The Evidence-Based Character of the Pulmonary Artery Catheter (in Cardiac Patients)

Since the beginning of the 1990s, the term ‘evidence-based’ has been used like a magic formula. Generally speaking, evidence-based signifies a clinical practice proven by good scientific (experimental) research. A critical editorial about the pulmonary artery catheter was impetus for a literature search as to the evidence-based character of the pulmonary artery catheter [1]. Invasive hemodynamic monitoring of patients on a coronary or intensive care unit is, according to the American Association of Critical Care Nurses, a core intervention of the critical care nurse. Venous catheterization of the right auricle and ventricle under X-ray was first performed by Dexter et al. in 1945 [2]. This new diagnostic procedure was mainly used for studies of congenital heart diseases like atrial septal defect, ventricular septal defect and Tetralogy of Fallot. Two years later, Dexter et al. [3] discovered the technique of measuring the pulmonary capillary wedge pressure (PCWP). Because fully saturated blood could be withdrawn from the distal part of the pulmonary artery, Dexter demonstrated the absence of valves in the pulmonary venous system. The implication of this finding was that the PCWP was, in theory, equal to the end diastolic pressure within the left ventricle. With this diagnostic technique, which still is used today, diseases like heart failure, mitral valve stenosis can be elucidated.It took 23 years before Swan [4] found in 1970 a new technique of catheterization of the pulmonary artery with the use of a flow-directed balloon-tipped catheter. The Swan–Ganz catheter (as the pulmonary artery catheter is sometimes referred to) made it possible to perform right heart-sided catheterizations without the use of X-ray (e.g. at the bedside in the coronary or intensive care unit). From that moment, a paradigm shift took place in the use of the pulmonary artery catheter (PAC). No longer merely a diagnostic tool, it became a monitoring tool for evaluating response to therapies in patients with diseases such as heart failure. Heart failure has been an important indication for the use of a PAC. Table 1 summarizes the cardiac indications for use of the PA catheter in critically ill patients as mentioned in several cardiac nursing texts [5–8]. Yet in 1980—10 years after the invention of the PAC—there still had not been experimental research published to determine the relative risk/benefit ratio of the PAC. The use of any invasive technology such as the PAC, can cause complications as we all know [9]. Potential complications are listed in Table 2. However, not only specific complications can happen. In addition, invasion procedures themselves may also cause undesirable cardiovascular—presumably reflex—responses in certain people. Everyone is familiar with the tendency of certain patients to faint during the simple procedure of a needle puncture to obtain blood samples. In an editorial in the American Journal of Cardiology, Spodick [9] describes that the circulation is influenced by invasive techniques and the psychological and physiological—reflex—factors caused by the invasive intervention. In this editorial, Spodick argues for a controlled clinical trial of invasive diagnostic and monitoring methods that can themselves modify the patient's condition. Patients in Killip class I (without hemodynamic complications) are, according to Spodick, the best patients for such a study to determine the helpful and harmful effects of the PAC and to define indications and contraindications for this procedure. As far as known, such a study did not take place. However, several studies on the efficacy of the PAC have been done. In 1987, Gore et al. [10] published the results of an observational study examining time trends in the incidence rates as well as in-hospital and long-term case-fatality rates of patients hospitalized with acute myocardial infarction (MI). A part of this study examined the use of the PAC in approximately 3000 patients. This study was performed over a period of 10 years (1975, 1978, 1981 and 1984) at 16 different hospitals in the Worcester, Massachusetts Standard Metropolitan Statistical Area. An increase was found in the use of the PAC from 7.2% (1975) to 19.9% (1984). The PAC was used more in teaching hospitals than in non-teaching hospitals and was placed in patients with larger MIs. Indications to place a PAC were in 96% of the cases congestive heart failure (CHF), cardiogenic shock or hypotension. In the group of patients in cardiogenic shock who received a PAC, the mortality did not differ from the group who did not receive a PAC. In both groups of patients with CHF and hypotension, the mortality rate was greater in the group in which a PAC was placed (P=0.001). The use of the PAC was also associated with an increase length of hospital stay. Although not significant, there was shown a better long-term survival (6 months and 5 years) in patients with CHF and cardiogenic shock who did receive a PAC. Gore et al. concluded that they could not demonstrate a statistically significant beneficial effect associated with the use of the PAC in the patients studied. A non-experimental study by Zion et al. [11] in 1990 showed similar results. In this study, performed in Israel, 5800 patients were included with an MI. A total of 371 patients received a PAC. Mortality in patients receiving a PAC was 59.4% vs. 33.5% for those without a PAC. In a subgroup analysis of patients with CHF with pulmonary oedema, the mortality was 47.8% in the group that received a PAC vs. 36.4% in the group that did not receive the catheter (P=NS). Table 3 shows the possible epidemiological implications of these data. The number needed to harm (NNH) is 9, which means that in this group of patients from nine patients who received a PAC, one of them dies secondary to complications of the PAC. However, we should be careful with these figures. This subgroup analyses was done on a very small group of patients: n=174 of which 67 patients did receive a PAC and 107 did not. It is dangerous to generalize from a small sample size to the population. Zion was muted in his conclusion. He theorized that it was unlikely that the PAC itself increased mortality. Patients receiving a PAC were more likely severely ill and therefore had a greater mortality rate. He also concluded that the mortality rate in patients without CHF who received a PAC was indeed higher than in similar patients who did not receive the catheter, but that the greater mortality was not caused by the catheter itself. The causes of death in these patients were electro-mechanical dissociation, ventricular septal rupture and brain death after resuscitation. It is remarkable that these mild conclusions of Zion rarely are cited in the literature after his publication. In later literature, the high difference in mortality is more frequently remarked upon. In 1991, 20 years after Swan discovered the flow directed pulmonary artery catheter, Guyatt et al. [12] undertook a randomized controlled trial of right heart catheterization in critically ill patients in Canada on two intensive care units. Patients included in the study were suffering from hypoxemia, hypotension and oliguria. Also, patients, not meeting these criteria, could be included when the attending physician believed the patient might benefit from PAC. One criterion for exclusion was the feeling on the part of the clinician that inserting a PAC was an ethical imperative. From among 1237 patients, there were 148 patients who met the inclusion criteria. From these 148 patients, 115 were not included of which 52 were excluded on the basis that the physician felt it was unethical not to insert the catheter. Ultimately, 33 patients were included in this study. This small number of patients was the reason the study was discontinued because funding was not renewed. The preference to insert a PAC itself appeared to be a more compelling rationale for the study than the determination of the benefit/risk ratio of that very same PAC. Also, in 1991, an expert panel from the European Society of Intensive Care Medicine formulated recommendations for the use of the PAC [13]. Indications for PAC use in cardiac patients are defined as follows: (1) cardiogenic shock (to measure CO and monitoring response to therapy), (2) severe, acute left ventricular failure (to assist in the diagnosis of acute mitral valve regurgitation or septal perforation), (3) acute right ventricular failure due to right ventricular infarction, (4) pulmonary oedema when there is doubt concerning the etiology (i.e. cardiogenic or not) and (5) cardiac tamponade when echocardiographic assessment is not available. Although the indications are well-founded, the exact mechanism, which led to defining these indications, is not described. Many questions remained unanswered. For example, no definitive answer was given to the question of whether prospective studies are necessary. Also, according to the judgment of the panel, the following questions must be answered: have there been an adequate collection and interpretation of hemodynamic data, is there consensus with regard to the therapy when the hemodynamic situation of the patient changes and is there a therapy adjustment that positively influences the hemodynamic status of the patient? With the study ‘The effectiveness of right heart catheterization in the initial care of critically ill patients’ Conners et al. [14] put the cat among the pigeons. In a prospective, non-experimental cohort study of more than 5700 patients, with nine different illnesses, approximately 2100 patients received a PAC. The conclusion of Conners was that patients receiving a PAC had a higher mortality and the severity of the illness alone did not account for that mortality increase. In an editorial accompanying this publication, a plea was made for multicenter, randomized controlled trials. Furthermore, until such a trial is been undertaken, it is time to pull the PAC from the armentarium of clinical tools and to issue a moratorium on its use [1]. Much controversy followed the publication of Conners' study [15,16]. The propensity score Conners used should be considered objectively and not compared with the reason either to insert or not to insert the catheter. There are published pre- and perioperative randomized studies which demonstrate benefits from the use of the PAC, Conners did not include patients in his study with the diagnosis of acute MI, among other patient groups (e.g. major trauma, fire injuries and major surgeries). Furthermore, the hazard of death was only 1.02 in patients with CHF. Furthermore, Conners did not comment on the knowledge and skills of the physicians and nurses obtaining and interpreting the PA catheter data, the specific therapies that were given, nor the causes of death. So far, this was the chief criticism of the Conners study. As mentioned earlier, the knowledge and skills of physicians and nurses concerning hemodynamic monitoring is critical to the efficacy of the PA catheter and patient outcomes. This aspect has been studied by Iberty [17,18] in the United States of America and by Gneagi [19] in Europe (France and the French speaking parts of Belgium and Switzerland). These studies were performed via a multiple-choice examination. Several aspects of hemodynamic monitoring were examined and these aspects are listed in Table 4. The conclusions of these three studies were in agreement. Knowledge concerning the PAC and hemodynamic monitoring shown to be extremely variable among physicians as well as nurses. Lack of knowledge can lead to misinterpretation of the collected data and to a wrong choice of therapy, which can increase patient's morbidity and mortality. Overall the nurses had a lower knowledge score than the physicians. However, concerning practical skills (e.g. interpretation of the pulmonary artery occlusion pressure and waveforms), the nurses performed slightly better. Physicians performed better in hemodynamic calculations and physiological interpretation. There was no difference in the results between the studies in the USA and in Europe. Variables that influenced the scores are mentioned in Table 5. Experience and level of training seems to be an import variable for a better score. It is important for nurses who work with hemodynamic monitoring to recognize that the physician's therapeutic decisions are very often based on the data collected by the nurses. Thus, indirectly, the nurse is also responsible for the choice of therapy! The authors point out that a misinterpretation of data or inability to recognize complications may possibly lead to worse patient outcomes. Recommendations have been put forward by the authors regarding credentialing nurses as competent to use the PAC and improving training and education of both nurses and physicians. Research on the effect of more aggressive efforts to improve tissue cardiac output and oxygen delivery and consumption in surgical patients who received a PAC, was done by Hayes et al. [20]. This prospective, randomized controlled study looked at a heterogeneous group of critically ill patients (n=109). Patients were included in whom, after receiving fluid administration, the following therapeutic goals were not achieved: a cardiac index (CI)≥4.5 l min−1 m−2, oxygen delivery ≤600 ml−1 min−1 m−2 and an oxygen consumption ≥17 ml−1 min−1 m−2. All patients received intravenous dopamine and norepinephrine. In the intervention group, dobutamine was administered to increase CI, oxygen delivery and consumption. In the control group dobutamine was administered when CI was less then 2.8 l min−1 m−2. To sum up, the three groups of patients could be categorized in this way: a group, which was not randomized, the intervention group and the control group. The group that was not randomized achieved the target therapeutic goals after fluid resuscitation alone. There was no mortality in this group of nine patients. In the intervention group (in which the patients received dobutamine, n=50), mortality was found to be higher than in the control group (n=50). Mortality in the intervention group was 50% vs. 30% in the control group (P=0.04). Also, oxygen consumption decreased in the intervention group while oxygen delivery increased. The explanation for this seemingly paradoxical finding, according to Hayes et al., was that the intervention therapy exacerbated the maldistribution of blood flow within the microcirculation, resulting in impaired perfusion of vital organs. This may also explain the higher incidence of multiple-organ-failure in the treatment group. Hayes concluded that the use of dobutamine to boost the CI and systemic oxygen delivery failed to improve the outcome in this group of patients. Contrary to what might have been expected, the results suggest that, in some cases, aggressive efforts to increase oxygen consumption may have been detrimental. A limitation of this study is that the results may not be generalizable to other patient groups for example, cardiac patients. Nonetheless, the results of this study should be taken in consideration when hemodynamic data leads to the decision to start a more aggressive therapy. The ongoing vigorous debate about the indications for and clinical utility of the PAC was the reason for a ‘Pulmonary Artery Catheter Consensus Conference’ in 1997. The purpose of the conference was to give an answer to the following questions: (a) review the state of knowledge concerning the PAC in specific patient populations; (b) identify specific critical unanswered questions related to the PAC in these settings; (c) seek consensus on these unanswered questions; and (d) make recommendations for clinical practice and for future clinical and epidemiological research in this area [21]. In the Consensus statement the evidence-based character of the use of the PAC in several patient populations is described. Five levels of evidence are distinguished (ranging from I to V). Level I are large randomized trials with clear-cut results, level V are case series, uncontrolled trials and experts opinions (Table 6.) These levels of evidence are used to grade the responses on questions. When an answer is given grade A, it means that the answer is supported by at least two level I investigations, when the answer is given grade E it means the answer is supported by level IV or level V evidence. When we look to the (nonsurgical) cardiovascular diseases, it is remarkable that the question ‘Does management with the pulmonary artery catheter improve outcome in patients?’ are consistently graded with an E. Congestive heart failure is graded with a D (Table 7.) Across all disease states, further randomized controlled studies of PAC efficacy are recommended. In the final recommendations the Conference stated that, although there is no reason for a moratorium, the clinicians should continue to carefully weigh the risks and benefits of the PAC. In addition, criteria for the appropriate use of the PAC should be developed and the knowledge of the clinician about use of the PAC and potential complications should be improved. More than half a century after Dexter performed the first right heart catheterization under X-ray, this technique is still a good diagnostic tool. Thirty years after Swan discovered the flow-directed PAC and the continuous monitoring of the hemodynamic status of the patient, it is still an item of discussion and the question ‘How evidence-based is the PAC?’ has not yet been properly answered. Even in the third millennium, research to answer this question is going on. Murdoch et al. [22] concluded in a trial (performed in the UK), similar to Conners' trial in 1996, that the use of the PAC is indeed safe but also that no substantial benefit was demonstrated. He also suggested that a large randomized controlled trial of PACs would have insufficient power to demonstrate an effect on mortality. In the British Journal of Medicine of February 2001 a large randomized trial was announced [23]. This ‘UK Pulmonary Artery Catheter Management Trial’ is a randomized controlled trial and the aim is to include 6000 patients. The results were expected in 2002 but to date, no final results or interim report has been published. A randomized controlled trial of the pulmonary artery catheter by Rhodes et al. [24] in 201 critically ill patients with different diseases was performed in the UK from 1997 to 1999 and the results were recently published. The results of this study suggest that the PAC is not associated with an increased mortality. At least, Sandham et al. [25] published in January 2003 the results of a randomized controlled trial of the use of the PAC in high-risk surgical patients. They found no benefit to therapy directed by PAC over standard care in the study population. Among patients in NYHA class III or IV in this study, in-hospital mortality was 13.8% in the standard-care group and 18.6% in the catheter group. This was not statistically significant. It is still an ongoing process to define the evidence-based character of the PAC. From observational studies in the 1980s there now seems to be a cautious start with randomized controlled trials. This has proven to be a difficult trial to conduct in the healthcare arena. Recall that in the study of Guyatt et al. (which failed) there were patients excluded because of physicians' belief that it was unethical not to utilize a PAC for patient care; this several in the trial of Rhodes Furthermore, in evaluating studies about the efficacy of the PAC, it is important to look not only at the different patient but also to take consideration the in between the USA and Europe. but not least, that evidence-based is based on three of In his in evidence-based describes these three (1) the evidence (which is not the same as (2) the preference of the patient and the clinician and (3) the of the these may the discussion of ‘evidence-based’ The increase in the use of PACs in the and 1980s by Gore seems to a technique by which and left ventricular can be and the better treatment for patients with an acute MI, the of use of the PAC. the controversy regarding the therapeutic benefit vs. of the of the PAC no doubt although less than a The to for his and during the of this