Pulmonary Artery Catheter

special reports Pulmonary Artery Catheter* Does the Problem Lie in the Users? Pierre Squara, MD; David Bennett, MD; and Claude Perret, MD

The aims of this study were to look for the variability in the treatment of circulatory shock and to assess the extent to which this variability was reduced by pulmonary artery catheterization (PAC). At three international conferences in 1997–1998 (European Society of Critical Care Medicine, French Language Society for Critical Care [Socie´te´ de Re´animation de Langue Franc¸aise], and Society of Critical Care Medicine), a real-life clinical case was discussed in meetings among physicians and a panel of experts, with assistance from an expert computer program. A total of 417 physicians took part in the discussions. Following the clinical presentation, only 38% of physicians suggested the same treatment as the experts, and 35% suggested potentially harmful treatments. Complete hemodynamic data from PAC significantly decreased the range of suggested treatments, improved agreement among physicians themselves as well as the agreement between physicians and experts, and decreased the number of potentially harmful propositions. However, whereas almost 80% of participants finally agreed on the treatment after one to three invasive hemodynamic sets of measurements, at least 10% persisted in suggesting potentially harmful treatments. PAC improved interphysician agreement, but our data suggest that yet greater agreement could be achieved by improving the theoretical training of practitioners. (CHEST 2002; 121:2009 –2015) Key words: hemodynamics; medical intelligence; right-heart catheterization Abbreviations: ESICM ⫽ European Society of Critical Care Medicine; PAC ⫽ pulmonary artery catheterization; PEEP ⫽ positive end-expiratory pressure; PWP ⫽ pulmonary wedge pressure; SCCM ⫽ Society of Critical Care Medicine; SRLF ⫽ Socie´te´ de Re´animation de Langue Franc¸aise; Svo2 ⫽ mixed venous saturation

⬎ 3,000 publications in the medical A lthough literature focus on pulmonary artery catheteriza-

tion (PAC) and ⬎ 45 million pulmonary artery catheters have been used since 1970, the method remains a matter of controversy.1,2 There is strong evidence that PAC is helpful in the management of circulatory disorders,3–5 especially when continuous monitoring is required.6,7 However, the lack of controlled studies, due to numerous methodologic difficulties,5,8,9 precludes a confident evaluation of the beneficial impact of PAC on survival. Four studies, all with serious methodologic weaknesses, found that *From Re´animation polyvalente, Hoˆpital Victor Dupouy, Argenteuil, France. Supported by a grant from Baxter–Critical Care Division. Manuscript received May 8, 2001; revision accepted October 24, 2001. Correspondence to: Pierre Squara, MD, Re´animation polyvalente, Hoˆpital Victor Dupouy, 69, rue du lieutenant Colonel Prudhon 95100, Argenteuil, France; e-mail: [email protected] www.chestjournal.org

PAC was associated with greater mortality.10 –13 Because of these uncertainties, and as part of a trend calling for evidence-based medicine and cost control, current practices regarding PAC have been challenged.2,14 Controlled studies are required to resolve this issue. Nevertheless, when analyzing the potential adverse effects of PAC, it is necessary to determine what is due to the tool and what is due to misuse of the tool.15 Ways to improve PAC use have been periodically reviewed.16 –18 Benefits may be greater when PAC is used by physicians who have the theoretical knowledge needed to interpret PAC data in an optimal manner.19,20 PAC may have fewer adverse effects in the hands of clinicians who follow guidelines and are proficient in interpreting PAC data.18,21 There are many reasons to believe that the reproducibility of diagnoses based on hemodynamic data are poor even among experienced intensivists.19,20,22 CHEST / 121 / 6 / JUNE, 2002

2009

It follows that physicians may differ regarding the management they feel is optimal in a given patient with or without PAC. We designed a study to look for variability in opinions about the best treatment in a real-life case and to assess the extent to which this variability was reduced by PAC.

Materials and Methods On three different occasions, we asked physicians attending a conference and experts to discuss a real-life clinical case, and also noted diagnoses and treatment objectives suggested by an expert computer program (Hemodyn; P. Squara; Enghien, France) during the meetings.22,23 The meetings took place as part of the official program of three international conferences held in 1997– 1998: the 10th meeting of the European Society of Critical Care Medicine (ESICM; Paris, France; September 6, 1997), the 26th meeting of the French Language Society for Critical Care (Socie´ te´ de Re´ animation de Langue Franc¸ aise [SRLF]; Paris, France; January 26, 1998), and the 27th meeting of the Society of Critical Care Medicine (SCCM; San Antonio, TX; February 6, 1998). Physicians attending these conferences were invited to register for the meetings, and during the meetings an interactive voting system was used to capture all participant answers in a computer. After a 10-min introduction, including a brief presentation of the experts (listed in the Appendix) and of the expert Hemodyn software, participants were familiarized with the interactive voting system using three questions with a limited list of responses displayed on a screen: (1) what is your homeland (one answer allowed), (2) what is your main medical specialty (one answer allowed), and (3) what is/are, in your opinion, the most important PAC parameter(s) [one or more answers allowed]? The results of the voting were displayed immediately. A real-life case was then presented. This case was the same at the three meetings and was retrospectively considered of low complexity by the experts. It was a typical history of acute hypertensive pulmonary edema in a patient with normovolemic hypertensive cardiomyopathy in whom excessively aggressive emergency treatment induced critical hypovolemia and venous vasodilatation. The case presentation was displayed on a screen using a video projector, and the computer captured all radio signal-transmitted answers. Participants were asked to vote after the presentation of clinical data (Table 1), and after the subsequent presentation of PAC data recorded 2 h, 14 h, and 27 h after hospital admission of the patient (PAC time points 1, 2, and 3, respectively; Table 2). Thus, votes were recorded on four occasions. On each of these occasions, participants were asked to choose one or more treatments among nine possibilities. The treatment actually administered and the clinical course between the two PAC time points were described (Table 3); based on this information, the participants were asked to suggest treatment changes. After each vote, the opinions of experts were displayed. The objectives and diagnoses generated by Hemodyn program were then displayed and freely discussed by the experts and participants. Finally, the experts were asked to classify the treatments suggested by the participants as “acceptable” or “potentially harmful.” Data Analysis Although the ␬ concordance test allows standardization of agreement between two judges, no statistical tool is available for standardizing agreement between two groups of judges. Consequently, we expressed between-group agreement as the ratio of 2010

Table 1—Summary of the Clinical Presentation* Medical history Age of 70 yr Chronic arterial hypertension with hypertensive cardiomyopathy Several episodes of acute pulmonary edema Sudden major nocturnal dyspnea First medical examination (at home) Acute pulmonary edema; Spo2 of 75% No peripheral edema; no fever BP, 190/120 mm Hg Initial treatment (at home) Sedation, intubation, and mechanical ventilation IV bolus of 40 mg furosemide IV bolus of 1 mg nitrate followed by a continuous IV infusion of 1 mg/h 10 ␮g/kg/min of dobutamine Admission to the emergency department Clinical signs of shock; oliguria BP, 90/60 mm Hg; heart rate, 120 beats/min; nitrate infusion stopped Chest radiograph showing bilateral infiltrates Pao2, 83 mm Hg; Sao2, 96% with Fio2 of 1; lactate, 3.4 mEq/L Electrocardiography showing sinus tachycardia; otherwise normal Echocardiography showing LV hyperkinesia and hypertrophy; no dilatation; RV normal *Spo2 ⫽ pulse oximetric saturation; Sao2 ⫽ arterial oxygen saturation; Fio2 ⫽ fraction of inspired oxygen; LV ⫽ left ventricular; RV ⫽ right ventricle.

the number of concordant votes over the total number of votes. The variability in suggested treatments was expressed as the proportion of participants who selected each answer. Categories (participants, characteristics, and suggested treatments) were compared using the ␹2 test. We used Bonferonni’s correction to compensate multiple comparisons.

Results A total of 560 physicians attended the meetings: 248 physicians at the ESICM meeting, 62 physicians at the SRLF meeting, and 250 physicians at the SCCM meeting. The small number of physicians at the SRLF meeting was due to a technical problem that delayed the meeting by 1 h. Of these 560 physicians, 417 physicians participated in the voting: 167 physicians at the ESICM meeting, 58 physicians at the SRLF meeting, and 192 physicians at the SCCM meeting. These 417 participants were from 29 countries, most of which were in Western Europe (n ⫽ 199), North America (n ⫽ 132), Eastern Europe (n ⫽ 20), Northern Europe (n ⫽ 18), Japan (n ⫽ 17), and South America (n ⫽ 14). Their areas of interest were distributed as follows: critical care (n ⫽ 220), anesthesiology (n ⫽ 91), pulmonology (n ⫽ 38), cardiology (n ⫽ 27), emergency medicine (n ⫽ 21), pediatrics (n ⫽ 7), and others (n ⫽ 13). Cardiac output was believed by the vast majority Special Report

Table 2—Summary of the Hemodynamic Data 2 h After Hospital Admission* PAC Time Points† Variables Hemodynamics RAP, mm Hg PWP, mm Hg PAPm, mm Hg Pam, mm Hg HR, beats/min CI, L/min/m2 Blood gas analysis Fio2, % Hb, g/dL Pao2, mm Hg Sao2, % Pvo2, mm Hg Svo2, % C(a-v)O2, % V˙o2, mL/min/m2 pH Paco2, mm Hg HCO⫺ 3 , mEq/L Lactate, mmol/L

1

2

3

4 4 12 70 123 2.04

7 9 19 90 110 2.19

10 14 20 80 99 2.40

1 14.2 111 0.99 27 0.51 9.4 191 7.31 41 21 4.1

1 13 157 1 38 0.58 7.7 165 7.38 38 23 1.9

0.5 12.8 145 1 45 0.69 5.6 135 7.41 36 23 0.7

*RAP ⫽ right arterial pressure; PAPm ⫽ mean pulmonary artery pressure; Pam ⫽ mean artery pressure; HR ⫽ heart rate; CI ⫽ cardiac index; Hb ⫽ hemoglobin; Pvo2 ⫽ mixed venous oxygen tension; C(a-v)O2 ⫽ arteriovenous oxygen content difference; V˙o2 ⫽ oxygen uptake. See Table 1 for expansion of other abbreviation. †The three PAC time points were at 2 h, 14 h, and 27 h, respectively, after hospital admission.

of participants to be among the most important hemodynamic parameters (n ⫽ 330, 79%). Pulmonary wedge pressure (PWP) was selected by 285 participants (68%), mixed venous saturation (Svo2) by 220 participants (53%), pulmonary artery pressure by 127 participants (30%), stroke volume by 100 participants (24%), right atrial pressure by 20 participants (5%), right ventricular ejection fraction by 20 participants (5%), and right ventricular end-diastolic volume by 18 participants (4%). No differences were found according to the country of origin or area of interest, except for the cardiologists, among whom

100%, 75%, 38%, 38%, and 13% believed felt that pulmonary capillary wedge pressure, cardiac output, pulmonary artery pressure, Svo2, and stroke volume were among the most important parameters, respectively (p ⬍ 0.05). None of the other variables were selected as important by the cardiologists. The treatments suggested by the participants and the experts before the presentation of PAC data and after each of the first PAC time points are summarized in Tables 4, 5. The mean number of treatments suggested by each physician (excluding “do not know” answers) decreased significantly after PAC: 1.3 treatments vs 2.0 treatments before PAC (p ⬍ 0.05). There were no significant differences in the proportion of suggested treatments across different meetings, country of origin, or area of interest. After the second and the third PAC time points (Tables 6, 7), there was a therapeutic consensus among experts. Participants agreement with this expert consensus was high; however, the differences in suggested treatments between experts and participants remained significant and the proportion of potentially harmful treatments remained high (⬎ 10%). Discussion We found considerable variability in the initial treatments suggested by physicians for a patient investigated using noninvasive means, including echocardiography. Only 38% of nonexperts suggested fluid administration, which was the treatment recommended by all the experts (100%). Our method did not enable us to identify the participants who suggested each treatment and, therefore, did not provide information on treatment combinations. However, ⬎ 35% of participants

Table 4 —Dispersion of Suggested Treatments Before PAC Insertion* Variables

Table 3—Summary of Course and Treatments Administered Treatment after the first PAC time point 500 mL of crystalloid Decrease in the dobutamine infusion rate from 10 to 5 ␮g/kg/min Treatment after the second PAC time point Additional 500 mL of crystalloid Dobutamine infusion stopped Nicardipine infusion started at 1 mg/h Treatment after the last PAC time point No change; weaning from mechanical ventilation

www.chestjournal.org

ESICM SRLF SCCM Experts

Potentially harmful suggestions Diuretics Dobutamine 1 Vasodilators PEEP Acceptable suggestions Do not know Vasopressors Dobutamine 7 Dobutamine 2 Filling

25 36 17 40

17 21 17 34

23 29 18 31

0† 0† 0† 10†

27 25 13 16 36

29 36 12 16 43

23 23 21 18 39

0† 5† 0† 20 100†

*Data are presented as %. 1 ⫽ increase titration; 2 ⫽ decrease titration; 7 ⫽ no change in titration. †p ⬍ 0.05 for intergroup comparisons. CHEST / 121 / 6 / JUNE, 2002

2011

Table 5—Dispersion of Suggested Treatments After PAC Time Point 1* Variables Potentially harmful suggestions Diuretics Dobutamine 1 Vasodilators PEEP Acceptable suggestions Do not know Vasopressors Dobutamine 7 Dobutamine 2 Filling

ESICM SRLF SCCM Experts 3‡ 10‡ 7‡ 17‡

2‡ 9‡ 7‡ 14

2‡ 8‡ 8‡ 13‡

0 0 0 0

2‡ 8‡ 6‡ 22 56‡

0‡ 12‡ 9 22 67

0‡ 7‡ 6‡ 22 66‡

0 0 0 20 100†

*Data are presented as %. See Table 4 for definitions. †p ⬍ 0.05 for intergroup comparisons. ‡p ⬍ 0.05 for comparisons between Table 4 (before PAC) and Table 5 (after PAC).

suggested at least one potentially harmful treatment. In particular, 35% of participants suggested positive end-expiratory pressure (PEEP), which was the potentially harmful treatment suggested most often. Providing PAC data in addition to the clinical presentation and echocardiography findings had the following effects: (1) the range of suggested treatments became narrower, (2) agreement between participants and experts improved, (3) agreement among participants improved, and (4) the proportion of potentially harmful treatments decreased. However, the proportion of participants who suggested at least one potentially harmful treatment remained ⬎ 15% after PAC time point 1, ⬎ 10% after PAC time point 2, and equal to 10% after PAC time point 3. Our results are in agreement with previous findings. It has been shown that PAC data are more accurate than clinical assessment in evaluating the hemodynamic status of patients,24 –26 that PAC data

Table 6 —Dispersion of Suggested Treatments After PAC Time Point 2* Variables Potentially harmful suggestions PEEP Diuretics Dobutamine 1 Dobutamine 7 Vasopressors Do not know Acceptable suggestions Filling Dobutamine 2 Vasodilators

ESICM SRLF SCCM Experts 12 1 7 15 7 7

9 0 10 15 9 9

8 3 13 13 5 8

0 0 0 0† 0 0

75 31 17

78 24 17

76 24 13

100† 20 20

*Data are presented as %. See Table 4 for definitions. †p ⬍ 0.05 for intergroup comparisons. 2012

Table 7—Dispersion of Suggested Treatments After PAC Time Point 3* Variables Potentially harmful suggestions Diuretics Dobutamine Vasopressors Do not know Acceptable suggestions Filling Vasodilators 1 Vasodilators 7 Vasodilators 2 PEEP

ESICM SRLF SCCM Experts 10 3 0 2

14 5 0 5

8 3 0 1

0† 0 0 0

9 12 69 15 0

9 17 60 17 0

8 18 63 13 0

0 0 90† 10 0

*Data are presented as %. See Table 4 for definitions. †p ⬍ 0.05 for intergroup comparisons.

may be more relevant than clinical investigation,23,27–30 and that PAC leads to major changes in therapeutic strategies.31,32 Our findings raise several questions. The first issue is the extent to which our participants were representative of the overall population of physicians who use PAC. We had no indication on the credentials of participants and/or on the number of times they used PAC annually. However, all participants identified themselves as practitioners, suggesting that they used PAC in their everyday practice. In addition, we found no differences in suggested treatments according to country of origin and/or area of interest. Last, it is unlikely that physicians who go to international meetings form a subgroup with less expertise than the overall PAC user population. Thus, our cohort of participants is probably more representative of real PAC users than investigators of controlled studies that most often represent a minority of international experts. A second point worthy of discussion is whether the data provided to the participants were sufficient to enable a full understanding of the case. All relevant data about the medical history, occurrence of the acute episode, and initial clinical examination were carefully explained, as were all findings from an extensive evaluation including lactate, blood gas analysis, and echocardiography. Participants were told at the beginning of the presentation that the presentation described a simple, real-life case, devoid of unusual complexity. We used slides, a method to which international conference participants are accustomed and believed to be receptive. In addition, the time spent on presenting the case was probably longer than the time usually spent on reviewing PAC data at the bedside. Third, to provide accurate information on how well physicians use PAC data, our case had to be Special Report

representative of the cases in which PAC is commonly used. The patient had a circulatory disorder requiring emergency treatment. There was a combination of shock and acute pulmonary edema, a situation that raises therapeutic challenges particularly with regard to whether fluid administration is warranted, and in which the use of PAC is widely accepted.5 Although, most experts immediately made the correct diagnosis and all suggested fluid administration, 10% also suggested PEEP, which was finally considered potentially harmful. The echocardiographic data from the patient were described to the participants and experts. Despite the evidence of hypertrophic cardiomyopathy with limited heart filling, it became obvious during the debate that many participants were reluctant to administer fluid to patients with severe acute pulmonary edema without prior PWP measurement and without coupled monitoring of PWP and cardiac output during a filling test. This observation illustrates the importance ascribed by physicians to PWP measurement as a valuable tool for medical decision making, although the limitations of this measurement are well known.17 The reasons for the significant level of persisting disagreement between participants and experts regarding the appropriate treatment deserve discussion. It has been shown that many PAC users have insufficient theoretical knowledge of hemodynamics20,33 and that data collection is frequently inadequate.19 However, the disagreement about the simple case presented in our study cannot be explained by mistakes or by a misunderstanding of some formulas such as resistance or work indexes. Rather it suggests poor reasoning. Although representative of PAC users, participants to our study were probably not all experienced intensivists. The answers to the question on which parameter(s) are most important support this hypothesis. The majority of participants selected cardiac output and PWP as the most important hemodynamic variables; few believed that “all variables” were important. Experts usually make a systematic analysis of a specific observed situation, then try to fit this analysis to several theoretical pathophysiologic models. They also test the fit of the case to the model by looking for additional relevant variables. There is strong evidence that expert problem solving in medicine is dependent on prior specific experience and on elaborated conceptual knowledge applicable to the occasional problematic situation.34 In contrast, “nonexperts” proceed with basic pattern recognition using few variables and limited knowledge, an approach that carries a large risk of error. In addition, when nonexperts www.chestjournal.org

correctly identify the appropriate pathophysiologic model, they may encounter difficulties in determining the appropriate hemodynamic objective in a given patient and situation (ie, the cardiac output needed, the oxygen uptake needed, etc.). This problem of interpractitioner variability is not limited to PAC but has been observed with many other diagnostic tools, such as radiography,35 CT,36,37 angiography,38,39 and echocardiography.40 For these tools, various computerized signal processing techniques have been developed to decrease interobserver variability. Use of a computer program to assist in PAC data interpretation is a similar approach.22 Some prospective randomized studies are expected to investigate the impact of PAC on mortality. Variability in PAC use may have a major influence on the results of these studies. In our study, only 80% of PAC-guided strategies provided patients with a chance of improvement, 10% had no effect, and 10% were harmful. A risk ratio of 0.8 for PAC-guided therapeutic strategies compared with non-PACguided strategies has been reported.8 If we hypothesize that treatments with no effects yield a risk ratio of 1 and harmful treatments yield a risk ratio of 1.2, an heterogeneity in the treatments as seen in our study may cause a 30% decrease in the potential beneficial effects of PAC on mortality. In addition, it is reasonable to hypothesize that the side effects of PAC are inversely related to the benefits, ie, that nonexperts obtain fewer benefits and more side effects. This illustrates the importance of defining criteria for optimal PAC use before starting multicenter studies. Thus, ⬎ 25 years after the introduction of PAC and despite thousands of scientific publications, our data showed unacceptable variability in treatments and an alarmingly high rate of potentially harmful treatment decisions in participants to three major international meetings in intensive care medicine. Postgraduate training authorities should take this finding into account. Current and future studies aimed at determining the value of PAC should address the issue of inadequate physician knowledge about PAC data interpretation. A recent report of the National Heart, Lung, and Blood Institute and the US Food and Drug Administration has emphasized the need for collaborative education of physicians and nurses in performing, obtaining, and interpreting information from the use of PAC.41 The use of a computer program that assists in the interpretation of hemodynamic data has been proposed to improve treatment decisions homogeneity.23 CHEST / 121 / 6 / JUNE, 2002

2013

ACKNOWLEDGMENT: We thank Dr. Patrick Godard for assistance, Professor Didier Payen for reviewing the article, and all the participants.

Appendix Study Participants ESCIM Meeting: Professor Antonio Artigas (Spain), Professor David Bennet (United Kingdom), Professor Alfred Connors (United States), Professor Didier Payen (France), Professor Claude Perret (Switzerland), Professor Peter Radermacher (Germany), Professor Heinz Steltzer (Austria), and Professor Jean Louis Vincent (Belgium). SRLF Meeting: Professor Jean Franc¸ ois Dhainaut (France), Professor Claude Martin (France), Professor Didier Payen (France), Professor Claude Perret (Switzerland), and Professor Jean Louis Teboul (France). SCCM Meeting: Professor David Bennet (United Kingdom), Professor Elaine Daily (United States), Professor Loren Nelson (United States), Professor Didier Payen (France), Professor William Peruzzi (United States), Professor William Sibbald (Canada), and Professor Jean Louis Vincent (Belgium).

References 1 Vincent J, Dhainault J, Perret C, et al. Is the pulmonary artery catheter misused? A European view. Crit Care Med 1998; 26:1283–1287 2 Sandham JD, Hull RD, Brant RF. The pulmonary artery catheter takes a great fall. Crit Care Med 1998; 26:1288 –1289 3 Swan H, Ganz W. Hemodynamic measurements in clinical practice: a decade in review. J Am Coll Cardiol 1983; 1:103–113 4 Sprung C, Eidelman L. The issue of a US Food and Drug Administration moratorium on the use of the pulmonary artery catheter. New Horiz 1997; 5:277–280 5 Pulmonary Artery Catheter Consensus conference: consensus statement. Crit Care Med 1997; 25:910 –925 6 Boldt J, Menges T, Wollbruck M, et al. Is continuous cardiac output measurement using thermodilution reliable in the critically ill patient? Crit Care Med 1994; 22:1913–1918 7 Parker MM, Peruzzi W. Pulmonary artery catheters in sepsis/ septic shock. New Horiz 1997; 5:228 –232 8 Ivanov R, Allen J, Sandham J, et al. Pulmonary artery catheterization: a narrative and systematic critique of randomized trials and recommendations for the future. New Horiz 1997; 5:268 –276 9 Sibbald W, Keenan S. Show me the evidence: a critical appraisal of the Pulmonary Artery Catheter Consensus conference and other musings on how critical practitioners need to improve the way to conduct business. Crit Care Med 1997; 25:2060 –2063 10 Connors A, Speroff T, Dawson N, et al. The effectiveness of right heart catheterization in the initial care of critically ill patients. JAMA 1996; 276:889 –918 11 Gore J, Goldberg R, Spodick D, et al. A community-wide assessment of the use of pulmonary artery catheters in patients with acute myocardial infarction. Chest 1987; 92: 721–727 12 Wu A, Rubin H, Rosen M. Are elderly people less responsive to intensive care. J Am Geriatr Soc 1990; 38:621– 627 13 Zion M, Balkin J, Rosenmann D, et al. Use of pulmonary artery catheterization with acute myocardial infarction: analysis of experience in 5,841 patients in the SPRINT Registry. Chest 1991; 101:589 –590 2014

14 Robin ED. The cult of the Swan Ganz catheter. Ann Intern Med 1985; 103:445– 449 15 Naylor C, Sibbald W, Sprung C, et al. Pulmonary artery catheterization: can there be an integrated strategy for guideline development and research promotion? JAMA 1993; 2407–2411 16 Feihl F, Perret C. Right heart catheterization at bedside: a note of cautious optimism. Intensive Care Med 1995; 21:296 – 298 17 Jardin F, Bourdarias J. Right heart catheterization at bedside: a critical review. Intensive Care Med 1995; 21:291–295 18 American Society of Anesthesiologists task force on pulmonary artery catheterization: practice guidelines for pulmonary artery catheterization. Anesthesiology 1993; 78:380 –394 19 Iberti T, Fischer E, Leibowitz A, et al. A multicenter study of physician’s knowledge of the pulmonary artery catheter. JAMA 1990; 12:2933–2940 20 Gnaegi A, Feihl F, Perret C. Intensive care physicians’ insufficient knowledge of right heart catheterization at the bedside: time to act? Crit Care Med 1997; 25:213–220 21 European Society of Intensive Care Medicine expert panel. The use of the pulmonary artery catheter. Intensive Care Med 1991; 17:1–7 22 Squara P, Dhainaut J, Lamy M, et al. Computer assistance for hemodynamic evaluation. J Crit Care 1989; 4:273–282 23 Squara P, Journois D, Formela F, et al. Value of elementary, calculated and modeled hemodynamic variables. J Crit Care 1994; 9:223–235 24 Forrester J, Diamond G, Swan H. Correlative classification of clinical and hemodynamic function after myocardial infarction. Am J Cardiol 1977; 39:177–189 25 Connors A, Dawson N, Mac Caffe D, et al. Assessing hemodynamic status in critically ill patients: do physicians use clinical information optimally? J Crit Care 1987; 2:174 –180 26 Eisenberg P, Jaffe A, Shuster D. Clinical evaluation compared to pulmonary artery catheterization in the hemodynamic assessment of critically ill patients. Crit Care Med 1984; 12:549 –553 27 Burggraf GW, Parker JO. Prognosis in coronary artery disease: angiographic, hemodynamic, and clinical factors. Circulation 1975; 51:146 –156 28 Fagard RH, Pardaens K, Staessen JA, et al. Prognostic value of invasive hemodynamic measurements at rest and during exercise in hypertensive men. Hypertension 1996; 28:31–36 29 Riedel M, Stanek V, Widimsky J, et al. Long-term follow-up of patients with pulmonary thromboembolism: late prognosis and evolution of hemodynamic and respiratory data. Chest 1982; 81:151–158 30 Squara P, Dhainaut JF, Artigas A, et al. Hemodynamic profile in severe ARDS: results of the European Collaborative ARDS Study [published erratum appears in Intensive Care Med 1999; 25:247]. Intensive Care Med 1998; 24:1018 –1028 31 Marinelli WA, Weinert CR, Gross CR, et al. Right heart catheterization in acute lung injury: an observational study. Am J Respir Crit Care Med 1999; 160:69 –76 32 Mimoz O, Rauss A, Rekik N, et al. Pulmonary artery catheterization in critically ill patients: a prospective analysis of outcome changes associated with catheter-prompted changes in therapy. Crit Care Med 1994; 22:573–579 33 Trottier SJ, Taylor RW. Physicians’ attitudes toward and knowledge of the pulmonary artery catheter: Society of Critical Care Medicine membership survey. New Horiz 1997; 5:201–206 34 Norman GR. Problem-solving skills, solving problems and problem-based learning. Med Educ 1988; 22:279 –286 35 Rubenfeld GD, Caldwell E, Granton J, et al. Interobserver Special Report

variability in applying a radiographic definition for ARDS. Chest 1999; 116:1347–1353 36 Fletcher BD, Glicksman AS, Gieser P. Interobserver variability in the detection of cervical-thoracic Hodgkin’s disease by computed tomography. J Clin Oncol 1999; 17:2153–2159 37 Webb WR, Sarin M, Zerhouni EA, et al. Interobserver variability in CT and MR staging of lung cancer. J Comput Assist Tomogr 1993; 17:841– 846 38 DeRouen TA, Murray JA, Owen W. Variability in the analysis of coronary arteriograms. Circulation 1977; 55:324 –328

www.chestjournal.org

39 Bounameaux H, Prins TR, Schmitt HE, et al. Venography of the lower limbs: pitfalls of the diagnostic standard; the ETTT Trial Investigators. Invest Radiol 1992; 27:1009 –1011 40 Lafitte S, Dos Santos P, Kerouani A, et al. Improved reliability for echocardiographic measurement of left ventricular volume using harmonic power imaging mode combined with contrast agent. Am J Cardiol 2000; 85:1234 –1238 41 National Heart, Lung, and Blood Institute, and Food and Drug Administration Workshop report: consensus statement. JAMA 2000; 283:2568 –2572

CHEST / 121 / 6 / JUNE, 2002

2015