in the intensive care unit versus a relatively fit patient undergoing cardiopulmonary bypass. The first priority is ensuring adequate pulmonary ventilation, which may require endobronchial intubation with either a singleor double-lumen endotracheal tube to allow selective unilateral lung ventilation and protect the unaffected lung. ^[404] Positive end-expiratory pressure (PEEP) applied to the affected lung may help control hemorrhage.^[398] Anticoagulation, if present, should be reversed unless the patient must remain on cardiopulmonary bypass. Bronchoscopy is performed for tracheobronchial toilet and to localize the site of bleeding. Other maneuvers may be performed at the bedside to provide temporary control of hemorrhage. A bronchial blocker may be guided into the involved bronchus to tamponade the bleeding and prevent continued contamination of the uninvolved lung.^[393] Management of the PAC itself in these circumstances is more controversial. Some authors recommend removing the catheter,^[394] but others suggest leaving the PAC in place to monitor PAP and guide antihypertensive therapy targeted at lowering this pressure and reducing bleeding.^[395] Others have suggested that the PAC balloon may be carefully reinflated and the catheter floated into the involved pulmonary artery to occlude the bleeding arterial segment as a temporizing measure.^[393] Although these conservative medical treatments may be effective in some cases, many patients will require definitive surgical therapy, such as oversewing the involved pulmonary artery or resecting the involved segment, lobe, or lung^{[393] [394] [398] [399]}

Patients who have this complication and are managed with conservative medical therapy alone are at risk for pseudoaneurysm formation and secondary hemorrhage.^{[393] [398] [402] [405] [406]} Because of these delayed risks and the high mortality associated with rebleeding, some authors have recommended follow-up pulmonary angiography and coil embolization or elective surgical resection in patients who were initially managed medically. ^{[393] [396] [398] [405] [407]}

A more insidious complication of PAC monitoring and perhaps the most common serious adverse effect from the use of this technology is misinterpretation of data leading to inappropriate, harmful therapy.^{[408] [409] [410] [411]} Although it is not clear what role this factor plays in creating complications in an individual patient, there is reason to believe that there are widespread knowledge deficits among practitioners who use PACs in clinical care. In 1990, Iberti and coauthors reported the results of a 31-question multiple-choice examination that tested knowledge of the PAC.^[412] The test was administered to 496 resident and staff physicians in the medicine, surgery, and anesthesiology departments of 13 North American medical centers. The authors found a poor overall level of knowledge of PACs, as evidenced by a mean test score of only 67% correct answers, and they concluded, "there is reason to be concerned about the effective use of the PAC." Although higher scores were achieved by individuals with more training and more experience inserting PACs and using catheter-derived information in patient management, none of these factors ensured a high level of knowledge. Similarly disappointing results were subsequently obtained when this examination was administered to 216 critical care nurses,^[413] 535 European intensive care physicians,^[414] and 1095 members of the U.S. Society of Critical Care Medicine.^[415] The authors of these reports seemed particularly concerned that PAWP measurement was performed incorrectly by 30% to 50% of the clinicians in these studies. Other authors have also noted the large interobserver variability in interpretation of PAC pressure tracings,^[416] and more recent reports continue to document the inability of practicing cardiovascular anesthesiologists to measure wedge pressure accurately^[417] and the failure of brief educational programs to improve this critical diagnostic skill.^[418] Taken together, these observations highlight the fact that effective use of PACs requires a great deal of expertise and clinical experience, and even measuring the most fundamental PAC-derived variable, namely, wedge pressure, may prove to be a complicated exercise.^{[419] [420] [421]} With these issues in mind, the physiologic basis for PAC monitoring will be reviewed, followed by a description of normal and pathologic PAC waveforms, with an emphasis of the physiologic mechanisms that produce them.