DIAGNOSTIC evaluation for acute pulmonary

Transcription

1 994 2-D ECHO Jackson et al. ECHOCARDIOGRAPHY FOR PULMONARY EMBOLISM Prospective Evaluation of Two-dimensional Transthoracic Echocardiography in Emergency Department Patients with Suspected Pulmonary Embolism RAYMOND E. JACKSON, MD, MS, RAYMOND R. RUDONI, MD, ANDREW M. HAUSER, MD, REBECCA G. PASCUAL, RN, MARY E. HUSSEY, RD, MS Abstract. Objective: To prospectively examine the diagnostic accuracy of two-dimensional transthoracic echocardiography (2-D echo) in emergency department (ED) patients being evaluated for acute pulmonary embolism (PE). Methods: This was a 14-month prospective observational trial of a convenience sample of ED patients undergoing evaluation for suspected PE at a suburban teaching hospital. The 2-D echo was defined as positive if any two of the following were noted: right ventricular dilation, abnormal septal motion, loss of right ventricular contractility, elevated pulmonary artery or right ventricular pressures, moderate to severe tricuspid regurgitation, or visualization of a clot seen in the right ventricle or pulmonary artery. The patient was considered to have a PE if one of the following was positive: a pulmonary angiogram, contrast helical computed tomography, a magnetic resonance angiogram, a high-probability ventilation/perfusion (V/Q) scan without contradictory evidence, or an intermediateprobability V/Q scan with ultrasonic evidence of deep venous thrombosis. Results: Of 225 cases identified, 39 met the defined criteria for PE (17%). A 2-D echo was performed on 124 patients (55%), of whom 27 (22%) had PE. In 20 patients the 2-D echo had at least two indicators of right ventricular strain; however, only 11 of these patients had confirmed pulmonary embolus. The 2-D echo had a sensitivity of 0.41 (95% CI = 0.32 to 0.49) and a specificity of 0.91 (95% CI = 0.86 to 0.96). The likelihood ratio positive was a moderately strong 4.4, with a weak likelihood ratio negative of 0.6. Conclusions: Bedside 2-D echo is not a sensitive test for the diagnosis of PE in ED patients. Positive findings moderately increase the suspicion for PE but are not diagnostic. Key words: two-dimensional echocardiography; transthoracic echocardiography; diagnosis; pulmonary embolism. ACADEMIC EMERGENCY MEDICINE 2000; 7: DIAGNOSTIC evaluation for acute pulmonary embolism (PE) continues to challenge experienced emergency medicine (EM) clinicians. A number of recent publications have observed that two-dimensional transthoracic echocardiography (2-D echo) may play a diagnostic role and provide important hemodynamic information reflecting the effect of a pulmonary embolus on the function of the right heart We have previously reported the diagnostic utility of 2-D echo for PE when compared with the criterion standard, pulmonary angiography. 11 However, concerns regarding these findings include the retrospective nature of the study based on a medical record review, and that the study population was biased by including patients who developed an embolism after admission, and then only those whose diagnosis was uncertain and for whom conditions permitted a pulmonary angiogram. To the best of our knowledge, there have been no prior prospective studies with 2-D echo of emergency department (ED) patients considered by the emergency physician as potentially having a PE. Because of these issues, we prospectively examined the diagnostic accuracy of 2-D echo in ED patients being evaluated for suspected acute PE. From the Department of Emergency Medicine (REJ, RRR, RGP) and the Division of Cardiology, (AMH, MEH), William Beaumont Hospital, Royal Oak, MI. Received February 17, 2000; revision received April 14, 2000; accepted April 20, Presented at the SAEM annual meeting, Boston, MA, May Address for correspondence and reprints: Raymond E. Jackson, MD, MS, Department of Emergency Medicine, William Beaumont Hospital, 3601 West 13 Mile Road, Royal Oak, MI Fax: ; rjerdoc@aol.com METHODS Study Design. We performed a 14-month prospective observational trial of a convenience sample of ED patients undergoing evaluation for suspected PE at a suburban teaching hospital. This study was approved by the William Beaumont Hospital Human Investigations Committee.

2 ACADEMIC EMERGENCY MEDICINE September 2000, Volume 7, Number Study Setting and Population. The study was performed in the ED of William Beaumont Hospital, Royal Oak, Michigan, a 929-bed suburban teaching facility with more than 100,000 yearly ED visits. Emergency department faculty either directly provide or supervise all patient care. The EM faculty and residents write admission orders for patients admitted to the internal medicine service. Study Protocol. A patient was eligible for study enrollment if the attending emergency physician determined that a ventilation/perfusion (V/Q) scan, or another diagnostic test, was needed to evaluate for the presence of a pulmonary embolus. Emergency physicians were encouraged to order a bedside 2-D echo for these patients. The ED bedside 2-D echo was performed by either an echocardiographic technician or a specially-trained cardiology nurse clinician in the ED, typically while waiting for the V/Q scan to be done. For a small number of cases, the EM faculty ordered both the V/Q scan and 2-D echo for the first day of admission. All studies were interpreted by board-certified cardiologists blinded to the results of standard PE tests. The echocardiographic findings were considered positive if any two of the following were noted: right ventricular dilation, abnormal septal motion, impaired right ventricular contractility, elevated pulmonary artery or right ventricular pressures, moderate to severe tricuspid regurgitation, or visualization of a clot within the right heart or pulmonary artery. The diagnosis of PE was deemed positive if one of the following was positive: a pulmonary angiogram, contrast helical computed tomography (CT), a magnetic resonance angiogram (MRA), a high-probability V/Q scan without contradictory evidence, or an intermediate-probability V/Q scan associated with ultrasonic evidence of deep venous thrombosis. Patients were assigned by the emergency physicians a pretest probability of PE (low, intermediate, high), and the intention of the emergency physicians to begin heparin before the V/Q scan was also noted. To determine the completeness of our surveillance and capture, nuclear medicine V/ Q scan logs were reviewed daily, capturing patients not identified by emergency physicians, and, in addition, we identified and reviewed all cases TABLE 1. Pulmonary Embolism (PE) Results for the 124 Patients Who Had Two-dimensional Echocardiography (Echo) PE No PE Echo positive Echo negative TOTAL with the final discharge diagnosis of PE over the study period and one month after termination of recruitment. These procedures were used to identify and characterize cases not captured prospectively. Medical records were reviewed by two authors for vital signs, presence of symptoms, various risk factors, results of studies, final diagnosis, outcome, and length of stay. Data Analysis. We report the sensitivity, specificity, positive and negative predictive values (PPV, NPV), and likelihood ratios positive and negative (LR, LR ), with 95% confidence intervals (95% CIs). We made comparisons between two groups using either Student s t-test, chi-square, or Fisher s exact test as appropriate. RESULTS During the 14-month enrollment period, 225 cases were identified in the ED. The mean SD age was years (range 21, 94), and 75 (33%) were male. The final ED disposition was hospital admission in 156 (69.3%). Of the 225 patients, 39 met the defined criteria for PE (17%). A 2-D echo was performed on 124 patients (55%), of whom 27 (22%) had PE (Table 1). The echo group had a slightly higher rate of PE than the no-echo group (p = 0.054). The group undergoing echocardiography was older, having a mean age of years (t = 3.55, p < 0.001). The group with 2-D echo had a higher admission rate (80.6%) than the group without 2-D echo (55%) (p < 0.001). Of 12 patients with systemic hypotension defined as a systolic blood pressure less than 100 mm Hg, ten had 2-D echo performed. The diagnosis was confirmed by pulmonary angiogram in four patients, contrast helical CT in three, magnetic resonance angiogram in two, a high-probability V/Q scan in 14, and an intermediate-probability V/Q scan with a venous Doppler study positive for deep venous thrombosis in four. Heparin therapy was started in the ED for 25 patients, and a fibrinolytic agent was given to three. In 20 patients the 2-D echo had at least two indicators of right ventricular strain; however, only 11 of these patients had confirmed PE. In our study cohort, the 2-D echo had a sensitivity of 0.41 (95% CI = 0.32 to 0.49) and a specificity of 0.91 (95% CI = 0.86 to 0.96). The PPV was 0.55 (95% CI = 0.46 to 0.64) and the NPV was 0.85 (95% CI = 0.78 to 0.9). The LR was a moderately strong 4.4, with a weak LR of 0.6 (Table 1). Detailed echocardiographic findings are listed in Table 2. We note that any finding moderately increases the posttest probability of PE, whereas the absence of the these findings does little to exclude the possibility of PE.

3 996 2-D ECHO Jackson et al. ECHOCARDIOGRAPHY FOR PULMONARY EMBOLISM TABLE 2. Echocardiographic Findings* Finding PE (n = 27) No PE (n = 97) Se Sp LR LR Right ventricular (RV) dilation Abnormal septal motion Abnormal RV contractility Clot in RV/pulmonary artery 0 0 Moderate to severe tricuspid insufficiency Elevated pulmonary artery pressure Elevated RV systolic pressure (>44 mm Hg) Any finding *PE = pulmonary embolism; Se = sensitivity; Sp = specificity; LR = likelihood ratio positive; LR = likelihood ratio negative. Of the 12 patients who died, ten were in the echocardiogram cohort, with two meeting their demise secondary to PE. Of these two, one had a positive echocardiogram. Of the remaining eight, two had positive echocardiograms. No patient died in the ED. With our surveillance procedures, we retrospectively identified an additional 66 patients who met our criteria for PE during the study period and were not in the prospective study cohort. The emergency physician suspected the diagnosis of PE in 44 of these 66 patients. A 2-D echo was performed on 35 (53%) of these 66 patients, of whom eight (22%) were positive by our criteria. We were able to identify only one case in which there was an ED visit within two weeks for a patient who had symptoms compatible with PE but was not initially admitted during the study period. DISCUSSION To the best of our knowledge, this is the first prospective study of ED patients undergoing evaluation for possible PE. We demonstrated that 2-D echo is an insensitive exam, having only a moderately strong LR and a weak LR for the diagnosis of PE. Therefore, as described in this study, it is not useful as a screening tool. In agreement with prior work, findings of right ventricular compromise were more common in patients with an acute PE. 1 3 Our observed LR for 2-D echo is moderately strong and is supported by the literature. Grifoni et al. noted that the 2-D echo was poorly sensitive (51%) but highly specific (87%) in a series of inpatients undergoing evaluation of possible embolus. 4 In their series any one of these findings was considered a positive 2-D echo: visualization of clot, right ventricular dilation, systolic flattening of the septum, and pulmonary hypertension. The prevalence of PE in their series was 54%, much higher than that which we observed in our ED cohort. Nevertheless, their estimated LR was 3.9, very similar to our observation. Steiner et al. also found that the overall sensitivity of 2-D echo was low (59%) with a modest specificity (77%) in a series of consecutive hospitalized patients. Their estimated LR was 2.6. They noted that for central emboli (i.e., emboli in the main pulmonary arteries or major segmental branches), the sensitivity and specificity increased to 82% and 92%, respectively, with an estimated LR of Nazeyrollas et al., 2 in an intensive care setting, found that the LR for specific echocardiographic findings ranged from 3.9 to 5.1, consistent with our findings. However, they noted that normal tricuspid peak flow velocity of less than 2.5 m/sec had a strong LR of 0.085, which essentially excludes the diagnosis. In our previous report of hospitalized patients undergoing pulmonary angiograms, 11 we reported an LR of 22, with a 33% prevalence of PE. Kasper et al., 7 in a cohort with a very high prevalence of 61%, also reported a strong LR of 25. Therefore, our findings in ED patients compare well with the literature on 2-D echo findings for hospitalized patients with the diagnosis of PE. Clinicians using this modality in combination with other tools may improve diagnostic accuracy. 1,2 From our data, we know the prevalence of PE in the echo cohort was 22% (i.e., the pretest odds were about 1:4). Given a positive 2-D echo, with an LR of 4.4, the posttest odds are now approximately 1:1, meaning that the patient s probability of PE is now 50%. The future role of 2-D echo in the assessment of PE may be in risk stratification. Ribeiro et al. 1 reported echocardiographic estimates of the extent of right ventricular dysfunction was a prognostic indicator of mortality. They divided the results of 2-D echo in 126 patients admitted with the diagnosis of PE into two groups: group A with no or little right ventricular dysfunction, and group B with moderate to severe right ventricular dysfunction. The interrater reliability for that assignment was excellent (kappa 0.94). Patients in group B experienced a significantly higher in-hospital mortality rate, with a relative risk of 6.0 (95% CI = 1.1 to 111.5) as all in-hospital deaths were in patients with moderate to severe right ventricular dysfunc-

4 ACADEMIC EMERGENCY MEDICINE September 2000, Volume 7, Number tion. These investigators also noted a significant association with mortality and findings of right ventricular dilation, and moderate to severe tricuspid regurgitation. Although a significant minority of their cohort received a thrombolytic agent, the effect of the treatment was not factored into the reporting of mortality. Kasper et al. 7 also noted higher total and PE-related in-hospital mortality for hospitalized patients with dilation of the right ventricle, odds ratio (OR) = 3.7 for total mortality and OR = 16.5 for mortality secondary to PE. Konstantinides et al. 9 noted that in patients with PE, right ventricular dilation had an OR of 2.6 for 30- day mortality. Kasper et al. 8 noted that 65% of patients from a multicenter PE registry had evidence of right ventricular pressure overload or pulmonary hypertension when presenting with arterial hypotension, with cardiogenic shock, or requiring cardiopulmonary resuscitation. Nass et al. 10 recently reported improvement of right ventricular function after treatment with thrombolytics. Therefore, early identification of patients with right ventricular dysfunction may be advantageous. Goldhaber et al. 12 recently reported from the ICOPER database a twofold increase in the 90-day mortality for patients noted to have right ventricular hypokinesis on echocardiography. LIMITATIONS AND FUTURE QUESTIONS Although this was a prospective study, whether a particular patient received a 2-D echo was the decision of the emergency physician. Of the 225 patients identified, the 124 who had a 2-D echo were older, were more likely to be admitted, and were in general more ill as reflected by the higher mortality. This reflects the bias of clinicians hesitant to order an additional test for patients deemed not acutely ill. This bias most likely resulted in underestimating the sensitivity while overestimating the specificity of the 2-D echo. However, this should have only a minimal effect on the estimated likelihood ratios as these biases offset each other. 13 The constellation of echocardiographic findings that are consistent with acute right ventricular failure secondary to an acute PE require further validation and testing for reliability. We prospectively chose to use two findings to optimize specificity at the expense of sensitivity. When we reanalyzed the data using any finding as a positive test, the sensitivity increased slightly to 0.56 (95% CI = 0.47 to 0.64), with a decrease in specificity to 0.79 (95% CI = 0.72 to 0.87) (Table 2). Though the cases were identified prospectively, the information on history and physical examination findings as well as the presence of risk factors were extracted retrospectively from a medical record review, though this information had no bearing on any factors pertaining to the objective of the study. There were seven cases in which the treating admitting physician accepted an intermediateprobability V/Q scan as diagnostic and performed no other confirmatory tests. As this practice did not meet our criteria for embolism, we may have underestimated the true rate of PE in our population. Further, an unknown number of patients may have been discharged from the ED with a PE without investigation. Future studies regarding ED 2-D echo may explore the added benefit of contrast echo for identifying patients with a patent foramen ovale. That condition, which is present in 25% of the population, may identify patients with a true right-to-left shunt secondary to the PE. As discussed above, the true value of echo may be in identifying patients who may benefit most from fibrinolytic therapy. Identification of these cases in the ED could optimize this benefit. CONCLUSIONS We report that bedside 2-D echo is not a sensitive test for diagnosis of PE in ED patients. Positive findings moderately increase the suspicion for PE in an individual patient, but are certainly not diagnostic. How emergency physicians best include this modality into their diagnostic, therapeutic prognostic assessment of patients suspected PE requires further study. References 1. Ribeiro A, Lindmarker P, Juhlin-Dannfelt A, Johnsson H, Jorfeldt L. Echocardiography Doppler in pulmonary embolism: right ventricular dysfunction as a predictor of mortality rate. Am Heart J 1997; 134: Nazeyrollas P, Metz D, Jolly D, et al. Use of transthoracic Doppler echocardiography combined with clinical and electrocardiographic data to predict acute pulmonary embolism. Eur Heart J. 1996; 17: Nazeyrollas P, Metz D, Chapoutot L, et al. Diagnostic accuracy of echocardiography-doppler in acute pulmonary embolism. Int J Cardiol. 1995; 47: Grifoni S, Olivotto I, Cecchini P, et al. Utility of an integrated clinical, echocardiographic, and venous ultraonographic approach for triage of patients with suspected pulmonary embolism. Am J Cardiol. 1998; 82: Steiner P, Lund GK, Debatin JF, et al. Acute pulmonary embolism: value of transthoracic and tranesophageal echocardiography in comparison with helical CT. Am J Roentgenol. 1996; 167: Kasper W, Geibel A, Tiede N, et al. Distinguishing between acute and subacute massive pulmonary embolism by conventional and Doppler echocardiography. Br Heart J. 1993; 70: Kasper W, Konstantinides S, Geibel A, Tiede N, Krause T, Just H. Prognostic significance of right ventricular afterload stress detected by echocardiography in patients with clinically suspected pulmonary embolism. Heart. 1997; 77: Kasper W, Konstantinides S, Geibel A, et al. Management strategies and determinants of outcome in acute major pulmonary embolism: results of a multicenter registry. J Am Coll

5 998 2-D ECHO Jackson et al. ECHOCARDIOGRAPHY FOR PULMONARY EMBOLISM Cardiol. 1997; 30: Konstantinides S, Geibel A, Olschewski M, et al. Association between thrombolytic treatment and the prognosis of hemodynamically stable patients with major pulmonary embolism. Circulation. 1997; 97: Nass N, McConnell MV, Goldhaber SZ, Chyu S, Solomon SD. Recovery of regional right ventricular function after thrombolysis for pulmonary embolism. Am J Cardiol. 1999; 83: 804 6, A Rudoni RR, Jackson RE, Godfrey GW, Bonfiglio AX, Hussey ME, Hauser AM. Use of two-dimensional echocardiography for the diagnosis of pulmonary embolus. J Emerg Med. 1998; 16: Goldhaber SZ, Visani L, DeRosa M for ICOPER. Acute pulmonary embolism: clinical outcomes in the International Cooperative Pulmonary Embolism Registry (ICOPER). Lancet. 1999; 353: Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical Epidemiology: A Basic Science for Clinical Medicine. Boston: Little, Brown, Instructions for Contributors to Clinical Pearls Clinical Pearls is a section of Academic Emergency Medicine that uses photographic images to provide visual clues for a case study presented as an unknown. Visual clinical findings make up a large part of the practice of emergency medicine (EM). Capturing these findings allows clinicians to share their experience and knowledge with others, making clinical photographs an excellent teaching tool. This section intends to stimulate academic emergency physicans to use clinical photography for augmenting their teaching of EM. Clinical Pearls manuscripts should be presented as case study unknowns and must be accompanied by a clinical photograph. Radiographs and other supporting data (ECGs, pathology specimens, Gram stains, etc.) are acceptable if they accompany a clinical photograph. A series of clinical photographs to demonstrate a progressive disease process is acceptable. Cases with a radiograph or ECG alone should be discussed with the section editor before submission. Manuscript preparation should follow the general Instructions for Authors found in AEM. Format of the section follows this general scheme: Title (usually the chief complaint of the patient), Chief Complaint, History of Present Illness, Physical Examination and Laboratory, Diagnosis, Discussion, Clinical Pearls (3 5 take-home points of the case), and References. The most original image available (slide, negative, or photograph) and two 5 7-inch color prints should accompany the manuscript. The original image will be returned. Arrows, symbols, or labels identifying structures should be marked on the second print if necessary. Each print and slide should be labeled with the last name(s) of the contributors and an arrow indicating the top of the image. Contributors must provide the names, highest academic degrees, addresses, and phone and fax numbers of the photographer and all contributors. Acknowledgment of manuscript and photograph acceptance will be made in writing to the contributor. The section editor will have the photograph critiqued by a professional medical photographer to provide suggestions for improving photographic technique. The critique will become part of the published article. By submitting to the Clinical Pearls, the contributor allows the section editor to distribute the case and image to all EM residency programs in the United States as part of a prejournal mail-out. This activity allows programs to preview the case in didactic situations to enhance the learning from the case. It is the responsibility of the contributor to obtain patient consent for use of the photograph in a publication if the patient is in any way identifiable. Send manuscripts and images to AEM, 901 North Washington Avenue, Lansing, MI For additional information or questions, contact Larry Stack, phone: ; fax: ; larry.stack@mcmail.vanderbilt.edu