The Story of the Pulmonary Artery Catheter: Five Decades in Critical Care Medicine.

Abstract

In the inaugural issue of Critical Care Medicine (CCM), Peter Safar (1) wrote an editorial, in which he described the purpose of the new journal: “1) to provide a vehicle for new information, 2) to review and synthesize information, 3) to provide peer review for original publications, 4) to publish recommendations, guidelines, and standards, 5) to report evaluations on new equipment and techniques, 6) to publish abstracts of world literature, 7) to provide a forum for open discussion of controversial topics, 8) to announce future meetings and reports of past meetings, and 9) to provide reports of matters concerning Society of Critical Care Medicine (SCCM) and the federation.” In this article, we will look at the evolution of the use of the pulmonary artery catheter (PAC) in the critically ill as it was reported in the Journal, reflecting the context of these purposes. This article is not meant to be an extensive history of the use of the PAC, as there are many articles in other journals that contribute to this history, but rather is intended to reflect how CCM has contributed to our field. Many important articles on the use of the PAC were published in other high impact journals, but much of the groundwork and subsequent debates on the topic were published in CCM. There are nearly 400 articles about the PAC in CCM over the past five decades, with over 2,000 articles incorporating data from the use of the PAC (Fig. 1). From the earliest days of critical care, measurement of cardiac output and hemodynamic monitoring were recognized to be important in understanding the physiology in critically ill patients, especially patients in shock. However, methods to measure cardiac output were cumbersome or not widely available. Early noninvasive efforts to evaluate cardiac performance using electrocardiography, phonocardiography, and carotid pulse tracings to describe systolic time intervals in 12 patients with hemorrhagic shock were described by Rodriguez et al (2) in the first volume of CCM. This early noninvasive approach to hemodynamic monitoring was not reliable or practical enough for general use in the ICU.Figure 1.: Number of publications in Critical Care Medicine per year over five decades.With the development of the flow directed, balloon-tipped catheter, routine hemodynamic measurements of pulmonary artery pressure and pulmonary artery occlusion pressure, cardiac output, and mixed venous oxygen saturation (Svo2) became easily feasible and widely used in the critically ill patient (3). In the early years, hemodynamic measurements using the PAC (commonly referred to as the Swan-Ganz catheter) were largely directed at understanding physiology, as well as predicting prognosis (4–7). Nearly every hemodynamically unstable adult in the ICU had a PAC placed in those early years. Our understanding of the physiology of shock, as well as approaches to treatment, grew dramatically. Shoemaker et al (5) reported detailed analysis of measured and derived hemodynamic variables and sorted them in groups related to blood volume, flow, stress response, oxygen transport, and tissue perfusion, looking at prediction of outcome in critically ill postoperative patients. In general, the perfusion-related variables were most predictive of outcome (5). Bland et al (6) further identified that critically ill postoperative nonsurvivors had reduced cardiac performance and pulmonary function compared with survivors. Using data from the PAC, the hemodynamic profile of adults with septic shock was defined and recognized to be hyperdynamic early in septic shock for both survivors and nonsurvivors (7). As the physiology in patients with shock became better understood, intensivists turned to using hemodynamic measurements from the use of the PAC to look at therapeutic responses (8,9). The PAC allowed measurements of Svo2, which help to interpret cardiac output measurements, as the Svo2 is determined by the balance between oxygen delivery and oxygen consumption, and a low Svo2 (e.g., <70%) reflects circulatory stress. Shoemaker et al (8) published a landmark article describing goal-directed therapy in surgical patients, applying the understanding of physiology to treatment. Using the hemodynamic parameters of survivors of postoperative life-threatening problems, patients were randomized to treatment to achieve cardiac index greater than 4.5 L/min/m2, oxygen delivery greater than 550 mL/min/m2, mean arterial pressure greater than 70 mm Hg, normoxia, and other normal physiologic parameters or to standard critical care treatment. Fluids, blood transfusion, vasopressors, or vasodilators were used to reach the predetermined goals in the protocol group. Mortality was significantly lower in the protocol group than in the control group. Although this study, as well as many subsequent trials, supported the potential benefit of goal-directed therapy, it should be noted that a “one-size-fits-all” approach to increasing oxygen delivery will not be of benefit to all patients and, in fact, will be harmful to some. With increasing understanding of physiology and widespread use of the PAC, it became apparent that clinical evaluation did not accurately predict hemodynamic parameters compared with measurements using the PAC (10). Eisenberg et al (10) reported that physicians only predicted pulmonary artery occlusion pressure 30% of the time and cardiac output 50% of the time. Therapy was often changed in response to measurements from the PAC. Debate arose, however, as to whether the use of the PAC might not result in improved outcomes and, in fact, might be harmful. These questions were addressed in a number of important articles published in CCM (11–14). Although there were certainly some complications associated with the use of the PAC, they were infrequent and largely related to the placement of central venous access (11). Mimoz et al (12) reported an improved prognosis in patients with shock in whom a change in therapy was instituted based on PAC measurements. Others reported no difference in mortality with the use of the PAC (13,14). More recently, Ivanov et al (15) reported a reduction in morbidity with PAC-guided treatment. Friese et al (16) used the National Trauma Data Bank to look at PAC use in severely injured patients and reported a reduction in mortality with the use of the PAC. Although there are some populations that likely benefit from the use of the PAC in experienced hands, in the general ICU population, the PAC is unlikely to change outcomes. The lack of impact on outcome in general patient populations has two explanations: either the information obtained with the PAC does not result in significant changes in management or the changes in management based on the PAC findings do not improve outcome (17). With the ongoing debate about the benefit, or lack thereof, of the use of the PAC, questions were raised about the interpretation of the pulmonary artery balloon-occluded pressure (18). It became apparent that clinician knowledge of the use of the PAC and accuracy of measurements obtained was suboptimal, potentially contributing to the lack of demonstrated benefit (19–23). In a questionnaire to ICU physicians, Gnaegi et al (20) reported that 50% were not able to correctly identify the pulmonary artery occlusion pressure from a clear tracing. A multiple-choice questionnaire was given to ICU nurses at a national meeting; the mean score on the examination was 48.5% (21). The nurses scored better on practical aspects of PAC use such as waveforms and complications than on interpretation of physiology but still had less than 60% correct interpretation of waveforms. Education of physicians and nurses on the correct use and interpretation of PAC measurements became an important focus. An online PAC Education Program was developed by the SCCM, without industry funding, following the 1997 PAC Consensus Conference organized by the SCCM (24). The hope was that the Education Program would standardize practice using PACs, but unfortunately, it was not as widely adopted as had been expected, and it was eventually discontinued. Even the PAC Consensus Conference Consensus Statement generated controversy, as demonstrated in two accompanying editorials by Sibbald and Keenan (25) and Taylor et al (26). Sibbald and Keenan (25) criticized the Consensus Statement as not following strict guidelines for review of evidence and concluded that the recommendations based on “expert opinion” were subject to bias. Taylor et al (26) countered that a different but also valid and rigorous process was used, along with grading of the evidence, for the Consensus Conference. Both authors agreed that further evidence, potentially from randomized controlled trials, is needed to clearly define the appropriate use of the PAC. The use of the PAC has continued to be the subject of controversy, and there has been a rich discussion over the years in CCM (27,28). The use of the PAC clearly declined over time (17,29,30). The PAC is more commonly used in surgical ICUs than in nonsurgical units, presumably because it was initially inserted in the operating room to monitor intraoperative status (30). There are many reasons for the decline in the use of the PAC (Table 1) (31). The lack of demonstrated benefit with the use of the PAC in randomized trials has led to diminished enthusiasm for routine use of the catheter. The development of alternative strategies and especially echocardiography and alternative methods to continuously measure cardiac output have been a major contributor to this decrease. Still, the reality exists that no monitoring device, no matter how insightful its measures and relevance to assessing patient pathophysiology, will improve patient outcomes unless coupled to a treatment, which itself improves outcomes. With decreased use of the PAC, there is less experience with the placement of the catheter and interpretation of the data provided, leading to concern that there may actually be increased risk associated with the PAC in inexperienced hands. There are few effective novel effective treatments for most critical illnesses that directly benefit from the panoply of physiological data supplied by the PAC. With the concern about lack of benefit and greater potential risks, the good clinician managing most ICU patients may feel that the PAC is not necessary. Currently, the PAC is most commonly used in cardiogenic shock, cardiac surgery, and for evaluation and management of pulmonary hypertension, but there continues to be controversy regarding the indications for use of the PAC. There is a need for ongoing research to define the specific patient populations most likely to benefit, but a randomized controlled trial in general ICU patients is unlikely to provide meaningful guidance. TABLE 1. - Some Potential Reasons for the Decreased Use of Pulmonary Artery Catheters Randomized, controlled trials have failed to demonstrate a decrease in mortality The PAC was overused in patients who would not benefit Financial rewards and medicolegal issues supporting its use have been revisited Technological developments and more widespread use of echocardiography have decreased the need Other less invasive techniques have been developed to measure cardiac output Reduced use of the PAC below a minimum increases the risks associated with its use PAC = pulmonary artery catheter. The evolving story of the use of the PAC demonstrates well the purposes that Safar (1) foresaw for the Journal: The first part of the story is the description of the physiology of shock and the learning curve for critical care physicians and nurses to use this monitoring technique without causing catheter-related complications. The second part of the story is learning the limitations of PAC use, including appropriate knowledge and expertise in obtaining accurate information, as well as learning, which patients might benefit from management with the use of a PAC. The third part of the story is the decline in use of the PAC at least in part due to a failure to interpret basic cardiovascular physiology at the bedside. The fourth part is the development of understanding that a large randomized controlled trial might not be the right way to evaluate monitoring technology. After all, the PAC is not a therapy in itself but a tool to obtain information that might guide therapy. CCM has contributed new knowledge, reviews, a consensus statement, evaluation of the PAC and its use, and, perhaps most importantly, a forum for rich discussion of the controversies surrounding the use of the PAC, probably the most controversial monitoring tool we have used in critical care. Safar (1) would be proud of the contributions of the SCCM’s “new” Journal.