Figure 1. Schematic diagram indicating transducer position on the right common femoral veins for the surgeon-performed, focused ultrasound examination.
Figure 2. Transverse ultrasound image of intraluminal thrombus of common femoral veins (CFVs). Arrow indicates echogenicity, which is consistent with thrombus. CFA indicates common femoral artery.
Rozycki GS, Tchorz KM, Riehle KJ, Hattaway A, Naidu DK. A Prospective Study of a Focused, Surgeon-Performed Ultrasound Examination for the Detection of Occult Common Femoral Vein Thrombosis in Critically Ill Patients. Arch Surg. 2004;139(3):275-280. doi:10.1001/archsurg.139.3.275
Copyright 2004 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2004
A focused, surgeon-performed ultrasound examination of the common femoral veins is an accurate screening tool for the detection of common femoral vein thrombosis in high-risk, critically ill patients.
A prospective study using a focused ultrasound examination for findings consistent with deep vein thrombosis of the common femoral veins. The results of these examinations were compared with those of duplex imaging or computed tomographic venography studies.
Surgical intensive care unit.
All critically ill patients who were admitted to the surgical intensive care unit and considered to be at high risk for the development of deep vein thrombosis.
Main Outcome Measure
Presence of deep vein thrombosis in the common femoral veins.
During a 16-month period, surgeons performed 306 ultrasound examinations on 220 critically ill surgical patients. The results included 295 true negative, 9 true positive, 1 false negative, and 1 false positive, yielding a 90.0% sensitivity, 99.6% specificity, and 99.3% accuracy.
A focused, surgeon-performed ultrasound examination is a rapid and accurate screening method to detect common femoral vein thrombosis in critically ill patients as well as to examine those patients in whom pulmonary embolism is strongly suspected.
The development of deep vein thrombosis (DVT) with its potential complications significantly increases morbidity and mortality for critically ill surgical patients. Although prophylactic agents and regimens have been shown to prevent DVT in most patients, they are not completely effective.1- 5 Therefore, additional measures such as sequential compression devices for prevention and serial duplex imaging of the lower extremities for early detection are used in select patients. Although sequential compression devices are widely available, serial duplex imaging, especially in a busy surgical intensive care unit (SICU), may be difficult to do routinely because of limited resources. Furthermore, the frequency with which such a study should be done to detect occult DVT is unknown. Considering the successful use of surgeon-performed, focused ultrasound examinations in patients with multiple acute conditions, it seems reasonable to use this technology as a screening tool for the detection of occult DVT in critically ill patients.6- 9
We hypothesized that serial focused ultrasound examinations performed by surgeons could accurately detect common femoral vein (CFV) thrombosis in critically ill surgical patients. The purpose of this study was to evaluate the efficacy of a limited ultrasound examination for the detection of occult CFV thrombosis.
During a 16-month period, surgeon-performed, focused ultrasound examinations were prospectively evaluated for the detection of CFV thrombosis at an urban level I trauma center. Entrance criteria included critically ill surgical patients who were considered to be at high risk for the development of DVT and whose anticipated SICU length of stay was 48 hours.3,5,10 Risk factors for development of DVT were being older than 45 years and having more than 3 days of bed rest, previous history of DVT or pulmonary embolus, spinal cord injury or spine fracture, severe closed head injury (Glasgow Coma Scale Score ≤8), pelvic or lower extremity fracture, complex wound of a lower extremity including major vascular injury, presence of femoral venous catheter, or massive transfusion. Surgeon-performed, focused ultrasound examinations were performed within 48 hours of the patient's SICU admission and then serially at a minimum of every 4 days. The focused ultrasound examination used only B (brightness)-mode ultrasound imaging defined as that in which the ultrasound waves or echo signals are electronically converted to intensity-modulated dots on the ultrasound screen. The higher the amplitude of the wave emitted from the transducer, the stronger the returning wave that forms the image on the ultrasound screen and the brighter the dot that makes up that image.11 Therefore, B-mode ultrasound, commonly used in other surgeon-performed, focused ultrasound examinations, defines anatomical but not the flow characteristics of a vessel. Results of the surgeon-performed ultrasound examinations were compared with the results of color-flow duplex imaging studies, which were performed by ultrasound technicians. Computed tomographic venography (CTV) was performed instead of the duplex study in a few patients who had acute signs and symptoms of pulmonary embolism and who were undergoing simultaneous computed tomographic angiographic testing. Both the duplex and CTV studies were interpreted by attending radiologists. Each patient had a daily physical examination by the trauma/surgical critical care attending surgeon (G. S. R.) or fellow (K. M. T.) who conducted this study. All patients received appropriate doses of either unfractionated heparin or low-molecular-weight heparin. Low-molecular-weight heparin was administered according to the Eastern Association for the Surgery of Trauma Practice Management Guidelines for the Management of Venous Thromboembolism in Trauma Patients.12 As per our Department of Surgery policy, all patients (except those who were actively bleeding or injured) received unfractionated heparin prior to their operations. The dose of unfractionated heparin was 5000 U administered subcutaneously every 12 hours or every 8 hours if the patient was morbidly obese. Patients who had a hip fracture or who were undergoing hip replacement surgery received 40 mg of enoxaparin sodium administered subcutaneously. Patients were followed up throughout their hospital course until discharge and through the rehabilitation phase. Patients whose anatomy precluded the performance of the ultrasound examinations were excluded from the study.
Experienced surgeon-sonographers (G. S. R. and K. M. T.) underwent additional ultrasound training under the direction of a senior registered vascular technician. The objectives of the training were to identify the common femoral artery, CFV, and saphenofemoral junction in longitudinal and transverse sections and to demonstrate the ease of compressibility of the normal CFV.
The focused ultrasound examinations were performed using an ultrasound machine with an 8-MHz linear-array transducer (2001 Leopard; B-K Medical Systems Inc, North Billerica, Mass) owned by the Department of Surgery of Emory University School of Medicine, Atlanta, Ga, and located in the SICU at Grady Memorial Hospital, Atlanta. Each ultrasound examination was conducted according to protocol, and each ultrasound image was appropriately labeled with the patient's medical record number. The patient was placed in the supine position with each leg slightly abducted and the knee slightly flexed. Ultrasound transmission gel was applied to the right side of the groin area, and the transducer was oriented for transverse sections and then placed approximately 1 cm below the inguinal ligament (Figure 1). The right common femoral artery was identified, and the gain setting on the ultrasound machine was adjusted, if needed, to ensure that the blood within the artery appeared dark (ie, echolucent). Next, the ultrasound transducer was moved slightly medially to identify the CFV at the level of the saphenofemoral junction. The CFV was examined for the presence or absence of intraluminal echogenicity (ie, internal echoes) consistent with a thrombus. Next, the surgeon applied pressure with the ultrasound transducer and noted the ease or lack of compressibility of the CFV. The normal CFV compresses easily and shows complete collapse of the lumen. When thrombus is present, however, coaptation does not occur13 (Figure 2). The transducer was rotated 90°, and the common femoral artery and CFV were imaged in sagittal (longitudinal) sections. Once again, the CFV was examined for the presence or absence of intraluminal echogenicity (thrombus) and compressibility. Attention was then turned to the left side of the groin where the examination was repeated. For this study, DVT was defined as an echogenic, constant intraluminal filling defect in the CFV that was visualized on both longitudinal and transverse views and incompressibility of the CFV.4
Ultrasound images were recorded in hard copy and videotape and later reviewed by an experienced surgeon who is also a registered diagnostic medical sonographer (G. S. R.). Results were categorized as true positive, true negative, false positive, and false negative (Table 1). This study was approved by the Emory University School of Medicine Human Investigations Committee.
During a 16-month period, 220 critically ill surgical patients (167 trauma and 53 nontrauma) underwent 306 surgeon-performed, focused ultrasound examinations. The mean ± SD age of the study population was 43.7 ± 17.9 years (range, 17-89 years), 132 patients (60%) were men, and their mean ± SD length of stay in the SICU was 20.2 ± 24.6 days (range, 2-266 days). The 167 injured patients (117 [70%] blunt; 50 [30%] penetrating) had a mean ± SD Injury Severity Score of 22.5 ± 11.1 (range, 4-51) and a mean ± SD SICU admission base deficit of −6.1 ± −6.3 (range, −33.3-7.4). The remaining 53 patients were admitted to the general, vascular, or subspecialty surgery services. The focused ultrasound examinations were performed at a mean ± SD interval of every 3.5 ± 2.8 days (range, 1-4 days), whereas the duplex studies were performed at a mean ± SD interval of every 6.3 ± 2.0 days (range, 4-7 days) but always within 4 days of the surgeon-performed, focused ultrasound examination.
Ultrasound examination results were 295 true negative, 9 true positive, 1 false negative, and 1 false positive, yielding a 90.0% sensitivity, 99.6% specificity, and 99.3% accuracy. The mean ± SD time for the surgeon-performed ultrasound examinations was 2.0 minutes ± 0.3 minutes (range, 1.8-4.0 minutes).
Despite the administration of prophylactic agents, asymptomatic DVT continues to occur in high-risk, critically ill patients. Although prevention is the best "treatment," emphasis on the early detection of DVT is worthwhile so that potential serious complications can be prevented. A variety of risk factors in the critically ill patient contribute to the development of DVT, and this limits the effectiveness of current preventive options.4,14- 17 Considering these issues, the early administration of prophylactic agents and routine screening for occult DVT have merit.
This study examined the efficacy of surgeon-performed, focused ultrasound examinations to detect occult thrombosis of the CFVs in high-risk, critically ill surgical patients. Although it is limited, our focused ultrasound examination provided some valuable information as an extension of the patient's physical examination while on rounds in the SICU.
Most of the ultrasound examinations performed by surgeons showed true-negative results, which was not unexpected because all of the patients received DVT prophylaxis. Because so few patients manifest findings consistent with DVT on their physical examinations, screening tests are not only worthwhile but necessary for early detection and treatment.2,3,18- 21
There were 9 true-positive examination results (9/220 [0.04%]), 6 confirmed by duplex imaging and 3 by CTV (Table 2). In general, the findings of an intraluminal echogenicity (thrombus) or incompressibility of the CFV by the surgeon prompted the immediate performance of a duplex study. The value of the focused examination in these patients therefore was that occult DVT was detected while on rounds in the SICU. This was usually sooner than the scheduled duplex imaging study. Furthermore, because of the positive findings on the surgeon-performed, focused ultrasound examination, anticoagulation was initiated in 1 patient (patient 3) (Table 3) before confirmation of his pulmonary embolism.
The 1 false-negative result in this study occurred in a patient who had undergone an aorto–right iliac and aorto–left common femoral artery bypass using a polytetrafluoroethylene graft. The surgeon-performed, focused ultrasound examination showed incompressibility of the patient's left CFV but absence of an intraluminal thrombus. A duplex imaging study performed on the same day was equivocal for DVT. Potential explanations for the false-negative result include the following: (1) The thrombus, although present, was not detectable on the focused ultrasound examination because it was an acute or fresh thrombus in early formation. The weak echogenicity of a fresh thrombus may preclude its visualization, but as the thrombus matures, it becomes more echogenic and is therefore more easily seen on the ultrasound examination.22 (2) A higher-frequency ultrasound transducer may have been used to perform the duplex study. The higher the frequency, the better the resolution of the thrombus. (3) The focus may have been improperly adjusted to detect the subtle echogenicity of the thrombus.
It is possible that all 3 factors contributed to the false-negative examination result, but the most likely cause was that the CFV contained a fresh or early thrombus. Ultrasound examinations (focused or duplex) were being performed at least every 4 days on these patients, and therefore, it is conceivable that most of these thrombi were detected at an early stage. These issues address the fact that the accuracy of ultrasound is user dependent and that the focused examination does not replace the duplex imaging study. Proper analysis of these events and continuing education on the principles of ultrasound physics may decrease similar false-negative findings in future examinations.
The 1 false-positive result in this study occurred in an obese yet muscular patient who developed multisystem organ dysfunction. The focused ultrasound examination showed that the left CFV was incompressible, and a duplex study performed the next day was equivocal for DVT. A radioactive fibrinogen uptake study showed normal findings and confirmed the absence of DVT.23
Ultrasound imaging of this obese yet muscular patient presented unique challenges that may have accounted for the false-positive examination result. The abundant adipose tissue and muscle in the patient's thigh made it difficult to compress the CFV. Although incompressibility of the vein is the main criterion for the diagnosis of DVT, the force necessary to cause the lumen to collapse completely is about 104 dyne/cm2, or about 3 lb.24 This amount of force, however, should not completely compress the artery. In the obese patient, more force may be needed to compress the vein, but that amount of force should be regulated by visualizing only a slight compression of the adjacent artery.
Although not related to this specific patient, another problem with the imaging of the obese patient is the difficulty in using a high-frequency transducer, which is needed to provide the fine resolution to visualize the intraluminal thrombus. The higher the frequency, the better the resolution but the less the penetration through the deep tissues. For patients with this body habitus, other diagnostic studies, such as the radioactive fibrinogen uptake study, may be needed.
The deficiencies of our study centered on the following areas: (1) limitations of the focused ultrasound examination, (2) the use of duplex imaging and CTV as the standard studies, (3) the variable frequency with which the screening studies were performed, and (4) the unique characteristics of our patient population.
The surgeon-performed, focused ultrasound examination used only B-mode imaging, which identifies the anatomical changes associated with DVT (ie, dilation of the vein [in comparison with the same vein in the opposite extremity and often a subtle difference], intraluminal thrombus, and incompressibility of the vein).13,25,26 The use of color Doppler may have improved the results of our study especially for those patients who had a nonocclusive thrombus. Although we considered using this modality, we wanted to design a study with the more affordable B-mode ultrasound, which is already used frequently by surgeons in the trauma and intensive care unit settings. Duplex imaging, however, identifies flow abnormalities, differentiates an acute (early) from a subacute thrombosis, evaluates nonocclusive thrombosis, and distinguishes between extrinsic compression and intrinsic obstruction.25,27 Because 1 patient was diagnosed with a popliteal vein thrombosis and another had only abnormal flow characteristics of the deep venous system, the value of the duplex test was underscored and it should not be eliminated from our diagnostic armamentarium.28,29 Our focused ultrasound examination, however, is rapid and emphasizes the incompressibility of the vein, which is considered to be the main criterion for the diagnosis of DVT.
In our study, duplex imaging was used as the standard for comparison of the surgeon-performed ultrasound examination results. In most patients, duplex imaging accurately detects thrombosis or flow abnormalities in the common femoral, superficial femoral, and popliteal veins. Its accuracy, however, is less for the detection of DVT in the superficial femoral vein in the Hunter canal and in the calf veins. Overall, the technical quality of the study is user dependent, and sensitivity varies depending on whether the patient is symptomatic. Collective reviews from Comerota et al30 and Agnelli et al31 showed that when duplex imaging was used to diagnose proximal DVT in symptomatic patients, the sensitivity and specificity were 96% but only about 61% (range, 51%-73%) and 97% (range, 95%-99%), respectively, for those patients who were asymptomatic. Furthermore, duplex imaging had an 80% sensitivity for the detection of calf DVT in symptomatic patients in whom the diagnosis of pulmonary embolism was being considered.
In a select group of patients, CTV was used to confirm the DVT. Because the patient was undergoing computed tomographic angiography to detect a pulmonary embolism, it was convenient and efficient to perform the CTV test at that time. Although CTV may be less sensitive than the duplex imaging study in detecting DVT, all of the CTV tests performed had positive results in this very select group of patients.32
The focused ultrasound examinations and the duplex studies were performed at various intervals but always within at least 4 days of each other. Ideally, both tests should have been performed on the same day to maximize the value of having the duplex as a standard test for comparison. This was not feasible, however, because only 2 surgeons performed the focused ultrasound examinations and duplex imaging was not available on weekends. Because of this issue, the study population was not a consecutive sample.
Our patient population was somewhat unique in that 3 of the patients whose examination results were positive had ligations of their external iliac veins. These patients had unilateral leg edema, and therefore, the positive examination results were not unexpected. Three other patients had previous central venous catheters in their CFVs. These patients represent a specific population who may need to undergo screening for DVT more often.
Finally, despite the use of these tests to screen for occult DVT, pulmonary embolism occurred in 4 patients (Table 3). One patient had signs and symptoms of a pulmonary embolus on admission to the SICU despite receiving appropriate preoperative prophylaxis. Considering the patient's multiple risk factors, he may have been a good candidate for preoperative screening for DVT, which could easily be accomplished in the office or clinic setting.
In general, data from this study underscore a conclusion from our previous work that ultrasound is a rapid and accurate diagnostic modality that can be used by surgeons as an extension of the physical examination.6,9,33 Similar to other studies, it has been well documented that surgeons who are properly trained in ultrasound can continue to build on their experience, perform focused examinations, and compare their results with those obtained with other diagnostic modalities. This modality is especially valuable in critically ill patients who present a challenge to the surgeon because the patient's altered sensorium, immobility, and multiple monitoring and traction devices can impede a thorough medical assessment.
Although our focused ultrasound examination is not as detailed as the more time-consuming duplex imaging study, it accurately detects CFV thrombosis and rapidly screens for DVT, as illustrated in the patient who was admitted emergently to the SICU with signs and symptoms of a pulmonary embolism.
From these data, we conclude that a focused ultrasound examination performed by surgeons is a rapid and accurate modality for detecting occult CFV thrombosis. We recommend that surgeons use this focused ultrasound examination to initially screen high-risk surgical patients for occult CFV thrombosis as well as to examine those patients in whom pulmonary embolism is strongly suspected.
Corresponding author and reprints: Grace S. Rozycki, MD, RDMS, Department of Surgery, Emory University School of Medicine, Suite 302, 69 Jesse Hill Dr SE, Atlanta, GA 30303 (e-mail: firstname.lastname@example.org).
Accepted for publication September 9, 2003.
We thank Samantha Bucknor, Grady Memorial Hospital, Atlanta, Ga, and Margaret Worley, Parkland Memorial Hospital, Dallas, Tex, for assistance in the preparation of this article. We also thank the surgical residents of Emory University School of Medicine and B-K Medical Systems Inc, North Billerica, Mass, ultrasonographers Ralph Winters, RDMS, and Dale Bentley, RVT, RDMS.
This study was presented at the Western Trauma Association Annual Meeting; February 27, 2002; Whistler-Blackcombe, Vancouver.