- Email: [email protected]
Detecting iliac vein thrombosis with current protocols of lower extremity venous duplex ultrasound
Detecting iliac vein thrombosis with current protocols of lower extremity venous duplex ultrasound Ashish K. Jain, MD,a Michael C. Soult, MD,a Scott A. Resnick, MD,b Kush Desai, MD,b Patricia Astleford, RN, RVT,c Mark K. Eskandari, MD,a and Heron E. Rodriguez, MD,a Chicago, Ill
ABSTRACT Background: Deep venous thrombosis isolated to the iliac veins is uncommon. Venous duplex ultrasound (DU) is widely accepted as the screening modality of choice for lower extremity deep venous thromboses. This investigation evaluated the accuracy and efficacy of DU in diagnosis of iliac vein thrombosis. Methods: We conducted a single-center retrospective review of patients who were diagnosed with iliac vein thrombosis between January 1, 2006, and December 31, 2015. Patients included in our analysis needed to have both DU and crosssectional imaging performed within a month of each other. The efficacy of DU in diagnosis of iliac vein thrombosis was determined using cross-sectional imaging as a standard for diagnosis. Results: In total, our query yielded 80 patients with a diagnosis of iliac vein thrombosis in the medical chart; 48 patients had both cross-sectional imaging and DU performed within 1 month of each other. There were 36 patients who had cross-sectional imaging positive for iliac vein thrombosis; only 10 (27.8%) of these patients were found to have iliac vein thrombosis by DU. Thus, 26 patients (72.2%) were not diagnosed accurately by DU. On the basis of our data, the sensitivity and positive predictive value of DU compared with cross-sectional imaging in diagnosis of iliac vein thrombosis were 27.8% and 76.9%, respectively. We did not identify any patient-specific factors that influenced the discrepancy between DU and cross-sectional imaging. Conclusions: Our current protocol of lower extremity venous DU is not an effective tool in diagnosis of iliac vein thrombosis. All patients with clinically suspected iliac vein thrombosis should be evaluated with specific pelvic ultrasound protocols or cross-sectional imaging. (J Vasc Surg: Venous and Lym Dis 2018;6:724-9.) Keywords: Iliac vein; Thrombosis; Diagnosis; Veins; Venous thrombosis
Deep venous thrombosis isolated to the common or external iliac veins is generally reported as an uncommon finding, representing anywhere from 1% to 4% of diagnosed deep venous thromboses.1 Yet, the morbidity associated with iliac vein thrombosis is significant; patients are at increased risk for development of pulmonary emboli and post-thrombotic syndrome, especially when they are treated with anticoagulation alone.2 Thus, accurate methods of diagnosis are important for treating iliac vein thrombosis effectively.
From the Division of Vascular Surgery,a Division of Interventional Radiology,b and Vascular Laboratory,c Northwestern Memorial Hospital. Research reported in this publication was supported, in part, by the National Institutes of Health’s National Center for Advancing Translational Sciences (Grant No. UL1TR001422) and the Northwestern Medicine Enterprise Data Warehouse. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Author conflict of interest: none. Presented in the plenary session of the Sixteenth Annual Meeting of the Society of Spanish Speaking Vascular Surgeons, Madrid, Spain, October 9-11, 2017. Correspondence: Heron E. Rodriguez, MD, Division of Vascular Surgery, Department of Surgery, Northwestern University, 676 N St Clair, Ste 650, Chicago, IL
Cross-sectional imaging studies, such as magnetic resonance (MR) imaging and computed tomography (CT), as well as invasive techniques like traditional venography are potentially more accurate imaging modalities for diagnosis of iliac vein thrombosis as they are not susceptible to bowel gas obscuring pelvic venous anatomy or the operator’s vulnerability. Whereas they are effective in making the diagnosis, these modalities are more expensive and involve risk to the patient, including administration of contrast material and ionizing radiation. Venous duplex ultrasound (DU) has been widely accepted as the method of choice for screening patients with acute deep venous thrombosis in the extremities. DU is associated with high sensitivity and specificity for detecting lower extremity deep venous thrombosis below the inguinal ligament. However, there are few data regarding the efficacy of standard, lower extremity venous DU in diagnosis of iliac vein thrombosis.3 Therefore, the purpose of this investigation was to evaluate the accuracy and efficacy of our current protocol of lower extremity venous DU in diagnosis of iliac vein thrombosis compared with cross-sectional CT and MR imaging techniques.
60611 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the Journal policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 2213-333X Published by Elsevier Inc. on behalf of the Society for Vascular Surgery. https://doi.org/10.1016/j.jvsv.2018.06.010
724
METHODS This study received review and approval from the Northwestern University Institutional Review Board; because data were deidentified, informed consent was waived. Data for this project were accessed through the
Jain et al
Journal of Vascular Surgery: Venous and Lymphatic Disorders
725
Volume 6, Number 6
Northwestern Medicine Enterprise Database Warehouse (NMEDW), an integrated repository of clinical and research data sources across the Northwestern University campus. The database was queried for all patients diagnosed with iliac vein thrombosis between January 1, 2006, and December 31, 2015. For the patients identified from the initial search, we then queried the database for various patient-related variables, including the patients’ demographics, relevant laboratory values, and relevant medical history. All queries were performed using International Classification of Diseases, Ninth Revision and Tenth Revision codes (Table I). We reviewed the electronic medical record (Cerner PowerChart; Cerner Corporation, Kansas City, Mo) at Northwestern Memorial Hospital for each of the patients identified in our NMEDW search. To obtain a valid comparison, patients included in our study needed to have both DU and cross-sectional imaging performed within a month of each other. The reports of both DU and cross-sectional imaging were reviewed, and the diagnosis of iliac vein thrombosis was recorded when present in the interpretation. Our current lower extremity venous DU examination is performed according to the International Accreditation Commission (IAC) Standards and Guidelines for Vascular Testing Accreditation.4 It is performed with the patient supine on a bed or a cart with the index leg externally rotated with the knee slightly bent. Both legs are examined with high-resolution DU systems. Transducers are selected on the basis of vessel depth (usually L9-3 or C5-1 MHz). Compression maneuvers are performed and
Table I. International Classification of Diseases, Ninth Revision (ICD-9) and Tenth Revision (ICD-10) codes used to query categorical variables in the Northwestern Medicine Enterprise Database Warehouse (NMEDW) Diagnosis
ICD-9 codes
ICD-10 codes
Iliac vein thrombosis
453.41, 453.51
I82.42, I82.52
585
N18
CKD Heart failure
428
I50
COPD
491, 492, 493, 494, 495, 496
J44
DVT or PE
V12.51, V12.55
I82.49, I82.59, I26, Z86.71
454, 459, 707.1
I83, I87.2, I87.3
VI Malignant disease
V10
Z85
PVD
443
I73.9
CVA
V12.54
I63, Z86.74
457.1
I89
250
E8, E10, E11, E13
V15.82
Z72.0
Lymphedema DM Smoking history
CKD, Chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; DM, diabetes mellitus; DVT, deep venous thrombosis; PE, pulmonary embolism; PVD, peripheral vascular disease; VI, venous insufficiency.
ARTICLE HIGHLIGHTS d
d
d
Type of Research: Retrospective, single-center cohort study Take Home Message: In 36 patients with iliac vein thrombosis found by cross-sectional imaging, only 10 (27.8%) were confirmed by duplex ultrasound (DU), providing a sensitivity of 27.8% and a positive predictive value of 76.9% for DU with the current protocol for detecting iliac vein thrombosis. Recommendation: We recommend that patients with suspected iliac vein thrombosis have a dedicated pelvic DU examination or cross-sectional imaging.
documented in a dual screen in gray-scale B mode at the common femoral, femoral, popliteal, and infrapopliteal vessels. The lumen of the vessel is evaluated for the presence of echogenic material. Color DU is used to assess the patency of the vessel. Doppler is then used to assess the signal within the lumen for spontaneity, phasicity, and augmentation. Not every patient undergoes examination of the pelvic vessels. If it is specifically requested or if an abnormality in the flow of the femoral veins is detected, pelvic vessel interrogation is performed. For this we use a curved array C5-1 MHz transducer. Gray scale and color DU are used to look for soft or bright echoes within the vessel lumen. Any extrinsic compression is noted. Compression is attempted but often is impossible because the vessel walls cannot be visualized throughout the compression maneuver. The images are reviewed and interpreted by a physician registered in vascular interpretation. A variety of cross-sectional studies were included in our analysis. CT venograms were obtained using multidetector scanners with delayed axial image acquisition 130 seconds after administration of an iodinated contrast agent (iohexol [Omnipaque; GE Healthcare, Chicago, Ill] and iopamidol [Isovue; Bracco, Monroe Township, NJ]); multiplanar reformatted images were routinely generated from the axial source images. Multiplanar MR venograms were obtained with a variety of pulse sequences before the administration of contrast material; fat-saturated T1 images were obtained 130 seconds after the administration of the intravenous contrast agent (gadopentetate dimeglumine [Magnevist] and gadobutrol [Gadavist]; Bayer, Whippany, NJ). All studies were interpreted by board-certified radiologists. The primary objective of our study was to determine how well DU results correlated with findings on crosssectional imaging. As an adjunct investigation, we performed univariate analysis of various patient-specific variables to determine whether any factors, such as obesity and lymphedema, were more likely to be associated with positive vs negative correlation between DU
726
Jain et al
Journal of Vascular Surgery: Venous and Lymphatic Disorders November 2018
and cross-sectional imaging. Continuous variables were summarized by means, standard errors, medians, and interquartile ranges; these were compared between groups by the Wilcoxon rank sum test. Categorical variables were summarized by frequency counts and percentages, which were compared between groups by the Fisher exact test.
RESULTS The initial NMEDW query yielded 80 patients with a diagnosis of iliac vein thrombosis in the medical record. After review of medical records, 32 patients were excluded because they did not have either crosssectional imaging or DU examination. The final study cohort consisted of 48 patients with a diagnosis of iliac vein thrombosis in the medical record; for all of these patients, cross-sectional imaging and DU examination were performed within a month of each other. In our study cohort, cross-sectional imaging demonstrated iliac vein thrombosis in 36 patients and no evidence of thrombosis in the remaining 12. In the group of patients who were found to have iliac vein thrombosis on cross-sectional imaging, only 10 patients (27.8%) had DU results that were consistent with iliac vein thrombosis. Thus, DU did not detect iliac vein thrombosis in 26 patients (72.2%) diagnosed with iliac vein thrombosis on the basis of cross-sectional imaging. In the remaining 12 patients with no evidence of iliac vein thrombosis by cross-sectional imaging, DU identified iliac vein thrombosis in three patients; the remaining nine patients had both DU and cross-sectional imaging findings negative for iliac vein thrombosis. On the basis of these results, the sensitivity of DU in diagnosis of iliac vein thrombosis is 27.8%, and the positive predictive value of the test is 76.9%. The study was not powered to interpret specificity and negative predictive value (Table II). In an adjunct analysis, we compared continuous and categorical patient-specific variables between patients who had both DU and cross-sectional imaging findings positive for iliac vein thrombosis vs those who had negative DU results in the setting of a positive cross-sectional Table II. Comparison of patients with iliac vein thrombosis diagnosed by cross-sectional imaging or venous duplex ultrasound (DU): Results and statistical significance Cross-sectional imaging Venous DU
Positive
Negative
Sensitivity
27.8%
Positive
10
3
Specificity
N/A
Negative
26
9a
PPV
76.9%
n ¼ 36
n ¼ 12
NPV
N/A
N/A, Not applicable; NPV, negative predictive value; PPV, positive predictive value. a This number does not represent the true negative because the study was designed to evaluate only patients with a diagnosis of iliac vein thrombosis; thus, specificity and negative predictive value cannot be determined.
imaging result. We found no significant differences to help explain the poor sensitivity of DU in comparison with cross-sectional imaging (Table III).
DISCUSSION The purpose of our investigation was to evaluate the efficacy of our current lower extremity venous DU protocol in the diagnosis of iliac vein thrombosis compared with cross-sectional imaging. Our data indicate that this protocol is not reliable in evaluating patients for iliac vein thrombosis. With a sensitivity of only 27.8% and a positive predictive value of 76.9%, DU is far from perfect with regard to diagnosis of iliac vein thrombosis. Many studies have previously evaluated the efficacy of DU in diagnosis of deep venous thrombosis. Sensitivity
Table III. Comparison of variables between patients who had venous duplex ultrasound (DU) results positive for iliac vein thrombus and those with negative venous DU results for iliac vein thrombus in the setting of a known iliac vein thrombus diagnosed with cross-sectional imaging
Continuous variables Weight, pounds
DU results positive, mean (n ¼ 10)
DU results negative, mean (n ¼ 26)
Wilcoxon rank sum, P value
166.8
171.8
.67
BMI, kg/m2
27.4
27.3
.99
BUN, mg/dL
16.7
15.5
Creatinine, mg/dL Fibrinogen, mg/dL Hemoglobin, g/dL INR Platelets, 103/mL
1.05 272.5
1.16 150.1
11.75
11.58
1.29 333.6
1.53 250.2
.96 .55 .19 .78 .76 .10
PT, seconds
13.8
15.4
.94
PTT, seconds
36.1
22.9
.07
Categorical variables
DU results positive (n ¼ 10)
DU results negative (n ¼ 26)
Fisher exact, P value
CKD
0
2
.99
Heart failure
0
0
N/A
COPD
1
2
.99
DVT or PE
1
9
.23
VI
1
4
.99
Malignant disease
5
6
.22
PVD
0
1
.99
CVA
0
1
.99
1
1
.48
DM
0
4
.56
Smoking history
3
3
.32
Lymphedema
BMI, Body mass index; BUN, blood urea nitrogen; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; DM, diabetes mellitus; DVT, deep venous thrombosis; INR, international normalized ratio; N/A, not applicable; PE, pulmonary embolism; PT, prothrombin time; PTT, partial thromboplastin time; PVD, peripheral vascular disease; VI, venous insufficiency.
Journal of Vascular Surgery: Venous and Lymphatic Disorders
Jain et al
727
Volume 6, Number 6
and specificity of ultrasound have been quoted as high as 100% and 97%, respectively, in the femoral and popliteal vessels5; however, literature specifically directed toward the diagnosis of iliac vein thrombosis is limited. Early studies compared DU with contrast venography and found that DU was significantly less accurate in the diagnosis of proximal leg vein thrombosis.6 As technology has improved, CT and MR venography have become more accessible modalities for diagnosis, given their less invasive nature compared with conventional venography. MR venography is resource intensive, and its quality also depends on local expertise. Still, studies have demonstrated that CT and MR venography are more accurate than DU in the diagnosis of proximal vein thrombosis.7-10 Lower extremity venous DU has become one of the most widely used medical tests in both the inpatient and outpatient setting. The superficial position of the veins in the leg below the inguinal ligament allows the performance of compression maneuvers with little difficulty. The inability to compress the common femoral vein or popliteal vein in patients with symptoms has a positive predictive value of 97% for a deep venous thrombosis. Similarly, full compressibility of these veins despite symptoms has a negative predictive value of 98%.11 This is particularly true in the hands of a skilled ultrasound technician.12 The same modality is much less reliable when patients are asymptomatic.13
The location of the iliac veins in the pelvis makes ultrasound examination more difficult. The presence of gas in the abdomen and the inability to perform compressive maneuvers in most cases are additional factors compromising the adequate evaluation of the iliac vessels. The operator’s skills and experience play a much more relevant role than in the case of infrainguinal vein examination. In our institution and likely in most vascular laboratories in the country, insonation of the iliac veins is attempted only when the technicians are specifically prompted to do so or when indirect signs of proximal obstruction are seen while the femoral vessels are being examined. Monophasic, continuous waveforms in the common femoral vein, implying loss of spontaneous flow, loss of phasicity, and lack of augmentation, are generally considered indicative of proximal venous disease.14,15 This may include external compression, hypoplastic venous anatomy, thrombosis, and a multitude of other causes. Studies have quoted the prevalence of iliac vein thrombosis in patients with monophasic common femoral waveforms ranging from 22% to 38%.16,17 Yet, in our analysis, even in cases in which DU demonstrated compressible femoral veins with respiratory variation and maintained phasicity, cross-sectional imaging demonstrated iliac vein thrombosis (Fig). Furthermore, waveform disturbance in the common femoral vein is no longer reliable when thrombus is seen within the
Fig. Venous duplex ultrasound (DU) and magnetic resonance imaging (MRI) in a 45-year-old woman with May-Thurner syndrome. A, The left common femoral vein (arrow) is easily compressible with the ultrasound probe, demonstrating no evidence of thrombosis. B, Normal respiratory variation in the common femoral vein. C, Bilateral common femoral veins (arrows) are clearly visible on MRI. D, Only the right external iliac vein (arrow) is visible on MRI; the left external iliac vein is occluded.
728
Jain et al
Journal of Vascular Surgery: Venous and Lymphatic Disorders November 2018
common femoral vein, thereby making a diagnosis of iliac vein thrombosis more difficult. For all those reasons, we think it likely that most vascular laboratories in the country are not reliably diagnosing iliac vein thrombosis during routine lower extremity venous DU. In our institution, our IAC-certified vascular laboratory performs approximately 10,000 lower extremity venous studies every year. Most of these studies are performed without fasting, with variable amounts of gas in the pelvis. Many are done at the bedside or in the emergency department. As in most vascular laboratories in the country, most routine lower extremity venous DU examinations do not achieve adequate insonation of the iliac veins. Furthermore, the 2018 IAC Standards and Guidelines for Vascular Testing Accreditation do not include the iliac veins in the list of required vessels to be imaged by protocol during lower extremity venous DU.4 Our study has limitations. First of all, our sample size of patients was very small, despite gathering of data during a 10-year period. The accuracy of statistical probabilities (ie, sensitivity) and ratios (ie, positive predictive value) is dependent on an accurate reflection of disease prevalence in the data set. Without a larger sample size of patients, the true prevalence of iliac vein thrombosis within the population is misrepresented. This limits our ability to make conclusive statements about the efficacy of DU in the general population. Second, this study was designed as a retrospective analysis of patients with a diagnosis of iliac vein thrombosis in their medical records. Thus, although sensitivity and positive predictive value can be determined on the basis of our results, we cannot comment on the specificity or negative predictive value of DU in diagnosis of iliac vein thrombosis. Nine patients with documented diagnosis of iliac vein thrombosis had negative results of imaging studies (both DU and cross-sectional imaging). Potential explanations for this include resolution of a previously diagnosed thrombus or simply an error in medical record documentation. To fully evaluate the accuracy of DU, it would be necessary to perform a prospective trial evaluating all patients with suspected iliac vein thrombosis with cross-sectional imaging and a specific pelvic venous protocol for DU; this would be the only way to accurately reflect the number of patients with both DU and crosssectional imaging findings negative for iliac vein thrombosis (true negative). Finally, DU is dependent on the experience and skill of the technician performing the study. A 10-year study period, even at the same institution, implies a lot of variability in the quality of the ultrasound technicians performing the studies. It is difficult to control for such variability, and it can certainly influence our results.
CONCLUSIONS This investigation raises the concern that current protocols of lower extremity venous DU are not efficient in
diagnosis of iliac vein thrombosis. Despite decades of experience with ultrasound and improvements in technical considerations and skills with this diagnostic modality, the poor sensitivity and subpar positive predictive value of our current protocol for lower extremity venous DU make it an unreliable tool for detecting iliac vein thrombosis. Patients with suspected iliac vein thrombosis should have a dedicated pelvic DU examination or cross-sectional imaging performed regardless of the findings seen on routine infrainguinal venous DU. We would like to acknowledge Irene B. Helenowski, PhD, of the Northwestern University Feinberg School of Medicine’s Department of Preventive Medicine for her contribution to the statistical analysis required in this study.
AUTHOR CONTRIBUTIONS Conception and design: AJ, SR, HR Analysis and interpretation: AJ, MS, KD, PA, ME, HR Data collection: AJ Writing the article: AJ, HR Critical revision of the article: AJ, MS, SR, KD, PA, ME, HR Final approval of the article: AJ, MS, SR, KD, PA, ME, HR Statistical analysis: AJ, HR Obtained funding: Not applicable Overall responsibility: HR
REFERENCES 1. Spritzer CE, Arata MA, Freed KS. Isolated pelvic deep venous thrombosis: relative frequency as detected with MR imaging. Radiology 2001;19:521-5. 2. Comerota AJ, Kamath V. Thrombolysis for iliofemoral deep venous thrombosis. Expert Rev Cardiovasc Ther 2013;11: 1631-8. 3. Dorfman GS, Cronan JJ. Venous ultrasonography. Radiol Clin North Am 1992;30:879-94. 4. Intersocietal Accreditation Commission. IAC standards and guidelines for vascular testing accreditation. Available at: https://www.intersocietal.org/vascular/seeking/vascular_ standards.htm. Accessed January 15, 2018. 5. Fletcher JP, Kershaw LZ, Barker DS, Koutts J, Varnava A. Ultrasound diagnosis of lower limb deep venous thrombosis. Med J Aust 1990;153:453-5. 6. Davidson BL, Elliott CG, Lensing AW. Low accuracy of color Doppler ultrasound in the detection of proximal leg vein thrombosis in asymptomatic high-risk patients. The RD Heparin Arthroplasty Group. Ann Intern Med 1992;117:735-8. 7. Laissy JP, Cinqualbre A, Loshkajian A, Henry-Feugeas MC, Crestani B, Riquelme C, et al. Assessment of deep venous thrombosis in the lower limbs and pelvis: MR venography versus duplex Doppler sonography. AJR Am J Roentgenol 1996;167:971-5. 8. Dronkers CE, Sramek A, Huisman MV, Klok FA. Accurate diagnosis of iliac vein thrombosis in pregnancy with magnetic resonance direct thrombus imaging (MRDTI). BMJ Case Rep 2016;2016. 9. Tamura K, Nakahara H. MR venography for the assessment of deep vein thrombosis in lower extremities with varicose veins. Ann Vasc Dis 2014;7:399-403.
Jain et al
Journal of Vascular Surgery: Venous and Lymphatic Disorders
729
Volume 6, Number 6 10. Lim KE, Hsu WC, Hsu YY, Chu PH, Ng CJ. Deep venous thrombosis: comparison of indirect multidetector CT venography and sonography of lower extremities in 26 patients. Clin Imaging 2004;28:439-44. 11. Kearon C, Ginsberg JS, Hirsh J. The role of venous ultrasonography in the diagnosis of suspected deep venous thrombosis and pulmonary embolism. Ann Intern Med 1998;129:1044-9. 12. Gupta R, Stouffer GA. Deep venous thrombosis: a review of the pathophysiology, clinical features, and diagnostic modalities. Am J Med Sci 2001;322:358-64. 13. Borris LC, Christiansen HM, Lassen MR, Olsen AD, Schott P. Comparison of real-time B-mode ultrasonography and bilateral ascending phlebography for detection of postoperative deep vein thrombosis following elective hip surgery. The Venous Thrombosis Group. Thromb Haemost 1989;61:363-5.
14. Shin SY, Kim H, Choi YS, Kim SW. Usefulness of Doppler waveform analysis before performing a complex procedure using femoral venous access. J Clin Ultrasound 2018;46:157-9. 15. Metzge PB, Rossi FH, Kambara AM, Izukawa NM, Saleh MH, Pinto IM, et al. Criteria for detecting significant chronic iliac venous obstructions with duplex ultrasound. J Vasc Surg Venous Lymphat Disord 2016;4:18-27. 16. Lin EP, Bhatt S, Rubens D, Dogra VS. The importance of monophasic Doppler waveforms in the common femoral vein: a retrospective study. J Ultrasound Med 2007;26:885-91. 17. Bach AM, Hann LE. When the common femoral vein is revealed as flattened on spectral Doppler sonography: is it a reliable sign for diagnosis of proximal venous obstruction? AJR Am J Roentgenol 1997;168:733-6.
Submitted Feb 20, 2018; accepted Jun 1, 2018.
ABSTRACT Background: Deep venous thrombosis isolated to the iliac veins is uncommon. Venous duplex ultrasound (DU) is widely accepted as the screening modality of choice for lower extremity deep venous thromboses. This investigation evaluated the accuracy and efficacy of DU in diagnosis of iliac vein thrombosis. Methods: We conducted a single-center retrospective review of patients who were diagnosed with iliac vein thrombosis between January 1, 2006, and December 31, 2015. Patients included in our analysis needed to have both DU and crosssectional imaging performed within a month of each other. The efficacy of DU in diagnosis of iliac vein thrombosis was determined using cross-sectional imaging as a standard for diagnosis. Results: In total, our query yielded 80 patients with a diagnosis of iliac vein thrombosis in the medical chart; 48 patients had both cross-sectional imaging and DU performed within 1 month of each other. There were 36 patients who had cross-sectional imaging positive for iliac vein thrombosis; only 10 (27.8%) of these patients were found to have iliac vein thrombosis by DU. Thus, 26 patients (72.2%) were not diagnosed accurately by DU. On the basis of our data, the sensitivity and positive predictive value of DU compared with cross-sectional imaging in diagnosis of iliac vein thrombosis were 27.8% and 76.9%, respectively. We did not identify any patient-specific factors that influenced the discrepancy between DU and cross-sectional imaging. Conclusions: Our current protocol of lower extremity venous DU is not an effective tool in diagnosis of iliac vein thrombosis. All patients with clinically suspected iliac vein thrombosis should be evaluated with specific pelvic ultrasound protocols or cross-sectional imaging. (J Vasc Surg: Venous and Lym Dis 2018;6:724-9.) Keywords: Iliac vein; Thrombosis; Diagnosis; Veins; Venous thrombosis
Deep venous thrombosis isolated to the common or external iliac veins is generally reported as an uncommon finding, representing anywhere from 1% to 4% of diagnosed deep venous thromboses.1 Yet, the morbidity associated with iliac vein thrombosis is significant; patients are at increased risk for development of pulmonary emboli and post-thrombotic syndrome, especially when they are treated with anticoagulation alone.2 Thus, accurate methods of diagnosis are important for treating iliac vein thrombosis effectively.
From the Division of Vascular Surgery,a Division of Interventional Radiology,b and Vascular Laboratory,c Northwestern Memorial Hospital. Research reported in this publication was supported, in part, by the National Institutes of Health’s National Center for Advancing Translational Sciences (Grant No. UL1TR001422) and the Northwestern Medicine Enterprise Data Warehouse. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Author conflict of interest: none. Presented in the plenary session of the Sixteenth Annual Meeting of the Society of Spanish Speaking Vascular Surgeons, Madrid, Spain, October 9-11, 2017. Correspondence: Heron E. Rodriguez, MD, Division of Vascular Surgery, Department of Surgery, Northwestern University, 676 N St Clair, Ste 650, Chicago, IL
Cross-sectional imaging studies, such as magnetic resonance (MR) imaging and computed tomography (CT), as well as invasive techniques like traditional venography are potentially more accurate imaging modalities for diagnosis of iliac vein thrombosis as they are not susceptible to bowel gas obscuring pelvic venous anatomy or the operator’s vulnerability. Whereas they are effective in making the diagnosis, these modalities are more expensive and involve risk to the patient, including administration of contrast material and ionizing radiation. Venous duplex ultrasound (DU) has been widely accepted as the method of choice for screening patients with acute deep venous thrombosis in the extremities. DU is associated with high sensitivity and specificity for detecting lower extremity deep venous thrombosis below the inguinal ligament. However, there are few data regarding the efficacy of standard, lower extremity venous DU in diagnosis of iliac vein thrombosis.3 Therefore, the purpose of this investigation was to evaluate the accuracy and efficacy of our current protocol of lower extremity venous DU in diagnosis of iliac vein thrombosis compared with cross-sectional CT and MR imaging techniques.
60611 (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the Journal policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 2213-333X Published by Elsevier Inc. on behalf of the Society for Vascular Surgery. https://doi.org/10.1016/j.jvsv.2018.06.010
724
METHODS This study received review and approval from the Northwestern University Institutional Review Board; because data were deidentified, informed consent was waived. Data for this project were accessed through the
Jain et al
Journal of Vascular Surgery: Venous and Lymphatic Disorders
725
Volume 6, Number 6
Northwestern Medicine Enterprise Database Warehouse (NMEDW), an integrated repository of clinical and research data sources across the Northwestern University campus. The database was queried for all patients diagnosed with iliac vein thrombosis between January 1, 2006, and December 31, 2015. For the patients identified from the initial search, we then queried the database for various patient-related variables, including the patients’ demographics, relevant laboratory values, and relevant medical history. All queries were performed using International Classification of Diseases, Ninth Revision and Tenth Revision codes (Table I). We reviewed the electronic medical record (Cerner PowerChart; Cerner Corporation, Kansas City, Mo) at Northwestern Memorial Hospital for each of the patients identified in our NMEDW search. To obtain a valid comparison, patients included in our study needed to have both DU and cross-sectional imaging performed within a month of each other. The reports of both DU and cross-sectional imaging were reviewed, and the diagnosis of iliac vein thrombosis was recorded when present in the interpretation. Our current lower extremity venous DU examination is performed according to the International Accreditation Commission (IAC) Standards and Guidelines for Vascular Testing Accreditation.4 It is performed with the patient supine on a bed or a cart with the index leg externally rotated with the knee slightly bent. Both legs are examined with high-resolution DU systems. Transducers are selected on the basis of vessel depth (usually L9-3 or C5-1 MHz). Compression maneuvers are performed and
Table I. International Classification of Diseases, Ninth Revision (ICD-9) and Tenth Revision (ICD-10) codes used to query categorical variables in the Northwestern Medicine Enterprise Database Warehouse (NMEDW) Diagnosis
ICD-9 codes
ICD-10 codes
Iliac vein thrombosis
453.41, 453.51
I82.42, I82.52
585
N18
CKD Heart failure
428
I50
COPD
491, 492, 493, 494, 495, 496
J44
DVT or PE
V12.51, V12.55
I82.49, I82.59, I26, Z86.71
454, 459, 707.1
I83, I87.2, I87.3
VI Malignant disease
V10
Z85
PVD
443
I73.9
CVA
V12.54
I63, Z86.74
457.1
I89
250
E8, E10, E11, E13
V15.82
Z72.0
Lymphedema DM Smoking history
CKD, Chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; DM, diabetes mellitus; DVT, deep venous thrombosis; PE, pulmonary embolism; PVD, peripheral vascular disease; VI, venous insufficiency.
ARTICLE HIGHLIGHTS d
d
d
Type of Research: Retrospective, single-center cohort study Take Home Message: In 36 patients with iliac vein thrombosis found by cross-sectional imaging, only 10 (27.8%) were confirmed by duplex ultrasound (DU), providing a sensitivity of 27.8% and a positive predictive value of 76.9% for DU with the current protocol for detecting iliac vein thrombosis. Recommendation: We recommend that patients with suspected iliac vein thrombosis have a dedicated pelvic DU examination or cross-sectional imaging.
documented in a dual screen in gray-scale B mode at the common femoral, femoral, popliteal, and infrapopliteal vessels. The lumen of the vessel is evaluated for the presence of echogenic material. Color DU is used to assess the patency of the vessel. Doppler is then used to assess the signal within the lumen for spontaneity, phasicity, and augmentation. Not every patient undergoes examination of the pelvic vessels. If it is specifically requested or if an abnormality in the flow of the femoral veins is detected, pelvic vessel interrogation is performed. For this we use a curved array C5-1 MHz transducer. Gray scale and color DU are used to look for soft or bright echoes within the vessel lumen. Any extrinsic compression is noted. Compression is attempted but often is impossible because the vessel walls cannot be visualized throughout the compression maneuver. The images are reviewed and interpreted by a physician registered in vascular interpretation. A variety of cross-sectional studies were included in our analysis. CT venograms were obtained using multidetector scanners with delayed axial image acquisition 130 seconds after administration of an iodinated contrast agent (iohexol [Omnipaque; GE Healthcare, Chicago, Ill] and iopamidol [Isovue; Bracco, Monroe Township, NJ]); multiplanar reformatted images were routinely generated from the axial source images. Multiplanar MR venograms were obtained with a variety of pulse sequences before the administration of contrast material; fat-saturated T1 images were obtained 130 seconds after the administration of the intravenous contrast agent (gadopentetate dimeglumine [Magnevist] and gadobutrol [Gadavist]; Bayer, Whippany, NJ). All studies were interpreted by board-certified radiologists. The primary objective of our study was to determine how well DU results correlated with findings on crosssectional imaging. As an adjunct investigation, we performed univariate analysis of various patient-specific variables to determine whether any factors, such as obesity and lymphedema, were more likely to be associated with positive vs negative correlation between DU
726
Jain et al
Journal of Vascular Surgery: Venous and Lymphatic Disorders November 2018
and cross-sectional imaging. Continuous variables were summarized by means, standard errors, medians, and interquartile ranges; these were compared between groups by the Wilcoxon rank sum test. Categorical variables were summarized by frequency counts and percentages, which were compared between groups by the Fisher exact test.
RESULTS The initial NMEDW query yielded 80 patients with a diagnosis of iliac vein thrombosis in the medical record. After review of medical records, 32 patients were excluded because they did not have either crosssectional imaging or DU examination. The final study cohort consisted of 48 patients with a diagnosis of iliac vein thrombosis in the medical record; for all of these patients, cross-sectional imaging and DU examination were performed within a month of each other. In our study cohort, cross-sectional imaging demonstrated iliac vein thrombosis in 36 patients and no evidence of thrombosis in the remaining 12. In the group of patients who were found to have iliac vein thrombosis on cross-sectional imaging, only 10 patients (27.8%) had DU results that were consistent with iliac vein thrombosis. Thus, DU did not detect iliac vein thrombosis in 26 patients (72.2%) diagnosed with iliac vein thrombosis on the basis of cross-sectional imaging. In the remaining 12 patients with no evidence of iliac vein thrombosis by cross-sectional imaging, DU identified iliac vein thrombosis in three patients; the remaining nine patients had both DU and cross-sectional imaging findings negative for iliac vein thrombosis. On the basis of these results, the sensitivity of DU in diagnosis of iliac vein thrombosis is 27.8%, and the positive predictive value of the test is 76.9%. The study was not powered to interpret specificity and negative predictive value (Table II). In an adjunct analysis, we compared continuous and categorical patient-specific variables between patients who had both DU and cross-sectional imaging findings positive for iliac vein thrombosis vs those who had negative DU results in the setting of a positive cross-sectional Table II. Comparison of patients with iliac vein thrombosis diagnosed by cross-sectional imaging or venous duplex ultrasound (DU): Results and statistical significance Cross-sectional imaging Venous DU
Positive
Negative
Sensitivity
27.8%
Positive
10
3
Specificity
N/A
Negative
26
9a
PPV
76.9%
n ¼ 36
n ¼ 12
NPV
N/A
N/A, Not applicable; NPV, negative predictive value; PPV, positive predictive value. a This number does not represent the true negative because the study was designed to evaluate only patients with a diagnosis of iliac vein thrombosis; thus, specificity and negative predictive value cannot be determined.
imaging result. We found no significant differences to help explain the poor sensitivity of DU in comparison with cross-sectional imaging (Table III).
DISCUSSION The purpose of our investigation was to evaluate the efficacy of our current lower extremity venous DU protocol in the diagnosis of iliac vein thrombosis compared with cross-sectional imaging. Our data indicate that this protocol is not reliable in evaluating patients for iliac vein thrombosis. With a sensitivity of only 27.8% and a positive predictive value of 76.9%, DU is far from perfect with regard to diagnosis of iliac vein thrombosis. Many studies have previously evaluated the efficacy of DU in diagnosis of deep venous thrombosis. Sensitivity
Table III. Comparison of variables between patients who had venous duplex ultrasound (DU) results positive for iliac vein thrombus and those with negative venous DU results for iliac vein thrombus in the setting of a known iliac vein thrombus diagnosed with cross-sectional imaging
Continuous variables Weight, pounds
DU results positive, mean (n ¼ 10)
DU results negative, mean (n ¼ 26)
Wilcoxon rank sum, P value
166.8
171.8
.67
BMI, kg/m2
27.4
27.3
.99
BUN, mg/dL
16.7
15.5
Creatinine, mg/dL Fibrinogen, mg/dL Hemoglobin, g/dL INR Platelets, 103/mL
1.05 272.5
1.16 150.1
11.75
11.58
1.29 333.6
1.53 250.2
.96 .55 .19 .78 .76 .10
PT, seconds
13.8
15.4
.94
PTT, seconds
36.1
22.9
.07
Categorical variables
DU results positive (n ¼ 10)
DU results negative (n ¼ 26)
Fisher exact, P value
CKD
0
2
.99
Heart failure
0
0
N/A
COPD
1
2
.99
DVT or PE
1
9
.23
VI
1
4
.99
Malignant disease
5
6
.22
PVD
0
1
.99
CVA
0
1
.99
1
1
.48
DM
0
4
.56
Smoking history
3
3
.32
Lymphedema
BMI, Body mass index; BUN, blood urea nitrogen; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVA, cerebrovascular accident; DM, diabetes mellitus; DVT, deep venous thrombosis; INR, international normalized ratio; N/A, not applicable; PE, pulmonary embolism; PT, prothrombin time; PTT, partial thromboplastin time; PVD, peripheral vascular disease; VI, venous insufficiency.
Journal of Vascular Surgery: Venous and Lymphatic Disorders
Jain et al
727
Volume 6, Number 6
and specificity of ultrasound have been quoted as high as 100% and 97%, respectively, in the femoral and popliteal vessels5; however, literature specifically directed toward the diagnosis of iliac vein thrombosis is limited. Early studies compared DU with contrast venography and found that DU was significantly less accurate in the diagnosis of proximal leg vein thrombosis.6 As technology has improved, CT and MR venography have become more accessible modalities for diagnosis, given their less invasive nature compared with conventional venography. MR venography is resource intensive, and its quality also depends on local expertise. Still, studies have demonstrated that CT and MR venography are more accurate than DU in the diagnosis of proximal vein thrombosis.7-10 Lower extremity venous DU has become one of the most widely used medical tests in both the inpatient and outpatient setting. The superficial position of the veins in the leg below the inguinal ligament allows the performance of compression maneuvers with little difficulty. The inability to compress the common femoral vein or popliteal vein in patients with symptoms has a positive predictive value of 97% for a deep venous thrombosis. Similarly, full compressibility of these veins despite symptoms has a negative predictive value of 98%.11 This is particularly true in the hands of a skilled ultrasound technician.12 The same modality is much less reliable when patients are asymptomatic.13
The location of the iliac veins in the pelvis makes ultrasound examination more difficult. The presence of gas in the abdomen and the inability to perform compressive maneuvers in most cases are additional factors compromising the adequate evaluation of the iliac vessels. The operator’s skills and experience play a much more relevant role than in the case of infrainguinal vein examination. In our institution and likely in most vascular laboratories in the country, insonation of the iliac veins is attempted only when the technicians are specifically prompted to do so or when indirect signs of proximal obstruction are seen while the femoral vessels are being examined. Monophasic, continuous waveforms in the common femoral vein, implying loss of spontaneous flow, loss of phasicity, and lack of augmentation, are generally considered indicative of proximal venous disease.14,15 This may include external compression, hypoplastic venous anatomy, thrombosis, and a multitude of other causes. Studies have quoted the prevalence of iliac vein thrombosis in patients with monophasic common femoral waveforms ranging from 22% to 38%.16,17 Yet, in our analysis, even in cases in which DU demonstrated compressible femoral veins with respiratory variation and maintained phasicity, cross-sectional imaging demonstrated iliac vein thrombosis (Fig). Furthermore, waveform disturbance in the common femoral vein is no longer reliable when thrombus is seen within the
Fig. Venous duplex ultrasound (DU) and magnetic resonance imaging (MRI) in a 45-year-old woman with May-Thurner syndrome. A, The left common femoral vein (arrow) is easily compressible with the ultrasound probe, demonstrating no evidence of thrombosis. B, Normal respiratory variation in the common femoral vein. C, Bilateral common femoral veins (arrows) are clearly visible on MRI. D, Only the right external iliac vein (arrow) is visible on MRI; the left external iliac vein is occluded.
728
Jain et al
Journal of Vascular Surgery: Venous and Lymphatic Disorders November 2018
common femoral vein, thereby making a diagnosis of iliac vein thrombosis more difficult. For all those reasons, we think it likely that most vascular laboratories in the country are not reliably diagnosing iliac vein thrombosis during routine lower extremity venous DU. In our institution, our IAC-certified vascular laboratory performs approximately 10,000 lower extremity venous studies every year. Most of these studies are performed without fasting, with variable amounts of gas in the pelvis. Many are done at the bedside or in the emergency department. As in most vascular laboratories in the country, most routine lower extremity venous DU examinations do not achieve adequate insonation of the iliac veins. Furthermore, the 2018 IAC Standards and Guidelines for Vascular Testing Accreditation do not include the iliac veins in the list of required vessels to be imaged by protocol during lower extremity venous DU.4 Our study has limitations. First of all, our sample size of patients was very small, despite gathering of data during a 10-year period. The accuracy of statistical probabilities (ie, sensitivity) and ratios (ie, positive predictive value) is dependent on an accurate reflection of disease prevalence in the data set. Without a larger sample size of patients, the true prevalence of iliac vein thrombosis within the population is misrepresented. This limits our ability to make conclusive statements about the efficacy of DU in the general population. Second, this study was designed as a retrospective analysis of patients with a diagnosis of iliac vein thrombosis in their medical records. Thus, although sensitivity and positive predictive value can be determined on the basis of our results, we cannot comment on the specificity or negative predictive value of DU in diagnosis of iliac vein thrombosis. Nine patients with documented diagnosis of iliac vein thrombosis had negative results of imaging studies (both DU and cross-sectional imaging). Potential explanations for this include resolution of a previously diagnosed thrombus or simply an error in medical record documentation. To fully evaluate the accuracy of DU, it would be necessary to perform a prospective trial evaluating all patients with suspected iliac vein thrombosis with cross-sectional imaging and a specific pelvic venous protocol for DU; this would be the only way to accurately reflect the number of patients with both DU and crosssectional imaging findings negative for iliac vein thrombosis (true negative). Finally, DU is dependent on the experience and skill of the technician performing the study. A 10-year study period, even at the same institution, implies a lot of variability in the quality of the ultrasound technicians performing the studies. It is difficult to control for such variability, and it can certainly influence our results.
CONCLUSIONS This investigation raises the concern that current protocols of lower extremity venous DU are not efficient in
diagnosis of iliac vein thrombosis. Despite decades of experience with ultrasound and improvements in technical considerations and skills with this diagnostic modality, the poor sensitivity and subpar positive predictive value of our current protocol for lower extremity venous DU make it an unreliable tool for detecting iliac vein thrombosis. Patients with suspected iliac vein thrombosis should have a dedicated pelvic DU examination or cross-sectional imaging performed regardless of the findings seen on routine infrainguinal venous DU. We would like to acknowledge Irene B. Helenowski, PhD, of the Northwestern University Feinberg School of Medicine’s Department of Preventive Medicine for her contribution to the statistical analysis required in this study.
AUTHOR CONTRIBUTIONS Conception and design: AJ, SR, HR Analysis and interpretation: AJ, MS, KD, PA, ME, HR Data collection: AJ Writing the article: AJ, HR Critical revision of the article: AJ, MS, SR, KD, PA, ME, HR Final approval of the article: AJ, MS, SR, KD, PA, ME, HR Statistical analysis: AJ, HR Obtained funding: Not applicable Overall responsibility: HR
REFERENCES 1. Spritzer CE, Arata MA, Freed KS. Isolated pelvic deep venous thrombosis: relative frequency as detected with MR imaging. Radiology 2001;19:521-5. 2. Comerota AJ, Kamath V. Thrombolysis for iliofemoral deep venous thrombosis. Expert Rev Cardiovasc Ther 2013;11: 1631-8. 3. Dorfman GS, Cronan JJ. Venous ultrasonography. Radiol Clin North Am 1992;30:879-94. 4. Intersocietal Accreditation Commission. IAC standards and guidelines for vascular testing accreditation. Available at: https://www.intersocietal.org/vascular/seeking/vascular_ standards.htm. Accessed January 15, 2018. 5. Fletcher JP, Kershaw LZ, Barker DS, Koutts J, Varnava A. Ultrasound diagnosis of lower limb deep venous thrombosis. Med J Aust 1990;153:453-5. 6. Davidson BL, Elliott CG, Lensing AW. Low accuracy of color Doppler ultrasound in the detection of proximal leg vein thrombosis in asymptomatic high-risk patients. The RD Heparin Arthroplasty Group. Ann Intern Med 1992;117:735-8. 7. Laissy JP, Cinqualbre A, Loshkajian A, Henry-Feugeas MC, Crestani B, Riquelme C, et al. Assessment of deep venous thrombosis in the lower limbs and pelvis: MR venography versus duplex Doppler sonography. AJR Am J Roentgenol 1996;167:971-5. 8. Dronkers CE, Sramek A, Huisman MV, Klok FA. Accurate diagnosis of iliac vein thrombosis in pregnancy with magnetic resonance direct thrombus imaging (MRDTI). BMJ Case Rep 2016;2016. 9. Tamura K, Nakahara H. MR venography for the assessment of deep vein thrombosis in lower extremities with varicose veins. Ann Vasc Dis 2014;7:399-403.
Jain et al
Journal of Vascular Surgery: Venous and Lymphatic Disorders
729
Volume 6, Number 6 10. Lim KE, Hsu WC, Hsu YY, Chu PH, Ng CJ. Deep venous thrombosis: comparison of indirect multidetector CT venography and sonography of lower extremities in 26 patients. Clin Imaging 2004;28:439-44. 11. Kearon C, Ginsberg JS, Hirsh J. The role of venous ultrasonography in the diagnosis of suspected deep venous thrombosis and pulmonary embolism. Ann Intern Med 1998;129:1044-9. 12. Gupta R, Stouffer GA. Deep venous thrombosis: a review of the pathophysiology, clinical features, and diagnostic modalities. Am J Med Sci 2001;322:358-64. 13. Borris LC, Christiansen HM, Lassen MR, Olsen AD, Schott P. Comparison of real-time B-mode ultrasonography and bilateral ascending phlebography for detection of postoperative deep vein thrombosis following elective hip surgery. The Venous Thrombosis Group. Thromb Haemost 1989;61:363-5.
14. Shin SY, Kim H, Choi YS, Kim SW. Usefulness of Doppler waveform analysis before performing a complex procedure using femoral venous access. J Clin Ultrasound 2018;46:157-9. 15. Metzge PB, Rossi FH, Kambara AM, Izukawa NM, Saleh MH, Pinto IM, et al. Criteria for detecting significant chronic iliac venous obstructions with duplex ultrasound. J Vasc Surg Venous Lymphat Disord 2016;4:18-27. 16. Lin EP, Bhatt S, Rubens D, Dogra VS. The importance of monophasic Doppler waveforms in the common femoral vein: a retrospective study. J Ultrasound Med 2007;26:885-91. 17. Bach AM, Hann LE. When the common femoral vein is revealed as flattened on spectral Doppler sonography: is it a reliable sign for diagnosis of proximal venous obstruction? AJR Am J Roentgenol 1997;168:733-6.
Submitted Feb 20, 2018; accepted Jun 1, 2018.