Adherence to PIOPED II Investigators Recommendations for Computed Tomography Pulmonary Angiography

IMAGING FOR THE CLINICIAN SPECIAL SECTION CLINICAL RESEARCH STUDY Robert G. Stern, MD, Section Editor Adherence to PIOPED II Investigators Recommendations for Computed Tomography Pulmonary Angiography Daniel M. Adams, MD, a Scott M. Stevens, MD, a,b Scott C. Woller, MD, a,b R. Scott Evans, PhD, c James F. Lloyd, BS, c Gregory L. Snow, PhD, d Todd L. Allen, MD, e Joseph R. Bledsoe, MD, e Lynette M. Brown, MD, PhD, f,b Denitza P. Blagev, MD, f Todd D. Lovelace, MD, g Talmage L. Shill, MD, g Karen E. Conner, MD, MBA, g Valerie T. Aston, RRT, f C. Gregory Elliott, MD a,b a Department of Medicine, Intermountain Medical Center, Murray, Utah; b Department of Medicine, University of Utah School of Medicine, Salt Lake City; c Medical Informatics, LDS Hospital, Salt Lake City, Utah; d Medical Statistics, LDS Hospital, Salt Lake City, Utah; e Department of Emergency Medicine, Intermountain Medical Center, Murray, Utah; f Division of Pulmonary and Critical Care Medicine, Intermountain Medical Center, Murray, Utah; g Department of Radiology, Intermountain Medical Center, Murray, Utah. ABSTRACT BACKGROUND: Computed tomography (CT) pulmonary angiography use has increased dramatically, raising concerns for patient safety. Adherence to recommendations and guidelines may protect patients. We measured adherence to the recommendations of Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED II) investigators for evaluation of suspected pulmonary embolism and the rate of potential false-positive pulmonary embolism diagnoses when recommendations of PIOPED II investigators were not followed. METHODS: We used a structured record review to identify 3500 consecutive CT pulmonary angiograms performed to investigate suspected pulmonary embolism in 2 urban emergency departments, calculating the revised Geneva score (RGS) to classify patients as pulmonary embolism unlikely (RGS 10) or pulmonary embolism likely (RGS 10). CT pulmonary angiograms were concordant with PIOPED II investigator recommendations if pulmonary embolism was likely or pulmonary embolism was unlikely and a highly sensitive D-dimer test result was positive. We independently reviewed 482 CT pulmonary angiograms to measure the rate of potential false-positive pulmonary embolism diagnoses. RESULTS: A total of 1592 of 3500 CT pulmonary angiograms (45.5%) followed the recommendations of PIOPED II investigators. The remaining 1908 CT pulmonary angiograms were performed on patients with an RGS 10 without a D-dimer test (n 1588) or after a negative D-dimer test result (n 320). The overall rate of pulmonary embolism was 9.7%. Potential false-positive diagnoses of pulmonary embolism occurred in 2 of 3 patients with an a negative D-dimer test result. CONCLUSIONS: Nonadherence to recommendations for CT pulmonary angiography is common and exposes patients to increased risks, including potential false-positive diagnoses of pulmonary embolism. 2013 Elsevier Inc. All rights reserved. The American Journal of Medicine (2013) 126, 36-42 KEYWORDS: Computed tomography; Guidelines; Pulmonary embolism SEE RELATED EDITORIAL p. 3 The widespread use of computed tomography (CT) pulmonary angiography for evaluation of suspected pulmonary embolism raises concern over potential harm to patients. 1 CT pulmonary angiography exposes patients to ionizing radiation and the risk of renal injury, soft tissue injury, or anaphylaxis from intravenous contrast. 2,3 Furthermore, estimates sug- Presented at: the American Thoracic Society meeting, May 16, 2011, Denver, Colorado. Funding: Supported in part by the North American Thrombosis Forum Traveling Fellowship. Conflict of Interest: None. Authorship: All authors had access to the data and played a role in writing this manuscript. Requests for reprints should be addressed to C. Gregory Elliott, MD, University of Utah School of Medicine, Department of Medicine, Intermountain Medical Center, 5121 S. Cottonwood, #307, Murray, UT 84107. E-mail address: greg.elliott@imail.org 0002-9343/$ -see front matter 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjmed.2012.05.028

Adams et al Adherence to Recommendations for CT Pulmonary Angiography 37 gest that more false-positive diagnoses of pulmonary embolism occur when CT pulmonary angiography is performed on populations with a low prevalence ( 5%) of pulmonary embolism. 4 False-positive interpretations of CT pulmonary angiograms expose patients to risks of antithrombotic therapies, including fatal and nonfatal major bleeding, and repeat CT CLINICAL SIGNIFICANCE pulmonary angiograms that further increase radiation exposure. 5 The nonspecific clinical presentation of pulmonary embolism and the potential risks associated with CT pulmonary angiography motivate physicians to seek optimal approaches for the investigation of suspected pulmonary embolism. Such approaches often take the form of expert recommendations and guidelines. In 2006, Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED II) investigators published recommendations for the diagnostic approach to patients with suspected pulmonary embolism. 6 In 2008, the European Society of Cardiology published guidelines on the diagnosis of pulmonary embolism, 7 and in 2011, the American College of Emergency Physicians published a clinical policy for evaluation of adults with suspected pulmonary embolism. 8 Current recommendations and guidelines to evaluate suspected pulmonary embolism agree that assessment of pretest probability is an essential step before ordering a CT pulmonary angiogram. 6-8 When pulmonary embolism is unlikely, a negative highly sensitive D-dimer test result allows anticoagulants to be withheld safely without CT pulmonary angiography. 6,9-12 Furthermore, when pulmonary embolism is unlikely and a highly sensitive D-dimer test result is negative, CT pulmonary angiography may lead to false diagnoses of pulmonary embolism more often than true diagnoses of pulmonary embolism. The primary objective of this study was to measure the proportion of CT pulmonary angiograms that were concordant with recommendations for the evaluation of suspected pulmonary embolism. 6 MATERIALS AND METHODS Study Site and Population Intermountain Medical Center is a university-affiliated tertiary medical center in Murray, Utah, and LDS Hospital is a community hospital in Salt Lake City, Utah. In 2009 and 2010, emergency department visits exceeded 83,000 and 25,000, respectively. The Intermountain Healthcare Institutional Review Board approved this Health Insurance Portability and Accountability Act-compliant study and waived written informed consent. We queried the Intermountain Healthcare Current evidence-based recommendations and guidelines agree that assessment of pretest probability is essential and that computed tomography (CT) pulmonary angiography is unnecessary when pretest probability is low and a sensitive D-dimer test result is negative. Physicians often order CT pulmonary angiography, although pulmonary embolism is unlikely and a D-dimer test result is negative. Nonadherence to recommendations for CT pulmonary angiography exposes patients to increased risk, including false-positive diagnoses of pulmonary embolism. Enterprise Data Warehouse and identified consecutive patients who had chest CT ordered in the emergency departments at LDS Hospital or Intermountain Medical Center. The emergency department physician s report was reviewed manually by one investigator (DA), who used a data abstraction form to verify that each CT pulmonary angiogram was performed to investigate suspected pulmonary embolism. 13-16 We analyzed 3500 consecutive CT pulmonary angiograms ordered to evaluate clinically suspected pulmonary embolism. No CT pulmonary angiograms were excluded. Data Collection and Processing Data elements were collected both by electronic query and by manual review of the electronic medical record. Electronically captured historical elements were verified for each patient by manual review of the emergency department physician s report by using a data abstraction form. To assess the accuracy of elements captured electronically, one investigator (DA) performed an independent manual review of data elements from 100 consecutive records and confirmed 100% agreement. Elements not captured electronically but apparent in the manual review (eg, surgery at an outside institution) were recorded in the study database. Appendix 1 lists the method of collection for each data element. Pretest probability for pulmonary embolism was assessed using the revised Geneva score (RGS). 17 This tool has been validated prospectively 11 and is composed of objective data elements. Prior venous thromboembolism was collected using a computer program developed at Intermountain Healthcare with high sensitivity and specificity for identification of prior venous thromboembolism. 18 Recent surgery was extracted from the Operating Room Management Information System. Recent fracture was identified by International Classification of Diseases, Ninth Revision codes 800-829. Cancer was identified by International Classification of Diseases, Ninth Revision codes 140-239, and active cancer was verified by manual record review. Unilateral leg pain and hemoptysis were recorded manually from the emergency physician s report. If these were not described in the emergency physician s report, we concluded that they were absent. Heart rate was defined as the highest recorded value from the vital signs before CT pulmonary angiography. Pain on palpation of the deep veins of the leg and unilateral edema (also accepted was the physician s interpretation that findings were consistent with deep

38 The American Journal of Medicine, Vol 126, No 1, January 2013 vein thrombosis) were recorded from the physical examination component of the physician s note, and if they were not recorded, they were considered absent. Both emergency departments used high-sensitivity D-dimer tests (Stago latex agglutination [Diagnostica Stago, Parsippany, NJ] and mini-vidas [biomerieux, Durham, NC]). Values from 0 to 499 ng/ml were considered negative. All CT pulmonary angiograms were performed on 64 detector row CT scanners. Intermountain Medical Center used Toshiba Aquilion CT scanners (Toshiba America Medical Systems, Inc, Tustin, Calif) for CT pulmonary angiography. LDS Hospital used both Toshiba Aquilion and General Electric VCT scanners (GE Healthcare, Waukesha, Wis). AutomA software (GE Healthcare) was used to deliver tailored radiation exposure. The standard CT pulmonary angiogram protocol used 80 ml of iso-osmolar contrast with an iodine concentration of 370 mg/ml injected at a rate of 4 to 5 ml/s and timed for maximum pulmonary artery opacification through bolus tracking (Toshiba America Medical Systems, Inc) or 20 ml timing bolus (GE Healthcare). Every CT pulmonary angiogram was interpreted by an in-house, board-certified radiologist. Primary Outcome Measure The primary measurement for this study was concordance of CT pulmonary angiogram examinations with recommendations published by the PIOPED II investigators. 6 We used the RGS 17 to quantify the probability of pulmonary embolism as pulmonary embolism likely (RGS 10) or pulmonary embolism unlikely (RGS 10). CT pulmonary angiograms were classified as concordant with PIOPED II investigator recommendations if the patient s pretest probability classification was pulmonary embolism likely or the pretest probability was pulmonary embolism unlikely and a D-dimer test result was positive. CT pulmonary angiograms were not concordant with PI- OPED II investigator recommendations if the patient s pretest probability was pulmonary embolism unlikely and no D-dimer test was performed or if the D-dimer test result was negative. Some clinical situations, such as pregnancy and recent trauma or surgery, decrease the likelihood of a negative D-dimer test result and may influence clinicians to forego D-dimer testing. To assess the impact of such circumstances, we recorded the number of patients who were classified as pulmonary embolism unlikely and did not have a D-dimer test performed but were pregnant or had recent (within 1 month) surgery or trauma. The rate of pulmonary embolism on CT pulmonary angiogram ( diagnostic yield ) was determined from the original interpretations of all CT pulmonary angiograms (number of positive pulmonary embolism interpretations/number of CT pulmonary angiograms). The rate of potential false-positive pulmonary embolism diagnoses was determined by an independent (blinded to D-dimer testing and initial interpretation of the CT pulmonary angiogram) interpretation of 482 CT pulmonary angiograms (all 241 CT pulmonary angiograms reported to show acute pulmonary embolism on the initial interpretation and an equal number, 241, of randomly selected CT pulmonary angiograms reported negative for pulmonary embolism on initial interpretation). The independent interpretation was performed by 1 of 3 fellowship-trained, board-certified chest radiologists (TL, TS, and KC). Kappa coefficients for agreement of the 3 chest radiologists were calculated by overlapping 20% of the sample and were found to be 1.0, 0.94, and 0.94. Data Analysis We descriptively analyzed the patients demographics and outcome measurements. Ninety-five percent confidence intervals were calculated for the proportion of CT pulmonary angiograms that were concordant or nonconcordant with PIOPED II investigator recommendations. The rate of pulmonary embolism with 95% confidence intervals was calculated for patients grouped by concordance with PIOPED II investigators recommendations. Median values and interquartile range were calculated for the RGS for the entire patient group and patient subgroups defined by recommendation concordance. The rate of potential false-positive pulmonary embolism diagnoses was determined according to subgroup on the basis of concordance with PIOPED II investigator recommendations. RESULTS Sample and Patient Characteristics The reports of 5220 consecutive chest CT scans ordered from the 2 emergency departments were reviewed to identify 3500 consecutive CT pulmonary angiograms ordered between May 22, 2009, and June 30, 2010, for suspected pulmonary embolism. A total of 2755 (78.7%) CT pulmonary angiograms were performed at Intermountain Medical Center, and 745 (21.3%) CT pulmonary angiograms were performed at LDS Hospital. Table 1 shows the clinical characteristics for all patients and subgroups prespecified by concordance with PIOPED II investigator recommendations. The pulmonary embolism likely subgroup differed from the pulmonary embolism unlikely subgroup in variables used to calculate the RGS. Patients in the pulmonary embolism unlikely (RGS 10) group (n 1588) for whom D-dimer testing was not ordered were more likely to have had recent surgery, trauma, or fracture; active cancer; or prior venous thromboembolism than patients in the pulmonary embolism unlikely group (n 1745) who had D-dimer levels measured. Other variables (eg, pregnancy or unilateral leg pain) were observed at similar rates for the patients in the pulmonary embolism unlikely group who did or did not have a D-dimer test ordered. Concordance With PIOPED II Investigators Recommendations A total of 1592 of 3500 (45.5%) CT pulmonary angiograms were concordant with the PIOPED II investigators recom-

Adams et al Adherence to Recommendations for CT Pulmonary Angiography 39 Table 1 Patient Characteristics* All RGS 10 D-dimer Positive D-dimer Negative D-dimer Done D-dimer Not Done N 3500 167 1425 320 1745 1588 Age, y, mean (SD) 52 (19.0) 56 (19.6) 54 (19.3) 44 (15.9) 52 (19.1) 51 (18.8) Female, n (%) 2204 (63.0) 101 (60.5) 928 (65.1) 200 (62.5) 1128 (64.6) 975 (61.4) Prior VTE, n (%) 682 (19.5) 97 (58.1) 153 (10.7) 50 (15.6) 203 (11.6) 382 (24.1) Surgery, n (%) 399 (11.4) 33 (19.8) 117 (8.2) 12 (3.8) 119 (6.8) 237 (14.9) Fracture, n (%) 54 (1.5) 7 (4.2) 13 (0.9) 2 (0.6) 15 (0.9) 32 (2.0) Trauma, n (%) 117 (3.3) 10 (6.0) 34 (2.4) 8 (2.5) 42 (2.4) 64 (4.0) Cancer, n (%) 244 (7.0) 29 (17.4) 59 (4.1) 9 (2.8) 68 (3.9) 147 (9.3) Pregnancy, n (%) 70 (2.0) 3 (1.8) 29 (2.0) 2 (0.6) 31 (1.8) 36 (2.3) Hemoptysis, n (%) 69 (2.0) 6 (3.6) 17 (1.2) 6 (1.9) 23 (1.3) 40 (2.5) Leg pain, n (%) 249 (7.1) 115 (68.9) 48 (3.4) 17 (5.3) 65 (3.7) 69 (4.3) Leg edema, n (%) 231 (6.6) 133 (79.6) 36 (2.5) 7 (2.2) 43 (2.5) 55 (3.5) Heart rate min 1, mean (SD) 94.0 (20.5) 104.0 (17.1) 93.0 (20.5) 91.0 (18.5) 92.7 (20.2) 94.4 (20.8) RGS revised Geneva score; 17 SD standard deviation; VTE venous thromboembolism. *Characteristics of the patient at the time of each CT pulmonary angiogram were included. Within 2 mo. Currently treated or cure 1y. mendations. Of these, 167 (4.8%) were pulmonary embolism likely and 1425 (40.7%) were pulmonary embolism unlikely with a positive highly sensitive D-dimer test result. The remaining 1908 (54.5%) CT pulmonary angiograms were not concordant with the PIOPED II investigators recommendations. Of these, 1588 (45.4%) were pulmonary embolism unlikely without a D-dimer test, and 320 (9.1%) were pulmonary embolism unlikely with a negative D-dimer test. Of the 1588 patients who were classified as pulmonary embolism unlikely without D-dimer tests, 319 (20.1%) had a clinical situation in which the clinician might forego D-dimer testing (pregnancy, recent surgery, or trauma). If these cases were considered concordant with recommendations of the PIOPED II investigators, the rate of CT pulmonary angiogram concordance with recommendations of the PIOPED II investigators increased from 45.5% to 54.6%. Diagnostic Yield Table 2 shows the diagnostic yield and RGS data related to concordance with the recommendations of the PIOPED II investigators. On the basis of the initial interpretations of 3500 CT pulmonary angiograms, the overall pulmonary embolism rate was 9.7%. The pulmonary embolism rate for patients in the pulmonary embolism likely (RGS 10) group was 37.1%, and the pulmonary embolism rate for patients in the pulmonary embolism unlikely (RGS 10) group with a negative D-dimer test result was 0.9%. The pulmonary embolism rate (10.0%) for patients in the pulmonary embolism unlikely (RGS 10) group with a positive D-dimer test result was higher than the pulmonary embolism rate (8.4%) for patients in the pulmonary embolism unlikely (RGS 10) group in whom D-dimer was not tested. CT pulmonary angiograms classified as concordant with recommendations were positive for pulmonary embo- Table 2 Recommendation Concordance, Revised Geneva Score, and Diagnostic Yield Revised Geneva Score Diagnostic Yield n % 95% CI Median (IQR)* n % PE 95% CI Concordant subgroups PE likely 167 4.8 4-6 12 (11-13) 62 37.1 29.8-44.5 PE unlikely and positive D-dimer 1425 40.7 39-42 5 (3-6) 142 10.0 8.4-11.5 Nonconcordant subgroups PE unlikely and no D-dimer 1588 45.4 44-47 5 (4-7) 134 8.4 7.1-9.8 PE unlikely and negative D-dimer 320 9.1 8-10 5 (3-5) 3 0.9 0.2-2.5 CI confidence interval; IQR interquartile range; PE pulmonary embolism; RGS revised Geneva score. 17 *Median RGS and IQR are for all patients in the group, not just those with a positive CT pulmonary angiogram. Diagnostic yield number of CT pulmonary angiogram examinations with PE/number of CT pulmonary angiogram examinations.

40 The American Journal of Medicine, Vol 126, No 1, January 2013 Table 3 Rate of Potential False-Positive Computed Tomography Pulmonary Angiogram Interpretations Concordant RGS 10 Positive D-dimer Not Concordant no D-dimer Negative D-dimer Potential false rate 0.00 0.06 0.03 0.67 (95% CI) (0.00-0.06) (0.03-0.09) (0.01-0.05) (0.23-0.96) CI confidence interval; RGS revised Geneva score. 17 lism 12.8% of the time (204/1592) compared with 7.2% of the nonconcordant CT pulmonary angiograms (137/1908). Potential False Diagnoses of Pulmonary Embolism Independent interpretation of 241 consecutive CT pulmonary angiograms interpreted initially to show acute pulmonary embolism identified 13 potential false-positive interpretations. The potential false-positive rate of initial CT pulmonary angiogram interpretation was high (0.67) for diagnoses of acute pulmonary embolism when RGS was 10 with a negative D-dimer test result (Table 3). DISCUSSION We found that more than half of 3500 CT pulmonary angiograms performed to investigate clinically suspected acute pulmonary embolism were not concordant with recommendations of the PIOPED II investigators. There are several possible explanations for the disparity we observed between practice and published recommendations. First, 95% of patients who underwent CT pulmonary angiography for suspected pulmonary embolism had a pretest probability of pulmonary embolism unlikely (RGS 10). Most emergency physicians assess pretest probability by gestalt rather than using a standardized tool. Although this has been reported to be safe, 6 gestalt assessment may lead to additional testing because it seems less reliable for less-experienced clinicians and has poor interobserver agreement compared with standardized prediction tools. 19,20 Furthermore, gestalt is less specific than validated pretest probability scores. 21 Clinicians, using a gestalt assessment of pretest probability, may be more likely to perform CT pulmonary angiography on the basis of their gestalt of intermediate pretest probability of pulmonary embolism, either bypassing or overriding D-dimer tests. Second, there are several risk assessment scores. It is possible some physicians used a different risk assessment score, which may have led to a different classification of pretest probability than the RGS. However, the most widely used of these, the Wells score, 22 classifies more patients in the low category than the RGS. 23 Therefore, if clinicians used the Wells score, we would expect even higher use of D-dimer assays. Third, the evidence is less compelling that a highly sensitive D-dimer test excludes pulmonary embolism safely when the RGS is 4 to 10. 8 This led the writing committee of the American College of Emergency Physicians to advise caution in excluding pulmonary embolism when the RGS is 4 to 10 and the D-dimer test result is negative. 8 It is possible that some physicians shared the perspective that strong evidence to support the use of D-dimer testing to exclude pulmonary embolism is lacking for patients with an intermediate pulmonary embolism risk. 8 This perspective could have led physicians to avoid D-dimer tests when their suspicion for pulmonary embolism was intermediate, leading to nonconcordance with recommendations of the PIOPED II investigators. Fourth, it is possible that physicians were influenced by clinical variables not represented in the RGS. Courtney et al 24 recently showed that non cancer-related thrombophilia, pleuritic chest pain, and a family history of venous thromboembolism increased the probability of pulmonary embolism. Consideration of these variables may have led physicians to bypass D-dimer testing. Although this possibility does not alter our measurement of concordance with recommendations of the PIOPED II investigators, it may lead some to conclude that we underestimated the rate of appropriate use of CT pulmonary angiography. Finally, some physicians may simply choose to perform CT pulmonary angiography in all patients with suspected pulmonary embolism, because they may be unfamiliar with recommendations and guidelines, find that assessing pretest probability is complex, or believe that CT pulmonary angiography protects them from malpractice claims. Variation exists between recommendations. In June 2011 (after the patient encounters in our study), the American College of Emergency Physicians Clinical Policies Subcommittee issued recommendations on the diagnosis of suspected pulmonary embolism, 8 which differ from the recommendations of the PIOPED II investigators. Differences include acceptance of gestalt to determine pretest probability of pulmonary embolism and a low (level C) recommendation for the use of a sensitive D-dimer test to exclude pulmonary embolism in patients with an intermediate (RGS 4-10) pretest probability of pulmonary embolism. 8 Irrespective of which recommendation is invoked, our data demonstrate nonconcordant use of CT pulmonary angiography for patients with a low pretest probability of pulmonary embolism (RGS 0-3) and negative sensitive D-dimer test results. Substantial research has focused on reducing the overuse of CT pulmonary angiography for suspected pulmonary

Adams et al Adherence to Recommendations for CT Pulmonary Angiography 41 embolism because CT pulmonary angiograms are costly and inconvenient, and engender risks, such as contrast dye injuries, 25,26 radiation exposure, 27-29 and overdiagnosis of pulmonary embolism. 30,31 A combination of assessment of pretest probability and use of a sensitive D-dimer assay has been shown to identify a group of patients with suspected pulmonary embolism for whom the diagnosis can be excluded and anticoagulants can be withheld safely without CT pulmonary angiography. 10,11,23 However, there are barriers to performing pretest probability assessment and sensitive D-dimer testing in emergency departments, including complexity of formalized pretest probability scoring, imperatives for shorter emergency department evaluations, and concern regarding the risks of excluding suspected pulmonary embolism without imaging. Our work confirms the safety of excluding pulmonary embolism on the basis of an objective pretest probability score of pulmonary embolism unlikely (RGS 10) and a negative D-dimer test result. We demonstrate that the diagnostic yield decreases when CT pulmonary angiography use is not concordant with recommendations of the PIOPED II investigators. Study Limitations Our study has several limitations. First, data were obtained retrospectively from medical records. Historical data were taken from the emergency physician s report and electronic data. It is possible that there were elements that the physician did not include in the report. We tried to compensate for this limitation by using objective and well-defined data elements. Second, because we assessed only patients who underwent CT pulmonary angiography, our methods did not identify patients in whom pulmonary embolism was suspected, but deep vein thrombosis was diagnosed by venous ultrasonography, without CT pulmonary angiography. Our approach may have underestimated concordance with recommendations by excluding these patients. However, our cohort included a similar proportion of patients with pain on palpation and unilateral edema (6.6%) as in the studies by Douma et al 23 (5.8%) and van Belle et al 10 (5.7%), in which a clinical prediction rule was determined prospectively and venous ultrasonography was not performed. Thus, we believe the diagnosis of deep vein thrombosis without CT pulmonary angiography when pulmonary embolism was suspected had little influence on our measurement of concordance with recommendations for CT pulmonary angiography. Finally, our study included 2 institutions in the same metropolitan area staffed by the same emergency physicians and radiologists. Therefore, our results may not be representative of other settings. CONCLUSIONS CT pulmonary angiogram examinations are often not concordant with expert recommendations and guidelines for investigation of suspected pulmonary embolism. Emergency department physicians order many CT pulmonary angiograms when pulmonary embolism is unlikely and sensitive D-dimer test results are negative. This practice lowers the diagnostic yield of CT pulmonary angiograms and exposes some patients to risks associated with CT pulmonary angiography, including false-positive test results and unnecessary treatment. References 1. Amis ES Jr, Butler PF, Applegate KE, et al. American College of Radiology white paper on radiation dose in medicine. J Am Coll Radiol. 2007;4:272-284. 2. Einstein AJ, Henzlova MJ, Rajagopalan S. Estimating risk of cancer associated with radiation exposure from 64-slice computed tomography coronary angiography. JAMA. 2007;298:317-323. 3. Lee CI, Haims AH, Monico EP, et al. Diagnostic CT scans: assessment of patient, physician, and radiologist awareness of radiation dose and possible risks. Radiology. 2004;231:393-398. 4. Corwin MT, Donohoo JH, Partridge R, et al. Do emergency physicians use serum D-dimer effectively to determine the need for CT when evaluating patients for pulmonary embolism? Review of 5,344 consecutive patients. AJR Am J Roentgenol. 2009;192:1319-1323. 5. Kline JA, Courtney DM, Beam DM, et al. Incidence and predictors of repeated computed tomographic pulmonary angiography in emergency department patients. Ann Emerg Med. 2009;54:41-48. 6. Stein PD, Woodard PK, Weg JG, et al. Diagnostic pathways in acute pulmonary embolism: recommendations of the PIOPED II investigators. Am J Med. 2006;119:1048-1055. 7. Torbicki A, Perrier A, Konstantinides S, et al. Guidelines on the diagnosis and management of acute pulmonary embolism: the Task Force for the Diagnosis and Management of Acute Pulmonary Embolism of the European Society of Cardiology (ESC). Eur Heart J. 2008;29:2276-2315. 8. Fesmire FM, Brown MD, Espinosa JA, et al. Critical issues in the evaluation and management of adult patients presenting to the emergency department with suspected pulmonary embolism. Ann Emerg Med. 2011;57:628-652 e675. 9. Anderson DR, Kahn SR, Rodger MA, et al. Computed tomographic pulmonary angiography vs ventilation-perfusion lung scanning in patients with suspected pulmonary embolism: a randomized controlled trial. JAMA. 2007;298:2743-2753. 10. van Belle A, Buller HR, Huisman MV, et al. Effectiveness of managing suspected pulmonary embolism using an algorithm combining clinical probability, D-dimer testing, and computed tomography. JAMA. 2006;295:172-179. 11. Righini M, Le Gal G, Aujesky D, et al. Diagnosis of pulmonary embolism by multidetector CT alone or combined with venous ultrasonography of the leg: a randomised non-inferiority trial. Lancet. 2008;371:1343-1352. 12. Perrier A, Roy PM, Sanchez O, et al. Multidetector-row computed tomography in suspected pulmonary embolism. N Engl J Med. 2005; 352:1760-1768. 13. Gilbert EH, Lowenstein SR, Koziol-McLain J, et al. Chart reviews in emergency medicine research: where are the methods? Ann Emerg Med. 1996;27:305-308. 14. Badcock D, Kelly A-M, Kerr D, Reade T. The quality of medical record review studies in the international emergency medicine literature. Ann Emerg Med. 2005;45:444-447. 15. Worster A, Bledsoe RD, Cleve P, et al. Reassessing the methods of medical record review studies in emergency medicine research. Ann Emerg Med. 2005;45:448-451. 16. Lowenstein SR. Medical record reviews in emergency medicine: the blessing and the curse. Ann Emerg Med. 2005;45:452-455. 17. Le Gal G, Righini M, Roy PM, et al. Prediction of pulmonary embolism in the emergency department: the revised Geneva score. Ann Intern Med. 2006;144:165-171. 18. Evans RS, Sharp JH, Linford LH, et al. Risk of symptomatic DVT associated with peripherally inserted central catheters. Chest. 2010; 138:803-810.

42 The American Journal of Medicine, Vol 126, No 1, January 2013 19. Chunilal SD, Eikelboom JW, Attia J, et al. Does this patient have pulmonary embolism? JAMA. 2003;290:2849-2858. 20. Rodger MA, Maser E, Stiell I, et al. The interobserver reliability of pretest probability assessment in patients with suspected pulmonary embolism. Thromb Res. 2005;116:101-107. 21. Lucassen W, Geersing GJ, Erkens PM, et al. Clinical decision rules for excluding pulmonary embolism: a meta-analysis. Ann Intern Med. 2011;155:448-460. 22. Wells PS, Anderson DR, Rodger M, et al. Derivation of a simple clinical model to categorize patients probability of pulmonary embolism: increasing the models utility with the SimpliRED D-dimer. Thromb Haemost. 2000;83:416-420. 23. Douma RA, Mos IC, Erkens PM, et al. Performance of 4 clinical decision rules in the diagnostic management of acute pulmonary embolism: a prospective cohort study. Ann Intern Med. 2011;154:709-718. 24. Courtney DM, Kline JA, Kabrhel C, et al. Clinical features from the history and physical examination that predict the presence or absence of pulmonary embolism in symptomatic emergency department patients: results of a prospective, multicenter study. Ann Emerg Med. 2010;55:307-315 e301. 25. Kooiman J, Klok FA, Mos IC, et al. Incidence and predictors of contrastinduced nephropathy following CT-angiography for clinically suspected acute pulmonary embolism. J Thromb Haemost. 2010;8:409-411. 26. Lieberman PL, Seigle RL. Reactions to radiocontrast material. Anaphylactoid events in radiology. Clin Rev Allergy Immunol. 1999;17: 469-496. 27. Berrington de Gonzalez A, Mahesh M, Kim KP, et al. Projected cancer risks from computed tomographic scans performed in the United States in 2007. Arch Intern Med. 2009;169:2071-2077. 28. Smith-Bindman R, Lipson J, Marcus R, et al. Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch Intern Med. 2009;169:2078-2086. 29. Shrimpton PC, Hillier MC, Lewis MA, Dunn M. National survey of doses from CT in the UK: 2003. Br J Radiol. 2006;79:968-980. 30. Ranji SR, Shojania KG, Trowbridge RL, Auerbach AD. Impact of reliance on CT pulmonary angiography on diagnosis of pulmonary embolism: a Bayesian analysis. J Hosp Med. 2006;1:81-87. 31. Cronin P, Kelly AM. Influence of population prevalences on numbers of false positives: an overlooked entity. Acad Radiol. 2011;18:1087-1093. Appendix 1 Revised Geneva Score and Other Data Elements and Method of Collection Data Element Electronic Collection Age Prior VTE Surgery or fracture in prior month Active malignant condition Unilateral leg pain Hemoptysis Heart rate Pain on palpation and unilateral edema Trauma Pregnancy VTE venous thromboembolism. Manual Review and Collection