Journal of Thrombosis and Haemostasis, 1: 652±657
IN FOCUS
Contribution of indirect computed tomography venography to computed tomography angiography of the chest for the diagnosis of thromboembolic disease in two United States emergency departments P. B. RICHMAN, J. WOOD, D. M. KASPER, J. M. COLLINS, R. W. PETRI, A. G. FIELD, D . N . C O W L E S and J . A . K L I N E y
Departments of Emergency Medicine and Radiology; and Mayo Clinic Hospital, Scottsdale, Arizona, yDepartment of Emergency Medicine;
Carolinas Medical Center, Charlotte, NC, USA Please see also Moll S. Use of combined CT venography and CT pulmonary arteriography. This issue, pp. 637±639.
Summary. Recent reports suggest that physicians in non-ambulatory settings can use indirect CT venography (CTV) of the lower extremities immediately following spiral CT angiography (CTA) of the chest to identify patients with a negative CTA who have thromboembolic disease identi®ed on CTV. We sought to determine the frequency of isolated deep venous thrombosis (DVT) discovered on CTV in emergency department (ED) patients with complaints suggestive of pulmonary embolism (PE) yet having a negative CTA. This study was conducted in a suburban and urban ED where patients with symptoms suspicious for PE were primarily evaluated with CTA and CTV. A total of 800 patients were studied, including 360 from the suburban ED and 440 from the urban ED. 88 (11%) patients were diagnosed with thromboembolic disease by CTA, or CTV, or both. Seventy-three patients had a CTA of the chest that was positive for PE, 42 (5.2%) of whom had evidence of both PE on CTA and DVT on CTV. Fifteen patients (2%, 95% CI 1±3%) had a negative CTA and were subsequently found to have isolated DVT on CTV, all of whom received anticoagulation therapy. These data suggest that indirect CT venography of immediately following CT angiography of the chest signi®cantly increased the frequency of diagnosed thromboembolic disease requiring anticoagulation in ED patients with suspected PE. Keywords: computed tomography, CT venography, deep venous thrombosis, pulmonary embolism.
challenging for Emergency Physicians. Many tests, including spiral CT angiography (CTA) of the chest, pulmonary ventilation-perfusion scanning, and lower extremity duplex imaging are routinely used in the emergent setting to evaluate patients with suspected PE. However, none of these tests alone can reliably exclude pulmonary embolism with adequate sensitivity to enable the clinician to withhold anticoagulation for patients without some element of uncertainty and, perhaps, risk of morbidity and mortality [1]. Recent evidence from the radiological literature suggests that indirect CT venography (CTV) of the lower extremities following spiral CTA can increase the sensitivity for the diagnosis of thromboembolic disease when compared with spiral CTA alone when studied inpatients or in combinations of inpatients and ambulatory patients [2±10]. However, to date, this ®nding has not been validated in the Emergency Department (ED) setting. Although exact statistics are lacking, we are aware that many centers evaluate patients with suspected PE using CTA, but do not routinely employ the CTV phase. A substantial body of evidence suggests that DVT is frequently discovered in patients with suspected PE, but for whom radiographic evidence of PE is absent (i.e. isolated DVT) [11±13]. No comparable data from a large study of outpatients are available to estimate the differential contribution of CTV to CTA. The speci®c aim of this study was to measure the frequency of isolated DVT discovered on CTV in ED patients with suspected PE and a negative CTA. Methods
Introduction The diagnosis of thromboembolic disease, including pulmonary embolism (PE) and deep venous thrombosis (DVT), remains Correspondence: Jeffrey A. Kline, Director of Research, Department of Emergency Medicine, 1000 Blythe Blvd, Charlotte NC 28203, USA. Tel.: 1 704 3558155; fax: 1 704 3557047; e-mail:
[email protected] Received 4 October 2002, accepted 7 October 2002
Study design
This study was conducted by a retrospective review of the electronic medical and radiological records of a cohort of ED patients with suspected PE. Setting
The study was conducted in both a suburban and urban ED. The suburban site at the Emergency Department of Mayo Clinic # 2003 International Society on Thrombosis and Haemostasis
CTV and pulmonary embolism 653
Hospital (Scottsdale, AZ, USA), is an academic, suburban ED, with an annual patient census of approximately 25 000. The urban ED is at Carolinas Medical Center (Charlotte, NC, USA), a teaching hospital with an annual ED census of 107 000 patients during the study period. Carolinas Medical Center is owned and operated by the Carolinas HealthCare (CHS) system, which also operates three of ®ve other hospitals in Charlotte and a network of outpatient clinics. Results of diagnostic testing and electronic medical records from all CHS hospitals and clinics are available through a centralized computer system. The study was approved by the respective institutional Review Boards at both centers prior to initiation. Population
In both settings, all ED patients with symptoms suspicious for PE for whom the Emergency Physician ordered a PE protocol (see description of radiological technique below) were eligible for analysis. All eligible patients were identi®ed by use of a computerized radiological database. The PE protocol is the diagnostic modality of choice at both institutions for the ED diagnosis of thromboembolic disease. Patients were selected on the basis of standard clinical history and physical examination by an emergency medicine resident or attending physician that uncovered enough suspicion of PE to prompt the order of a CT scan to rule out PE. Only patients with suspected PE who were pregnant, had contraindications to contrast, or were too heavy to be moved by the CT table (> 200 kg) were evaluated by other imaging techniques. At the suburban center, ED patients have been evaluated for PE with CTA and CTV since the inception of the hospital in 1999. At the urban center, patients have been evaluated for PE with CTA and CTV since January 2001. At the suburban center, CTA and CTV studies from January 2001±December 2001 were identi®ed, and at the urban center, studies from January 2001±August 2001 were identi®ed. Study protocol
Suburban setting An electronic radiological database was queried to identify all eligible patients and the ®nal PE protocol interpretation by the radiology department for each case. Final CT readings were abstracted by a trained, physician data abstractor on a structured data collection instrument. Via search of the Mayo Clinic Hospital electronic medical record, the ED and hospital course for all study patients were reviewed in a structured fashion by a trained, physician data abstractor. Relevant demographic, historical, and physical examination, ®ndings, as well as clinical course, were reported on a closedquestion data collection instrument. A ®nal diagnosis of pulmonary embolism was based on the discharge diagnosis obtained from the medical record. Urban setting As a standing policy, ED physicians prospectively completed a one-page data collection form to record clinical data (vital signs, symptoms and risk factors) and the preliminary results of imaging studies on all patients # 2003 International Society on Thrombosis and Haemostasis
evaluated for PE. To ensure identi®cation of all ED patients who underwent CTA-CTV to rule out PE, we used three methods: (i) evaluation of the prospective data sheets; (ii) Query of the QuadRIS# electronic radiological archiving system; and (iii) A medical student (DNC) examined the logbooks kept by the CT technicians that identi®es each CT examination requested, the imaging protocol used, the physician who ordered the test and where the patient was physically located when the test was ordered. Chart review was performed by a medical student to complete the clinical dataforms on patients who were discovered to have had CTA-CTV for PE ordered in the ED, but for whom the clinical dataform was not prospectively completed. We did not examine interobserver variability for this subgroup. Length of stay in the ED was determined from the ED computer tracking system as the difference between sign-in time and the time the patient physically left the ED. Scans were read as positive for PE or DVT based upon presence of a ®lling defect in the pulmonary arterial tree as described by Remy-Jardin et al. [14] or a venous ®lling defect on CTV as described by Cham et al. [9] and Garg et al. [15] All data were recorded in spreadsheet form to allow summary computations and basic statistical analysis (Excel 2000, Microsoft Inc). Unless otherwise noted, the imaging protocols for CTA and CTV were the same at both institutions. The PE protocols included CTA of the chest immediately followed by indirect lower extremity CTV in all cases. All radiographic images were obtained using the Siemens Volume Zoom Four Detector Row CT Scanner or a GE QX/I light speed scanner. For CT chest angiography, depending on the patient's ability to hold their breath, either 1 mm collimation for a 30-s breath hold or 2.5 mm collimation for a 15-s breath hold was chosen with an adaptive pitch (kVp 140). mAs selection is proportional to patient's body size. For average size patient with FOV 35, selection for mAs 122, and CTDI 20 mGy. Gantry rotation speed was 0.75 s. From either 1 mm or 2.5 mm collimation, reconstructions at 3 mm slice thickness at 3 mm increments are made for interpretation. Images were then reconstructed at 1.25 mm. Depending on the length of the breath hold, the volume of dye injected was limited to maintain the bolus during the scan acquisition time. The radiologist or technician injected up to 150 cc Omnipaque 350 at 3 cc per second. In younger patients under the age of 65 that have normal cardiac function the scan was begun at 18 s after the start of the injection. In patients over the age of 65 or any patient with poor cardiac function, the scan was begun at 22 s after the start of the injection. For indirect CTV, in patients younger than 65 years with good cardiac output the lower extremities were scanned at 3 min from the mid-calves to the top of acetabulum using 2.5 mm collimation to create 10 mm slice thickness at 20 mm increments. For patients older than 65 years or any patient with poor cardiac output, the legs were scanned and lower pelvis at 5 min. All CT images were reviewed by an experienced, board certi®ed radiologist with fellowship training in body CT.
654 P. B. Richman et al Data analysis
Data were entered into Microsoft Excel 2000 for Windows (Microsoft, Redmond, WA, USA) and transported into SPSS for Windows (SPSS Inc., Chicago, IL, USA) for statistical analysis. Categorical data is reported as frequency of occurrence; continuous data as means standard deviations. 95% Con®dence Intervals for means and proportions were calculated using the exact normal or binomial methods. Interobserver agreement between data abstractors was assess by calculation of kappa values. The following qualitative terms were correlated to kappas: 0.0±0.19 `slight'; 0.2±0.39 `fair'; 0.4±0.59 `moderate'; 0.6±0.79 `substantial'; and 0.8± 1.0 `almost perfect'. The primary outcome parameter of the study was the frequency of positive indirect CTV of the lower extremities following negative CTA of the chest. Results During the 12-month study period, 360 patients were evaluated in the suburban ED by the PE protocol (1.4% of all ED visits). In 20 cases, radiological records were incomplete. None of these 20 were diagnosed with thromboembolic disease. Complete radiological readings were available within the electronic database for 340 of these patients who comprised the ®nal study group. The mean age was 61 18 years, 58% were female, and the mean initial pulse oximetry recording on room air was 95 5%. 228 (67%, 95% CI 61% to 72%) of patients were recorded to have had chest pain, 265 (78%, 95% CI 73% to 82%) had shortness of breath, and 160 (47%, 95% CI 42% to 53%) had both chest pain and shortness of breath. 173 (51%, 95% CI 46% to 56%) of the patients were ultimately admitted to the hospital; 313 (92%, 95% CI 87% to 96%) survived until discharge. At the urban ED, 440 patients were evaluated by the CTACTV protocol for PE, representing 0.62% of all ED visits during the 8-month study period. An additional 13 ED patients underwent VQ scanning only. These 13 were not included in data analysis. Prospectively completed data forms were available for 393 patients (including all patients ultimately diagnosed with PE) and chart review was completed for 47. The mean age was 48 16 years, 58% were female, and the mean initial pulse oximetry recording on room air was 96.5 4.3%. 301 (68%, CI 64% to 73% of patients were recorded to have had pleuritic or substernal chest pain 300 (68%, CI 64% to 73%) had shortness of breath, and 227 (52%, 95% CI 47% to 56%) had both chest pain and shortness of breath. 247 patients were discharged home primarily (median length of ED stay 368 min) and 41 patients were observed in a chest pain rule out center overnight in the ED and then discharged home, 152 (35%, 95% CI 30% to 39%) were ultimately admitted to the hospital. Fifty-one (15%; 95% CI 11±19%) patients from the suburban ED and 37(8.4%; 95% CI 6±11%) in the urban ED were, respectively, diagnosed with thromboembolic disease by the CT protocol (Table 1). Forty-two (12%; 95% CI 9±16%)
Table 1 Diagnosis of PE and DVT by contrast-enhanced CT angiography (CTA) and CT venography (CTV) CT results
Suburban Center
Urban Center
Combined
CTA and CTV CTA and CTV CTA and CTV CTA and CTV
25 17 9 309
17 14 6 403
42 31 15 712
Total
360
(7%) (5%) (3%) (85%)
(4%) (3%) (1%) (92%)
440
(5%) (4%) (2%) (89%)
800
Table 2 Anatomic location of PE in patients with CT angiography read as positive for PE Location
n
%
Right lung only Left lung only Both lungs
22 8 43
30 11 59
Total
73
100
95% CI 20±42 5±21 47±70 ±
patients in the suburban ED and 31 (7%; 95% CI 5±10%) in the urban ED had a CT angiogram of the chest that was positive for PE for a total of 73 patients with PE. Twenty-®ve (7%; 95% CI 5±11%) patients in the suburban ED and 17 (4%; 95% CI 2±6%) patients in the urban ED had positive ®ndings for both PE on CTA and DVT on CTV. Table 2 summarizes the anatomic location of PEs in the 73 patients with con®rmed ®lling defects in the pulmonary vasculature. Nine patients (3%; 95% CI 1±5%) in the suburban group and six patients (1%; 95% CI 1±3%) in the urban center patients with a CTA negative for PE but were found to have a DVT on CTV. Thus, nearly 18% of the patients in the suburban group and 16% of patients in the urban center that were diagnosed with thromboembolic disease by the CT protocol had negative chest CTs. In total, at both centers, 800 patients were examined for suspected PE by a combined CTA-CTV study. The CTACTV demonstrated thromboembolic disease in 88 patients. In 15 cases (2%, 95% CI 1±3%), the CTA-CTV demonstrated isolated deep venous thrombosis with no evidence of a PE in the pulmonary vasculature. All 15 patients were treated with heparin and warfarin anticoagulation. Discussion The limitations of the diagnostic algorithm examined in the PIOPED study (ventilation-perfusion scanning with adjunctive pulmonary angiography as needed) have been well described previously [1,16]. We believe that the ventilation-perfusion lung scan as a primary method to evaluate for PE in the ED has many insuf®ciencies in terms of operational ef®ciency and diagnostic accuracy. Over the past several years, chest CTA has gradually replaced VQ scanning as the initial procedure of choice at many institutions in the initial radiological evaluation of a patient with suspected PE. In a recent European trial, 410 patients were evaluated with CTA and VQ scanning, with # 2003 International Society on Thrombosis and Haemostasis
CTV and pulmonary embolism 655
pulmonary angiography performed in non-diagnostic and contradictory studies. CTA alone was found to be 88% sensitive and 94% speci®c for the diagnosis of PE [17]. Similar, sensitivities and speci®cities have been observed in all but one of the series found in the literature [14]. To date, however, investigators have not validated a strategy to utilize CTA in an algorithm to exclude thromboembolic disease [14,18±29]. In a recent survey of 30 academic EDs, we found that 53% used CTA only as the primary method to evaluate for suspected PE, whereas only 18% reported use of CTA with CTV [30]. A substantial body of evidence suggests that DVT is frequently discovered in patients with suspected PE, but for whom radiographic evidence of PE is absent (i.e. isolated DVT). In a metaanalysis of inpatient and outpatients with suspected PE, Van Rossum et al. computed a 3% and 7% rate of simultaneous, isolated DVT in patients with a normal ventilation-perfusion lung scans or normal pulmonary angiograms, respectively [11]. Studies of outpatients evaluated for PE with ventilation-perfusion lung scanning and selected use of pulmonary angiography found between a 2% and 4% rate of isolated DVT [12,13] without visual evidence of PE. No comparable data from a large study of outpatients are available to estimate the differential contribution of CTV to CTA. Recently, several investigators, reporting in the radiological literature in primarily hospitalized patient populations, have described evidence that CTV performed after a single dose of contrast for chest angiography allows for accurate simultaneous evaluation for both DVT and PE, perhaps, obviating the need for duplex ultrasonography in high-risk patients [3±10]. These studies have shown the accuracy and reliability of CTV to equal that of venous duplex ultrasonography for the lower extremities. In the largest report, to date, Loud et al. reported their ®ndings in 650 patients for whom venous phase images were acquired from the diaphragm to the upper calves after completion of CT pulmonary angiography [10]. For 308 of the patients within the study group results of the CTV were also compared with those of bilateral-lower extremity venous sonography. A total of 116 patients within the overall study group were diagnosed with thromboembolic disease including 27 patients with PE alone, 31 patients with DVT alone, and 58 patients with both. In the group for which lower extremity venous ultrasound was also obtained, CTV had a sensitivity of 97% and speci®city of 100% for femoropoliteal DVT. In the other large series previously reported, Cham et al. prospectively enrolled 541 consecutive patients who received pulmonary CTA followed by indirect CT venography from the pelvis to the popliteal fossa [9]. Among the 45 patients with DVT who were diagnosed, 16 had no evidence of PE on pulmonary CTA, revealing an increase in the diagnosis of thromboembolic disease of 18%. In addition, among a subset of patients (116) who also underwent lower extremity ultrasound, 15 had DVT on both US ultrasound and indirect CTV. In four other cases DVT was seen only on indirect CTV. Our study differs from that which has been reported previously for several reasons. First, our population represents one # 2003 International Society on Thrombosis and Haemostasis
of the largest study groups of patients with symptoms suspicious for PE who have undergone CTA and indirect CTV. The samples were drawn from two disparate populations in terms of geography and demographics. We reviewed a total of 800 ED patients with suspected PE at two centers, of which 88 were diagnosed with thromboembolic disease, including 73 with PE and 15 with isolated DVT. We believe that these data and associated 95% con®dence intervals help to show that the venous phase of the CT protocol signi®cantly increases the overall rate of detection of thromboembolic disease in ED patients with suspected PE. Second, our study group is also unique because it represents only ED patients. In a similar fashion to many other disease processes, the risks and clinical presentation of patients with thromboembolic disease may be different from that of inpatient and other ambulatory populations. In particular, we believe it was necessary to document the frequency with which the CTA-CTV protocol would disclose isolated DVT in patients with suspected PE. Knowledge of this frequency is important to help radiology and ED department managers decide if the additional radiation incurred by the indirect venography imaging sequence and the extra time and cost to acquire these images can be justi®ed. In the ED setting, CTV may provide logistical advantages over ultrasound imaging of the lower extremities. First, addition of indirect CTV to CTA requires only that additional CT slices be obtained through the legs, but does not require additional contrast. Therefore, a patient who is in the CT suite does not need to be transferred to another section of the hospital to have an ultrasound performed. Rather, additional images of the lower extremities may be obtained just minutes after the CTA is performed. Second, even in tertiary care institutions (including the study centers), ultrasound technicians are not `in-house' 24 h a day, 7 days a week. Thus, particularly during evening and weekend hours use of CTA followed by CTV allows for increased diagnostic sensitivity and, yet, requires the presence of only one radiological technician on-site. We believe these bene®ts of CTV outweigh the risk of additional gonadal radiation induced during the CTV imaging sequence. Many patients are discharged to home from the ED after an evaluation for the diagnosis of pulmonary embolism. In modern practice, many ED patients do not have certain ability to avail themselves of medical follow-up. Thus, the diagnostic accuracy of tests used for ED evaluation must be speci®cally investigated under the assumption that the physician has one chance to `get it right'. In the present sample, 15 of 800, or 2% of patients with negative helical CTA of the chest alone would have been wrongly classi®ed as not having thromboembolic disease if no further testing were performed. This ®nding agrees with the published 2±4% rate of isolated DVT discovered in outpatients with suspected PE, but who had PE ruled out on the basis of ventilation-perfusion scintigraphy with selective use of pulmonary angiogaphy [12,13]. The present study did not determine if the patients with isolated DVT discovered on CTV would have been discharged in the absence of the information provided by CTV. Nonetheless, if we assume that each additional case of DVT that was
656 P. B. Richman et al
diagnosed on the basis of CTV represents an additional true positive case that warranted anticoagulation, the relative increase in diagnostic sensitivity of CTV was 15 out of 88, or 17% (95% CI 10±27%). This datum suggests a signi®cant contribution of CTV to CTA of the chest in the decision of whether or not administer anticoagulation therapy to a patient with suspected PE. This report summarizes the ®rst of a two-phase study. This phase of our investigation helps lay the scienti®c groundwork for testing of a diagnostic algorithm that includes CT chest angiography and includes CTV of the lower extremities to exclude the diagnosis of PE in patients with validated low-risk categorizations. This study had several limitations. Our retrospective design undoubtedly led to unintended selection bias in both centers. The decision to consider the diagnosis of pulmonary embolism for a given patient is highly individual to the practitioner. To our knowledge and to date, investigators have yet to de®ne standardized, entry criteria for patients in thromboembolic research that are easily reproduced. Preliminary studies suggest that, when Emergency Physicians utilize risk assessment criteria de®ned by specialists from other ®elds of medicine, their ability to appropriately apply the rules are variable [31,32]. Thus, the assessment of risk for pulmonary embolism in ED patients for the purposes of research remains somewhat subjective even in prospective investigations. At the suburban center we did not track patients with clinical symptoms suspicious for PE who underwent other diagnostic modalities. We suspect that very few cases were evaluated by alternative means. The PE protocol was the diagnostic modality of choice at both EDs to evaluate patients for possible PE. Only pregnant patients and patients with contraindications to contrast were evaluated by ventilation-perfusion scanning. Another limitation of our study is that we have no criterion standard to de®ne sensitivity and speci®city. Only 0.6% patients in the suburban series of patients had a subsequent standard pulmonary angiogram during their evaluation. Thus, the ®nal de®nition of pulmonary embolism for nearly all of our patients was based on the ®nal diagnosis from the medical record for which we found no divergence when compared with the CT diagnoses. In the urban center we employed prospective and retrospective methods to ®nd all known cases of PE or DVT that were diagnosed within the subsequent 3 months. Limitations with the data collection method include the absence of interobserver evaluation for the interpretation of the CT scan. Blachere et al. found excellent interobserver agreement (kappa 0.72) for the diagnosis of PE via CTA [27]. Similarly, Garg et al. found moderately good concordance for indirect CTV (kappa 0.59) [7]. Also, we did not explicitly and thoroughly track the rate of false positive readings for the initial CT results interpreted by board-certi®ed radiologists who did not have specialty training in body CT interpretation. However, all patients at the urban center had clinical dataforms prospectively completed by the ED clinicians before and after knowledge of CT scan results. The clinicians indicated the result of the initial reading of the CT on this dataform. In six
cases, clinicians wrote that the CT scans were initially interpreted as positive for ®lling defects in the lung vasculature and initiated anticoagulation, but our review found that these six were subsequently overread as negative for PE. In no case was an initially positive DVT overread as negative. Finally, this study design is limited to de®ning whether CTV after CTA increases the diagnostic frequency of thrombembolic disease but will not adequately address whether the protocol can be used to guide the decision to withhold anticoagulation. This study does not deterimine if CTV can substitute for a decision rule to rule out DVT, or test if CTV is safer, cheaper or more ef®cient that the combination of CTA followed by duplex venous ultrasonography. These questions will require further study. Conclusion
Indirect CTV of the lower extremities immediately following spiral CTA of the chest signi®cantly increased the frequency of diagnosed thromboembolic disease that required anticoagulation in ED patients with suspected PE. Acknowledgments The authors would like to thank Joseph Hentz, MS and Jose L. Hernandez, BA for their assistance with data analysis. References 1 Wolfe TR, Hartsell SC. Pulmonary embolism: making sense of the diagnostic algorithm. Ann Emerg Med 2001; 37: 504±14. 2 Garg K, Kemp JL, Wojcik D, Hoehn S, Johnston RJ, Macey LC, Baron AE. Thromboembolic disease: comparison of combined CT pulmonary angiography and venography with bilateral leg sonography in 70 patients. AJR 2000; 175: 997±1001. 3 Ferretti GR, Ayanian D, Ranchoup Y, Thony F, Bosson JL, Coulomb M. CT X-ray evaluation of abdominal and pelvic veins in patients suspected of acute pulmonary embolism with negative Doppler sonography. J Radiol 1998; 79: 327±30. 4 Au VW, Walsh G, Fon G. Computed tomography pulmonary angiography with pelvic venography in the evaluation of thrombo-embolic disease. Australas Radiol 2001; 45: 141±5. 5 Loud P, Grossman Z, Klippenstein D, Ray C. Combined CT venography and pulmonary angiography: a new diagnostic technique for suspected thromboembolic disease. AJR 1998; 170: 951±4. 6 Coche E, Hamoir X, Hammer F, Hainaut P, Goffette P. Using dualdetector helical CT angiography to detect deep venous thrombosis in patients with suspicion of pulmonary embolism: Diagnostic value and additional ®ndings. AJR 2001; 176: 1035±9. 7 Garg K Kemp JL, Russ PD, Baron AE. Thromboembolic disease. variability of interobserver agreement in the interpretation of CT venography with CT pulmonary angiography. Am J Roentgenol 2001; 176: 1043±7. 8 Ghaye B, Szapiro D, Willems V, Dondelinger RF. Combined CT venography of the lower limbs and spiral CT angiography of pulmonary arteries in acute pulmonary embolism: preliminary results of a prospective study. JBR-BTR 2000; 83: 271±8. 9 Cham M, Yankelevitz D, Shaham D, Shah A, Sherman L, Lewis A, Rademaker J, Pearson GCJ, Wolff W, Prabhu P, Galanski M, Clark R, Sostman H, Henschke C. Deep venous thrombosis: Detection by using indirect CT venography. Radiology 2000; 216: 744±51. # 2003 International Society on Thrombosis and Haemostasis
CTV and pulmonary embolism 657 10 Loud P, Katz D, Bruce D, Klippenstein D, Grossman Z. Deep venous thrombosis with suspected pulmonary embolism: detection with combined CT venography and pulmonary angiography. Radiology 2001; 9: 498±502. 11 van Rossum AB, van Houwelingen HC, Kieft GJ, Pattynama PM. Prevalence of deep vein thrombosis in suspected and proven pulmonary embolism: a meta-analysis. Br J Radiol 1998; 71: 1260±5. 12 MacGillary MR, Sanson B-J, Buller HR, Brandjes DPM. Compression ultrasonography of the leg veins in patients with clinically suspected pulmonary embolism. Is a more extensive assessment of compressibility useful? Thromb Haemost 2002; 84: 973±6. 13 Daniel KR, Jackson RE, Kline JA. Utility of the lower extremity venous ultrasound in the diagnosis and exclusion of pulmonary embolism in outpatients. Ann Emerg Med 2000; 35: 547. 14 Remy-Jardin M, Remy J, Deschildre F, Artaud D, Beregi JP, HosseinFoucher C, Marchandise X, Duhamel A. Diagnosis of pulmonary embolism with spiral CT: comparison with pulmonary angiography and scintigraphy. Radiology 1996; 200: 699±706. 15 Garg K, Mao J. Deep venous thrombosis. Spectrum of ®ndings and pitfalls in interpretation on CT venography. Am J Roentgenol 2001; 177: 319±23. 16 The PIOPED Investigators. Value of the ventilation/perfusion scan in acute pulmonary embolism. JAMA 1990; 263: 2753±9. 17 Holbert JM, Costello PFMP. Role of spiral computed tomography in the diagnosis of pulmonary embolism in the Emergency Department. Ann Emerg Med 1999; 33: 520±8. 18 van Rossum AB, Pattynama PMT, Ton ERTA, Treurniet FEE, Arndt JW, van Eck B, Kieft GJ. Pulmonary embolism: validation of spiral CT angiography in 149 patients. Radiology 1996; 201: 467±70. 19 Goodman LR, Curtin JJ, Mewissen MW, Foley WD, Lipchik RJ, Crain MR et al. Detection of pulmonary embolism in patients with unresolved clinical and scintigraphic diagnosis: helical CT versus angiography. AJR 1995; 164: 1369±74. 20 Teigen CL, Maus TP, Sheedy PF, Stanson AW, Johnson CM, Breen JF, McKusick MA. Pulmonary Embolism: Diagnosis with contrast-enhanced electron-beam CT and comparison with pulmonary angiography. Radiology 1995; 194: 313±9. 21 van Rossum AB, Treurniet FEE, Kieft GJ, Smith SJ, Shepers-Bok R. Role of spiral volume computed tomographic scanning in the assessment of patients with clinical suspicion of pulmonary embolism
# 2003 International Society on Thrombosis and Haemostasis
22
23
24 25
26 27
28
29
30 31 32
and an abnormal ventilation-perfusion lung scan. Thorax 1996; 51: 23±8. Mayo JR, Remy-Jardin M, Muller NL, Remy J, Worsley DF, HosseinFoucher C et al. Pulmonary Embolism: Prospective Comparison of Spiral CT with Ventilation-Perfusion Scintigraphy. Thoracic Radiol 1997; 205: 447±52. Ferretti GR, Bosson JL, Buffaz PD, Ayanian D, Pison C, Blanc F, Carpentier F, Carpentier P, Coulomb M. Acute pulmonary embolism: role of helical CT in 164 patients with intermediate probability at ventilation-perfusion scintigraphy and normal results at duplex US of the legs. Radiology 1997; 205: 453±8. Cross JJL, Kemp PM, Walsh CG, Flower CDR, Dixon AK. A randomized trial of spiral CT and ventilation perfusion scintigraphy for the diagnosis of pulmonary embolism. Clin Radiol 1998; 53: 177±82. Drucker EA, Rivitz SM, Shepard JO, Boiselle PM, Trotman-Dickenson B, Welch TJ, Maus TP, Miller SW, Kaufman JA, Waltman AC, McLoud TC, Athanasoulis CA. Acute pulmonary embolism: assessment of helical CT for diagnosis. Radiology 1998; 209: 235±41. Kim K, Muller NL, Mayo JR. Clinically suspected pulmonary embolism: utility of spiral CT. Radiology 1999; 210: 693±7. Garg K, Sieler H, Welsh CH, Johnston RJ, Russ PD. Clinical validity of helical CT being interpreted as negative for pulmonary embolism: implications for patient treatment. Am J Roentgenol 1999; 172: 1627±31. Blachere H, Latrabe V, Montaudon M, Valli N, Couf®nhal T, Raherisson C, Leccia F, Laurent F. Pulmonary embolism revealed on helical CT angiogrpahy: comparison with ventilation-perfusion radio nuclide lung scanning. Am J Roentgenol 2000; 174: 1041±7. Perrier A, Howarth N, Didler D, Loubeyre P, Unger PF, deMoerloose P, Slosman D, Junod A, Bounameaux H. Performance of helical computed tomography in unselected outpatients with suspected pulmonary embolism. Ann Inter Med 2001; 135: 88±97. Kline JA, Jones AE. Availability of technology to image and risk stratify patients with pulmonary embolism in academic emergency departments. Acad Emerg Med 2003; 10 (Abstract). Jackson RE, Rudoni RR, Pascual R. Emergency physician assessment of the pre-test probability of pulmonary embolism (Abstract). Acad Emerg Med 1999; 6: 437. Wolf SJ, McCubbin TR, Feldhaus KM, Faragher JP, Adcock D. The role of D-dimer and Well's criteria in patients with suspected pulmonary embolism. Acad Emerg Med 2002; 9: 362 (Abstract).
