 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1998 The Journal of Bone and Joint Surgery, Inc. Ultrasound Surveillance for Asymptomatic Deep Venous Thrombosis after Total Joint Replacement* , HERSHEY , ROCHESTER, NEW YORK, HERSHEY, PENNSYLVANIAInvestigation performed at the Departments of Orthopaedics and Rehabilitation, Radiology, and Surgery, Pennsylvania State University, College of Medicine, The Milton S. Hershey Medical Center, Hershey, and the Department of Radiology, University of Rochester, Strong Memorial Hospital, Rochester
Prospective data on 202 consecutive patients who had a total of 123 total hip and ninety-four total knee arthroplasties were collected from two university medical centers. The findings of routine surveillance for deep venous thrombosis performed with ascending contrast venography were compared with those of surveillance with duplex ultrasonography complemented with color-flow Doppler imaging. All of the studies were performed between the third and seventh postoperative days. Of the 202 patients (342 extremities) who were examined, fifty-five (27 per cent) were found to have deep venous thrombosis; fifty-two (95 per cent) of the thrombi were in the calf and three (5 per cent) were in the proximal veins. All of the thrombi were clinically asymptomatic and all were non-occlusive, allowing passage of contrast medium around an intraluminal filling defect. Duplex ultrasonography with color-flow Doppler imaging correctly identified two of the three proximal thrombi and five of the fifty-two thrombi in the calf (sensitivity, 10 per cent). The sensitivity for the detection of thrombi in the calf was zero of sixteen at one of the institutions involved in the study and 14 per cent (five of thirty-six) at the other. There were two false-positive findings on ultrasonographic examination; one involved a proximal thrombus and one, a distal thrombus. We believe that the interinstitutional variability and insensitivity of duplex ultrasonography with color-flow Doppler imaging for the detection of asymptomatic deep venous thrombi in the calf after total joint replacement make it unreliable as a routine surveillance tool after total hip or knee arthroplasty.
Deep venous thrombosis is the most common complication after total hip or knee arthroplasty. The reported prevalence has ranged from 45 to 70 per cent after total hip arthroplasty and from 50 to 84 per cent after total knee arthroplasty in studies of 100 to 500 patients who had elective joint replacement without prophylaxis against thrombosis9,10,26,37. With currently accepted prophylaxis, the reported prevalence of deep venous thrombosis in series of 103 to 2161 patients10-12,15,27,37 has ranged from 20 to 30 per cent after total hip arthroplasty and from 22 to 50 per cent after total knee arthroplasty. Studies involving forty-eight to as many as 2000 patients who had documented deep venous thrombosis have shown pulmonary embolism to be responsible for mortality rates of 1.7 to 3.4 per cent5,17,28,37. There is an important relationship between the location of the deep venous thrombosis in the lower extremity and the likelihood of subsequent pulmonary embolism. It is widely accepted that proximal (femoropopliteal) deep venous thrombosis is associated with the greatest risk of embolization to the lung28. However, in series involving 100 to more than 900 patients14,29,34, 17 to 23 per cent of distal thrombi propagated to the thigh, and approximately 50 per cent of those resulted in pulmonary embolism. Multiple studies involving 100 to 200 patients each have shown that 40 to 60 per cent of all thrombi after total hip replacement, and as many as 95 per cent of those after total knee arthroplasty, involve the deep veins of the calf. With currently accepted prophylactic measures, the relative proportion of distal deep venous thrombi after total hip replacement has increased9,12,17,25,29,33. Furthermore, the decreasing duration of hospitalization after total joint replacement has correspondingly shortened the duration of prophylaxis that patients receive. In this setting, the need to recognize asymptomatic postoperative deep venous thrombosis accurately before the patient is discharged is greater. Because of its low morbidity and non-invasive nature, ultrasonography has become a popular modality for the surveillance of deep venous thrombosis after total hip and knee arthroplasty. Both duplex compression ultrasonography and color-flow Doppler imaging have been employed for this purpose. Although in some centers ultrasonography is considered effective for the diagnosis of spontaneous deep femoropopliteal venous thrombosis in symptomatic patients after total hip or knee replacement8,19,35,40, its general usefulness for postoperative surveillance for asymptomatic deep venous thrombosis has not been specifically proved. Few investigators have compared ultrasonography with contrast venography for the identification of deep venous thrombosis distal to the venous trifurcation. Wells et al., in a meta-analysis, reported sensitivities ranging from 20 to 88 per cent in studies that met criteria for a low level of bias, and the sensitivities appeared to be highly institution and operator-dependent39. The ability of ultrasonography to identify clinically important deep venous thrombosis in the postoperative setting remains controversial. The purpose of the present study was to compare the use of ultrasonography with that of contrast venography as a definitive surveillance tool for asymptomatic deep venous thrombosis after total joint arthroplasty.
Two hundred and twenty-six patients had a primary or revision total hip or knee replacement at one of two university medical centers during the study period. Twenty-four patients (11 per cent) were not examined with contrast venography because of an allergy to the contrast medium, a medical contraindication to contrast-medium load, or poor venous access in the lower extremity. The remaining 202 patients had surveillance for deep venous thrombosis performed with both ascending contrast venography and duplex ultrasonography complemented with color-flow Doppler imaging. These patients make up the present study group. One hundred and thirty-six patients (210 extremities) were managed at The Milton S. Hershey Medical Center, and sixty-six (132 extremities) were managed at the University of Rochester. Informed consent for bilateral surveillance studies was obtained from all patients. Current popular techniques for prophylaxis against deep venous thrombosis were used. Only the extremities for which both ultrasonography and venography were performed were included in the analysis. Bilateral ultrasonographic examination was performed for all of the patients. All of the patients from the University of Rochester were involved in another trial evaluating prophylaxis against deep venous thrombosis that dictated bilateral venographic screening. The surveillance studies were performed between the third and seventh postoperative days, and the two examinations were performed within twenty-four hours of each other. By design, ultrasonography was performed before venography. It was verified that the patients had normal renal function, and the patients were adequately hydrated before ascending venography was performed. Iso-osmolar non-ionic contrast medium was used in all patients. All of the venograms were made and interpreted by certified vascular radiologists at each institution, according to the criteria of DeWeese and Rogoff7. Each venogram was reviewed at the time of the evaluation and again at a later date in a blinded fashion according to study protocol. No intraobserver discrepancies with regard to the locations of thrombi were encountered after the repeat review of the venograms. The ultrasonographic examinations were performed by experienced vascular ultrasonographers at each institution in departments that had been certified by the Intersocietal Commission for Accreditation of Vascular Laboratories.
Three hundred and forty-two lower extremities (202 patients) in which both contrast venography and ultrasonography had been performed after an arthroplasty were analyzed. Two hundred and two extremities of 123 patients (seventy-six from The Milton S. Hershey Medical Center and forty-seven from the University of Rochester) were studied after the patient had been managed with total hip arthroplasty. The average age of these patients was sixty-three years; the average age of the fifty-one men was 56.1 years (range, forty to eighty-two years), and the average age of the seventy-two women was 67.5 years (range, fifty-five to eighty-five years). Ninety-four total knee replacements were performed in seventy-nine patients (sixty from the Hershey center and nineteen from the University of Rochester). Nine men and six women had a bilateral total knee arthroplasty performed under the same anesthetic. One hundred and forty of the extremities were studied after the patient had been managed with total knee arthroplasty. The average age of these patients was 66.6 years; the average age for the thirty-five men was 66.8 years (range, forty-eight to eighty-three years), and the average age for the forty-four women was 66.2 years (range, fifty-three to seventy-nine years).
Individual Institutional Data
The overall sensitivity of ultrasonography with regard to the detection of thrombi was 16 per cent (six of thirty-eight thrombi that were identified with venography were detected with ultrasonography) at the Hershey center, and the overall sensitivity for the detection of thrombi was one of seventeen at the University of Rochester. The sensitivity for thrombi in the calf was poor at both institutions; it was 14 per cent (five of thirty-six) at the Hershey center and zero of sixteen at the University of Rochester. One of two proximal thrombi were identified at the Hershey center and one of one proximal thrombus was identified at the University of Rochester. The positive and negative predictive values for ultrasonography were six of eight and 85 per cent, respectively, at the Hershey center, and one of one and 88 per cent, respectively, at the University of Rochester. At the Hershey center, there were two false-positive findings on ultrasonographic examination. Duplex ultrasonography identified a non-occluding thrombus in the popliteal vein in one patient after total hip arthroplasty, whereas venography revealed chronic deep venous thrombosis in the posterior and anterior tibial veins without evidence of acute thrombosis. A thrombus in the calf was identified with ultrasonography after a primary total knee arthroplasty in the other patient; the venogram showed normal findings. Both patients were discharged without prophylaxis, on the basis of the venographic findings, and no complications ensued during the six-month clinical follow-up interval.
Combined Data Of the 202 extremities (123 patients) that had surveillance for deep venous thrombosis after total hip arthroplasty, fourteen were found to have thrombosis with venography: one thrombus was in the thigh, and thirteen were in the calf. Nine of these thrombi were found in the involved extremity, and five were found in the uninvolved extremity. Ultrasonography correctly identified one proximal thrombus and two thrombi in the calf (Table I). In the patients who had total hip arthroplasty, ultrasonography demonstrated a sensitivity of one of one, a specificity of 99.5 per cent, and an accuracy of 99.5 per cent in the diagnosis of deep venous thrombosis in the thigh compared with a sensitivity of two of thirteen, a specificity of 100 per cent, and an accuracy of 94.6 per cent in the diagnosis of deep venous thrombosis in the calf (Table II). Overall, ultrasonography demonstrated a sensitivity of three of fourteen, a specificity of 99.5 per cent, and an accuracy of 94.1 per cent when used to diagnose deep venous thrombosis after total hip arthroplasty (Fig. 1).
Forty-one thrombi were detected with venography in the seventy-nine patients who had a total knee arthroplasty (fifteen of whom had a bilateral procedure). Two thrombi were found in the thigh and thirty-nine, in the calf. Thirty-four thrombi (83 per cent) were found in the involved extremity, and seven (17 per cent) were found in the uninvolved extremity. Ultrasonography accurately detected one proximal thrombus and only three thrombi in the calf (Table I). Compared with venography, ultrasonography demonstrated a sensitivity of one of two, a specificity of 100 per cent, and an accuracy of 99.3 per cent when used to diagnose proximal deep venous thrombosis and a sensitivity of 8 per cent (three of thirty-nine), a specificity of 99.0 per cent, and an accuracy of 73.6 per cent when used to diagnose deep venous thrombosis in the calf (Table II). Overall, ultrasonography demonstrated a sensitivity of 10 per cent (four of forty-one thrombi), a specificity of 98 per cent, and an accuracy of 72 per cent when used to detect deep venous thrombosis after total knee arthroplasty (Fig. 1). With all data combined, ultrasonography demonstrated a sensitivity of two of three, a specificity of 99.7 per cent, an accuracy of 99.4 per cent, a positive predictive value of two of three, and a negative predictive value of 99.7 per cent, compared with ascending contrast venography, in the diagnosis of three proximal deep venous thrombi (Figs. 2 and 4). It demonstrated a sensitivity of 10 per cent (five of fifty-two), a specificity of 99.7 per cent, an accuracy of 86.0 per cent, a positive predictive value of five of six, and a negative predictive value of 86 per cent in the diagnosis of fifty-two deep venous thrombi in the calf (Figs. 3 and 4). With regard to all deep venous thrombi, ultrasonography demonstrated a sensitivity of 13 per cent (seven of fifty-five), a specificity of 99.7 per cent, an accuracy of 85.4 per cent, a positive predictive value of seven of nine, and a negative predictive value of 86 per cent when used for surveillance after total hip or knee arthroplasty (Fig. 4).
The purpose of surveillance for deep venous thrombosis after total joint arthroplasty is to identify patients who are at risk for pulmonary embolism because of thrombosis and therefore need continued anticoagulation therapy as well as those patients who do not have thrombosis and therefore might not need continued anticoagulation therapy. In reality, no prophylaxis provides complete protection against thromboembolic disease and therefore every treatment strategy represents a balance between relative risks and benefits. Despite currently accepted prophylaxis with warfarin, thrombi still developed in the deep venous system of 22 per cent (seventy-eight) of 347 patients33 after total hip replacement and in 35 per cent (seventeen of forty-eight patients17) to 57 per cent (267 of 468 patients37) after total knee replacement. Also, in today's cost-conscious health-care environment, truncated hospital stays sharply decrease the duration of monitored prophylaxis against deep venous thrombosis on an inpatient basis; thus, prompt recognition of venous thrombosis at the time of discharge is essential to the strategy of selective postdischarge anticoagulation. Ultrasonography is an attractive and commonly employed tool for surveillance because it is associated with low morbidity and is non-invasive. While duplex ultrasonography alone allows compression augmentation for the identification of deep venous thrombosis, color-flow Doppler imaging also demonstrates the direction of blood flow and the reduction in luminal diameter, which can potentially enhance sensitivity in the diagnosis of deep venous thrombosis. However, we have demonstrated the inability of ultrasonography with color-flow Doppler imaging to detect deep venous thrombi in the calf that were demonstrated by venography. Thus, ultrasonography is a poor method for the identification of asymptomatic deep venous thrombosis in the early postoperative period. This analysis has some shortcomings. The limited number of institutions involved may have biased the results; however, both institutions are fully accredited and nationally recognized vascular ultrasonography teaching centers. Furthermore, The Milton S. Hershey Medical Center is responsible for the continuing education and licensing of technologists across the country. We expect the quality of ultrasonography at these institutions to be better than average; however, the sensitivities still did not meet the criteria of acceptability. Our early surveillance—that is, at three to seven days postoperatively—was necessary because of the current trend of reduced duration of hospital stays after joint replacement. Ultrasonographic examinations were performed routinely before venography to reduce the possibility of the ultrasonographic detection of contrast-medium-induced thrombi. Our experience has demonstrated no difference in the prevalence of thrombosis at the early time-period compared with that at the more traditional screening interval of seven to ten days postoperatively. In any event, the two screening studies were performed within twenty-four hours of one another so the early screening interval should not have introduced a bias for either screening modality. The overall prevalence of deep venous thrombosis of 27 per cent (fifty-five of 202 patients) in the present study is consistent with prevalences reported in the literature15,41; however, there was a difference in the distribution of the thrombi. Historically, 50 to 60 per cent of thrombi are identified in the proximal veins after total hip arthroplasty compared with only 10 per cent after total knee arthroplasty14,16,24,29,37. In the present study, only one of fourteen thrombi occurred in the proximal veins and thirteen occurred distal to the popliteal vein after total hip arthroplasty. The same trend was seen after total knee arthroplasty; two (5 per cent) of forty-one thrombi occurred in the proximal veins and thirty-nine (95 per cent) occurred in the calf veins. The implication is that current methods of prophylaxis have altered the natural distribution of deep venous thrombosis and are more effective in preventing thrombosis in the thigh than in reducing the prevalence of thrombosis in the calf11,31. As more than 90 per cent of the deep venous thrombi occurred in the calf, it is evident that any tool used for contemporary surveillance must be adequately sensitive to detect such thrombi. The treatment of distal thrombi remains controversial. Lotke et al.25 noted deep venous thrombosis after 72 per cent (126) of 175 total knee arthroplasties, with most of the thrombi located in the calf. However, the low prevalence of clinically important pulmonary emboli in that study led the authors to recommend that therapeutic anticoagulation be deferred when patients have thrombi in the calf alone. Other authors have refuted this opinion. One of us and colleagues33 observed that four of thirteen patients who had an untreated thrombus in the calf were seen for a pulmonary embolus and estimated that one in five pulmonary emboli arises from the calf. Similarly, patients who had distal deep venous thrombosis were observed to have a higher risk of pulmonary embolism than those who did not have deep venous thrombosis14. In more recent studies, distal thrombi propagated to the thigh in 17 per cent (seven) of forty-one patients to 23 per cent (five) of twenty-two patients14,28,29. These data prompted several authors to recommend that patients who have distal deep venous thrombosis be managed with anticoagulant prophylaxis in order to prevent proximal propagation15,32,34. The high prevalence of asymptomatic thrombi in the calf after hip and knee replacement places a large number of patients at risk for proximal propagation of the clot as well as for pulmonary embolism if the thrombi are left undiagnosed and untreated. Although our preference is to treat asymptomatic thrombi in the calf, we believe that extended thromboembolic prophylaxis should be confined to patients who have confirmed thrombosis. Paiement et al.30 recommended that all patients routinely receive a twelve-week postoperative course of low-dose warfarin in lieu of routine surveillance. While it effectively reduces the rate of thromboembolic disease, this strategy exposes patients to the risks of outpatient anticoagulant therapy. Coventry et al.5 reported bleeding complications in eighty-three (4.3 per cent) of 1950 patients after total hip arthroplasty. More recently, Landefeld and Beyth22, in a meta-analysis of more than 4000 patients from twenty-five studies, noted a 3 per cent average annual frequency of major bleeding complications and also found that this risk was elevated tenfold during the first month of anticoagulant therapy. In addition, while the role of increased age as a risk factor for warfarin-related bleeding has not been determined, the increased prevalence of adverse drug reactions, the presence of comorbitities and polypharmacy, and the increased vascular and endothelial fragility in older patients have been associated with an increased risk of bleeding complications3. These data exemplify the desirability of obtaining an accurate diagnosis of thrombosis, before the patient is discharged from the hospital, in order to limit unnecessary anticoagulant prophylaxis. Ascending venography has been the time-honored so-called gold standard for the diagnosis of deep venous thrombosis. While the use of ionic contrast agents produced occasional thrombotic complications in the past, the recent introduction of iso-osmolar, non-ionic contrast media has reduced this risk20,21. The advantage of venography compared with currently available non-invasive techniques is its ability to identify distal thrombi and non-occlusive thrombi. Although the risk of contrast-medium-induced phlebitis is diminished, allergic reactions to the contrast medium, renal intolerance of the contrast-medium load, and the inability to cannulate the veins of a swollen foot postoperatively still contribute to the morbidity associated with contrast venography. Contrast venography could not be performed for twenty-four (11 per cent) of our original 226 patients for a number of reasons; this dropout rate is representative of our experience over the years, with 10 to 15 per cent of patients being unable to be screened with venography32. Such a dropout rate for venography compares favorably with reported rates of technically inadequate images of the calf with color-flow Doppler imaging; those rates have ranged from 18 per cent (thirty-two of 177)8 to 44 per cent (thirty-four of seventy-eight)35. The sensitivity of magnetic resonance venography in detecting deep venous thrombosis has been evaluated in recent studies4,36, but accuracy, cost, and availability must all be addressed before this modality can be considered as the surveillance method of choice. The role of ultrasonography in the diagnosis of deep venous thrombosis has been examined previously2,13,14,23,42. With a sensitivity in the detection of proximal thrombosis in symptomatic patients approaching 97 per cent, and suggestions of similar success in detecting symptomatic thrombosis in the calf, the technique has enjoyed a rapid increase in popularity. Extension of the indications for this modality to postoperative surveillance was only natural, if not inevitable, but these indications have not been rigorously validated. In a study of 152 extremities after total hip replacement, Woolson et al.43 noted that ultrasonography identified seventeen of nineteen asymptomatic proximal deep venous thrombi that were demonstrated by venography. There were two false-negative ultrasonographic examinations, and two additional ultrasonographic scans were considered uninterpretable. Eight additional limbs had an isolated deep venous thrombus in the calf that was evident on venography but not on ultrasonography. Grady-Benson et al.14 established the postoperative reliability of duplex ultrasonographic technology in an even smaller series of patients; seven of seven proximal thrombi and seven of eight thrombi in the calf that were seen with venography were correctly identified with ultrasonography. Two additional thrombi in the calf were suspected on the basis of ultrasonography, but the findings on venography were negative (positive predictive value, seven of nine). Garino et al.13 demonstrated a so-called learning curve for the use of ultrasonography; initially, none of seven proximal thrombi were seen with ultrasonography and there were three false-positive studies for patients who had negative findings on venography. With experience, the technician was able to identify five of five proximal thrombi with ultrasonography and two examinations were false-positive (positive predictive value, five of seven). Collectively, these three widely cited studies13,14,43 account for a total of thirty-eight proximal thrombi (sensitivity, 76 per cent; twenty-nine of thirty-eight) and only sixteen thrombi in the calf (sensitivity, seven of sixteen). However, despite the absence of more substantive data, the orthopaedic community has adopted ultrasonography as a reliable screening examination for the assessment of thromboembolic risk after total joint arthroplasty. A more critical analysis was recently conducted by Wells et al.39, who performed a meta-analysis of seventeen studies to determine the accuracy of ultrasonographic screening of asymptomatic patients after orthopaedic procedures. The reported sensitivity for the detection of asymptomatic proximal thrombosis ranged widely, from 38 to 100 per cent. The reported sensitivity of ultrasonography in the detection of distal thrombosis ranged from 20 to 88 per cent6,24,29,35,38,43, but only two studies had adequate venographic controls29,35. It has been documented that the sensitivity of ultrasonography in the detection of deep venous thrombosis improves with the experience of the technician13,24,42, but few technicians in vascular laboratories in the United States have the experience to achieve this reported improvement in sensitivity. At the institutions in the present study, both of which are nationally recognized teaching centers for technologists, the overall sensitivity in the detection of deep venous thrombosis was 13 per cent (seven of fifty-five): two of three for proximal thrombosis and 10 per cent (five of fifty-two) for distal thrombosis. These discouraging results may be explained by the observation that asymptomatic postoperative thrombi in the calf are frequently only partially occlusive and do not obstruct flow1,6,43. Jongbloets et al.18, in a study of the use of ultrasonography in the diagnosis of deep venous thrombosis, found that most thrombi that were missed with ultrasonography were non-occlusive and were less than five centimeters long. This low sensitivity and interinstitutional variability make ultrasonography an unreliable screening modality for asymptomatic deep venous thrombosis after total joint replacement. The increased prevalence of deep venous thrombosis in the calf after either total hip or total knee arthroplasty, with its attendant risk of proximal propagation, warrants an appropriate treatment strategy to minimize the risk of embolic events. While extended prophylaxis for all patients effectively reduces the likelihood of embolic complications, we prefer to avoid exposing most patients who do not have any venous thrombosis to the risks of protracted anticoagulation. This approach of using postdischarge therapeutic anticoagulation only for patients who have a known deep venous thrombosis requires surveillance techniques that are sensitive to distal venous thrombosis. Of the fifty-five thrombi that were confirmed with venography in the present study, fifty-two (95 per cent) were confined to the calf and only five (10 per cent) of these were identified with duplex ultrasonography with color-flow Doppler imaging. While ultrasonography identified two of the three proximal thrombi, the small number of proximal thrombi precludes any meaningful statement concerning the utility of ultrasonography with color-flow Doppler imaging in detecting such thrombi. Nonetheless, we have not found ultrasonography to be sufficiently sensitive in detecting distal thrombi for it to be useful as a surveillance tool after total joint arthroplasty. It follows that, without individual institutional validation, ultrasonography cannot be endorsed as an acceptable method of surveillance for deep venous thrombosis in the postoperative period. This has important implications both for patient management and for the conduct of trials concerned with prophylaxis against deep venous thrombosis. Should routine surveillance and selective postdischarge anticoagulation therapy be chosen as part of a treatment strategy for prophylaxis against thromboembolism after total joint replacement, ascending contrast venography remains the most reliable screening modality until better non-invasive techniques become available. Alternatively, a strategy of extended outpatient anticoagulant prophylaxis may be adopted, along with its attendant risks of bleeding complications, obviating the need for routine surveillance studies.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
  CiteULike  Connotea  Del.icio.us  Technorati What's this?
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |